mirrors/git-annex
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
git-annex/Test.hs

1728 lines
70 KiB
Haskell
Raw Normal View History

added some toplevel git-annex subcommand tests note that test coverage doesn't work for those yet
2011-01-07 00:09:49 +00:00
{- git-annex test suite
-
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
- Copyright 2010-2018 Joey Hess <id@joeyh.name>
added some toplevel git-annex subcommand tests note that test coverage doesn't work for those yet
2011-01-07 00:09:49 +00:00
-
- Licensed under the GNU GPL version 3 or higher.
-}
try building test suite on windows
2013-05-17 20:45:44 +00:00
{-# LANGUAGE CPP #-}
embed test suite into git annex; available by running: git annex test I have seen some other programs do this, and think it's pretty cool. Means you can test wherever it's deployed, as well as at build time. My other reason for doing it is less happy. Cabal's handling of test suites sucks, requiring duplicated info, and even when that's done, it fails to preprocess hsc files here. Building it in avoids that and avoids having to explicitly tell cabal to enable test suites, which would then make it link the test executable every time, which is unnecessarily slow. This also has the benefit that now "make fast test" does a max speed build and tests it.
2013-02-27 19:35:07 +00:00
module Test where
test: Added --keep-failures option.
2016-01-06 17:44:12 +00:00
import Types.Test
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
import Types.RepoVersion
Split Test.hs and avoid optimising it much, to need less memory to compile. The ghc options were found by Sean Whitton; the debian arm autobuilders need those to build w/o OOM, and it seems to involve llvm using too much memory to optimize Test. This commit was sponsored by Boyd Stephen Smith Jr. on Patreon.
2018-02-18 15:48:48 +00:00
import Test.Framework
wire tasty's option parser into the main program option parser This makes bash completion work for git-annex test, and is generally cleaner.
2015-07-13 17:19:20 +00:00
import Options.Applicative.Types
test suite partially converted to use tasty test framework This is a win. Will need to wait for tasty getting into Debian, and do a trivial conversion of the remainder of the hunit tests.
2013-08-06 04:12:06 +00:00
import Test.Tasty
import Test.Tasty.Runners
import Test.Tasty.HUnit
import Test.Tasty.QuickCheck
Revert "temporarily revert tasty-rerun support for this release" This reverts commit 8fcf041159d06986b5a11a3eddd115b4c71a39d8.
2014-01-27 18:22:44 +00:00
import Test.Tasty.Ingredients.Rerun
test suite infra for testing mocked ssh remotes This commit was supported by the NSF-funded DataLad project.
2017-03-17 23:13:52 +00:00
import Options.Applicative (switch, long, help, internal)
move quickcheck Arbitrary declaration into test suite So git-annex can build w/o quickcheck installed.
2011-04-25 19:28:41 +00:00
test unused et al
2011-01-12 05:58:23 +00:00
import qualified Data.Map as M
Removed dependency on json library; all JSON is now handled by aeson. I've eyeballed all --json commands, and the only difference should be that some fields are re-ordered.
2016-07-26 23:15:34 +00:00
import qualified Data.ByteString.Lazy.UTF8 as BU8
test: Avoid most situations involving failure to delete test directories By forking a worker process and only deleting the test directory once it exits. This way, if a test leaves files open, they'll get closed when the worker exits, so avoiding failure to delete open files on Windows, and failure to delete directories due to NFS lock files. If a test leaves a git worker process running, the closed pipes should cause the worker to exit too, also avoiding the problem there. The 10 second sleep ought to give plenty of time for such worker processes to exit, although this is of course a race. Finally, even if test directory fails to be deleted still, it won't appear as if the last test in the test suite failed; the error will be displayed at the very end. This commit was supported by the NSF-funded DataLad project.
2017-08-14 20:27:35 +00:00
import System.Environment
better test framework
2010-11-02 21:08:16 +00:00
factor out common imports no code changes
2011-10-04 02:24:57 +00:00
import Common
test suite infra for testing mocked ssh remotes This commit was supported by the NSF-funded DataLad project.
2017-03-17 23:13:52 +00:00
import CmdLine.GitAnnex.Options
split groups of related functions out of Utility
2011-08-22 20:14:12 +00:00
factor out common imports no code changes
2011-10-04 02:24:57 +00:00
import qualified Utility.SafeCommand
test locationlog contents too
2011-01-11 19:43:29 +00:00
import qualified Annex
starting to work on test suite for v6
2015-12-15 21:19:26 +00:00
import qualified Annex.Version
split out two more Git modules
2011-12-13 19:22:43 +00:00
import qualified Git.Filename
add regression test for symlink calculation Note: Test reordered because running git-annex sync early broke the environment for some other tests.
2014-02-07 00:33:30 +00:00
import qualified Git.Types
much improved test and real fix for FAT symlink loss on conflicted merge I think that 751f496c11b421f498d19d4ea939bfd84125d349 didn't quite manage to actually fix the bug, although I have not checked since its "fix" got redone. The test suite now actually checks the file staged in git is a symlink, rather than relying on the bug casing a later sync failure. This seems a more reliable way to detect it, and probably avoids a heisenbug in the test suite.
2014-03-04 18:31:26 +00:00
import qualified Git.Ref
import qualified Git.LsTree
fix test case to convert slashes for DOS
2014-03-04 19:06:40 +00:00
import qualified Git.FilePath
remove 163 lines of code without changing anything except imports
2016-01-20 20:36:33 +00:00
import qualified Annex.Locations
avoid build warning on windows
2017-10-24 16:24:06 +00:00
#ifndef mingw32_HOST_OS
add annex-ignore-command and annex-sync-command configs Added remote configuration settings annex-ignore-command and annex-sync-command, which are dynamic equivilants of the annex-ignore and annex-sync configurations. For this I needed a new DynamicConfig infrastructure. Its implementation should be as fast as before when there is no dynamic config, and it caches so shell commands are only run once. Note that annex-ignore-command exits nonzero when the remote should be ignored. While that may seem backwards, it allows using the same command for it as for annex-sync-command when you want to disable both. This commit was sponsored by Trenton Cronholm on Patreon.
2017-08-17 16:26:14 +00:00
import qualified Types.GitConfig
avoid build warning on windows
2017-10-24 16:24:06 +00:00
#endif
test
2012-12-20 03:43:15 +00:00
import qualified Types.TrustLevel
successfully split Annex and AnnexState out of TypeInternals
2011-01-26 01:49:04 +00:00
import qualified Types
factored out a generic MapLog from uuid-based logs UUIDBased is just a MapLog with a UUID for the field.
2014-03-15 17:44:31 +00:00
import qualified Logs.MapLog
reorganize log modules no code changes
2011-10-15 20:21:08 +00:00
import qualified Logs.Trust
import qualified Logs.Remote
fix build
2012-05-06 00:11:08 +00:00
import qualified Logs.Unused
quickcheck test for transfer info read/write code Fixed a bug the quickcheck turned up.
2012-12-19 20:15:39 +00:00
import qualified Logs.Transfer
more quickcheck fun and the code gets better..
2012-12-20 02:14:12 +00:00
import qualified Logs.Presence
disable debug output in test suite Added during a dev session in Jan and kept accidentially.
2015-06-17 17:49:29 +00:00
import qualified Logs.PreferredContent
data types and serialization for metadata A very haskell commit! Just data types, instances to serialize the metadata to a nice format, and QuickCheck tests. This commit was sponsored by Andreas Leha.
2014-02-12 21:54:28 +00:00
import qualified Types.MetaData
convert test to Remote
2011-03-27 20:58:28 +00:00
import qualified Remote
factor non-type stuff out of Key
2017-02-24 17:42:30 +00:00
import qualified Key
a few fixes for the test suite in direct mode Still far from working.
2013-05-15 21:44:30 +00:00
import qualified Config
split cost out into its own module Added a function to insert a new cost into a list, which could be used to asjust costs after a drag and drop.
2013-03-13 20:16:01 +00:00
import qualified Config.Cost
add test to ensure hmac remains stable
2011-04-21 20:56:24 +00:00
import qualified Crypto
Fix a git-annex test failure when run on NFS due to NFS lock files preventing directory removal. Should fix this: lock (v6 --force): FAIL Exception: .git/annex/keys: removeDirectoryRecursive: unsatisfied constraints (Directory not empty) Verified that the test case still catches the regression it's meant to. This commit was supported by the NSF-funded DataLad project.
2017-08-14 19:11:42 +00:00
import qualified Database.Keys
reorg
2015-12-15 19:34:28 +00:00
import qualified Annex.WorkTree
reorg
2014-01-26 20:36:31 +00:00
import qualified Annex.Init
add regression test for symlink calculation Note: Test reordered because running git-annex sync early broke the environment for some other tests.
2014-02-07 00:33:30 +00:00
import qualified Annex.CatFile
don't assume git-annex is in path when calling itself from test suite
2016-02-16 20:29:04 +00:00
import qualified Annex.Path
Make test suite work better when the temp directory is on NFS. Deleting directories is one of the great unsolved problems of CS, thanks to abominations like NFS lock files and Windows and races with other processes cleaning up after themselves in the background. The gpg test harness sometimes failed to delete its temp directory on NFS. Avoid the problem class by not deleting it at all, and putting it inside the tmp repo being tested. The test suite's more robust (and/or nonsensical) workarounds for deleting its test dir will thus be used, hopefully avoiding the problem until an OS finds a new way to violate POSIX and the laws of nature. Note that this means that the .gnupg directory will be on whatever filesystem the test suite is being run on, which may be a lesser quality filesystem than gpg is really expecting. Gpg does not seem to need to write sockets etc to there so this seems ok. The only known problem is that if the filesystem forces a directory mode like 777, gpg will warn about unsafe home directory perms, but it still works.
2018-12-19 15:56:39 +00:00
import qualified Annex.Perms
Added GIT_ANNEX_VECTOR_CLOCK environment variable Can be used to override the default timestamps used in log files in the git-annex branch. This is a dangerous environment variable; use with caution. Note that this only affects writing to the logs on the git-annex branch. It is not used for metadata in git commits (other env vars can be set for that). There are many other places where timestamps are still used, that don't get committed to git, but do touch disk. Including regular timestamps of files, and timestamps embedded in some files in .git/annex/, including the last fsck timestamp and timestamps in transfer log files. A good way to find such things in git-annex is to get for getPOSIXTime and getCurrentTime, although some of the results are of course false positives that never hit disk (unless git-annex gets swapped out..) So this commit does NOT necessarily make git-annex comply with some HIPPA privacy regulations; it's up to the user to determine if they can use it in a way compliant with such regulations. Benchmarking: It takes 0.00114 milliseconds to call getEnv "GIT_ANNEX_VECTOR_CLOCK" when that env var is not set. So, 100 thousand log files can be written with an added overhead of only 0.114 seconds. That should be by far swamped by the actual overhead of writing the log files and making the commit containing them. This commit was supported by the NSF-funded DataLad project.
2017-08-14 17:55:38 +00:00
import qualified Annex.VectorClock
filter branches (incomplete) Promosing work toward metadata driven filter branches. A few methods to construct them are stubbed out; all the data types and pure code seems good. This commit was sponsored by Walter Somerville.
2014-02-16 21:39:54 +00:00
import qualified Annex.View
change directory encoding in ViewedFile such that the original directory can be extracted from it
2014-02-22 18:54:53 +00:00
import qualified Annex.View.ViewedFile
add git annex view command (And a vpop command, which is still a bit buggy.) Still need to do vadd and vrm, though this also adds their documentation. Currently not very happy with the view log data serialization. I had to lose the TDFA regexps temporarily, so I can have Read/Show instances of View. I expect the view log format will change in some incompatable way later, probably adding last known refs for the parent branch to View or something like that. Anyway, it basically works, although it's a bit slow looking up the metadata. The actual git branch construction is about as fast as it can be using the current git plumbing. This commit was sponsored by Peter Hogg.
2014-02-18 21:38:23 +00:00
import qualified Logs.View
split groups of related functions out of Utility
2011-08-22 20:14:12 +00:00
import qualified Utility.Path
fix
2011-09-28 19:17:45 +00:00
import qualified Utility.FileMode
fold Build/SysConfig.hs into BuildInfo via include This avoids warnings from stack about the module not being listed in the cabal file. So, the generated file is also renamed to Build/SysConfig. Note that the setup program seems to be cached despite these changes; I had to cabal clean to get cabal to update it so that Build/SysConfig was written. This commit was sponsored by Jochen Bartl on Patreon.
2017-12-14 16:46:57 +00:00
import qualified BuildInfo
handle C-style escapes in Format I was happily able to repurpose some code from Git.Filename to handle this. I remember writing that code... a whole afternoon at a coffee shop, after which I felt I'd struggled with Haskell and git, and sorta lost, in needing to write this nasty peice of code. But was also pleased at the use of a pair of functions and quickcheck that allowed me to get it 100% right. So, turns out I not only got it right, but the code wasn't as special-purpose as I'd feared. Yay!
2011-12-23 00:14:35 +00:00
import qualified Utility.Format
broke out Verifiable to a utility library, and added a quickcheck test
2012-09-08 03:23:52 +00:00
import qualified Utility.Verifiable
test: Fix threaded runtime hang. There was one forkProcess lurking in test.hs, and that seems to be responsible for recent buildd failures on amd64 and armhf. I was able to reproduce it pretty easily on amd64, and even once on i386, and it was clearly that same bad old threaded runtime hang. So removing this forkProcess should fix it. Odd that it lurked for some months before popping up.
2012-10-11 17:45:09 +00:00
import qualified Utility.Process
Bug fix: A recent change caused git-annex-shell to crash.
2012-10-16 02:22:40 +00:00
import qualified Utility.Misc
split out Utility.InodeCache
2013-02-14 20:17:40 +00:00
import qualified Utility.InodeCache
clean up from windows porting
2013-05-11 22:23:41 +00:00
import qualified Utility.Env
finally really add back custom-setup stanza Fourth or fifth try at this and finally found a way to make it work. Absurd amount of busy-work forced on me by change in cabal's behavior. Split up Utility modules that need posix stuff out of ones used by Setup. Various other hacks around inability for Setup to use anything that ifdefs a use of unix. Probably lost a full day of my life to this. This is how build systems make their users hate them. Just saying.
2017-12-31 20:08:31 +00:00
import qualified Utility.Env.Set
Fix bug in parsing of parens in some preferred content expressions. This fixes the behavior of the manual mode group. The current manual mode preferred content expression is: "present and (((exclude=*/archive/* and exclude=archive/*) or (not (copies=archive:1 or copies=smallarchive:1))) or (not copies=semitrusted+:1))" The old matcher misparsed this, to basically: OR (present and (...)) (not copies=semitrusted+:1)) The paren handling and indeed the whole conversion from tokens to the matcher was just wrong. The new way may not be the cleverest, but I think it is correct, and you can see how it pattern matches structurally against the expressions when parsing them. That expression is now parsed to: MAnd (MOp <function>) (MOr (MOr (MAnd (MOp <function>) (MOp <function>)) (MNot (MOr (MOp <function>) (MOp <function>)))) (MNot (MOp <function>))) Which appears correct, and behaves correct in testing. Also threw in a simplifier, so the final generated Matcher has less unnecessary clutter in it. Mostly so that I could more easily read & confirm them. Also, added a simple test of the Matcher to the test suite. There is a small chance of badly formed preferred content expressions behaving differently than before due to this rewrite.
2013-05-25 01:33:54 +00:00
import qualified Utility.Matcher
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
import qualified Utility.Hash
import qualified Utility.Scheduled
reorg quickcheck to a separate module
2015-11-17 19:49:22 +00:00
import qualified Utility.Scheduled.QuickCheck
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
import qualified Utility.HumanTime
metadata: Fix encoding problem that led to mojibake when storing metadata strings that contained both unicode characters and a space (or '!') character. The fix is to stop using w82s, which does not properly reconstitute unicode strings. Instrad, use utf8 bytestring to get the [Word8] to base64. This passes unicode through perfectly, including any invalid filesystem encoded characters. Note that toB64 / fromB64 are also used for creds and cipher embedding. It would be unfortunate if this change broke those uses. For cipher embedding, note that ciphers can contain arbitrary bytes (should really be using ByteString.Char8 there). Testing indicated it's not safe to use the new fromB64 there; I think that characters were incorrectly combined. For credpair embedding, the username or password could contain unicode. Before, that unicode would fail to round-trip through the b64. So, I guess this is not going to break any embedded creds that worked before. This bug may have affected some creds before, and if so, this change will not fix old ones, but should fix new ones at least.
2015-03-04 15:16:03 +00:00
import qualified Utility.Base64
finally really add back custom-setup stanza Fourth or fifth try at this and finally found a way to make it work. Absurd amount of busy-work forced on me by change in cabal's behavior. Split up Utility modules that need posix stuff out of ones used by Setup. Various other hacks around inability for Setup to use anything that ifdefs a use of unix. Probably lost a full day of my life to this. This is how build systems make their users hate them. Just saying.
2017-12-31 20:08:31 +00:00
import qualified Utility.Tmp.Dir
Always use filesystem encoding for all file and handle reads and writes. This is a big scary change. I have convinced myself it should be safe. I hope!
2016-12-24 18:46:31 +00:00
import qualified Utility.FileSystemEncoding
Fix mangling of --json output of utf-8 characters when not running in a utf-8 locale As long as all code imports Utility.Aeson rather than Data.Aeson, and no Strings that may contain utf-8 characters are used for eg, object keys via T.pack, this is guaranteed to fix the problem everywhere that git-annex generates json. It's kind of annoying to need to wrap ToJSON with a ToJSON', especially since every data type that has a ToJSON instance has to be ported over. However, that only took 50 lines of code, which is worth it to ensure full coverage. I initially tried an alternative approach of a newtype FileEncoded, which had to be used everywhere a String was fed into aeson, and chasing down all the sites would have been far too hard. Did consider creating an intentionally overlapping instance ToJSON String, and letting ghc fail to build anything that passed in a String, but am not sure that wouldn't pollute some library that git-annex depends on that happens to use ToJSON String internally. This commit was supported by the NSF-funded DataLad project.
2018-04-16 19:42:45 +00:00
import qualified Utility.Aeson
fix windows build
2014-10-16 19:54:02 +00:00
#ifndef mingw32_HOST_OS
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
import qualified Remote.Helper.Encryptable
import qualified Types.Crypto
import qualified Utility.Gpg
squelch build warning on windows
2016-09-06 18:59:32 +00:00
import qualified Types.Remote
ssh warnings
2013-05-17 20:51:46 +00:00
#endif
add a stupid test harness
2010-11-02 20:49:35 +00:00
test: Added --keep-failures option.
2016-01-06 17:44:12 +00:00
optParser :: Parser TestOptions
optParser = TestOptions
fix numerous problem with test suite on crippled filesystems etc
2016-02-16 19:30:59 +00:00
<$> suiteOptionParser ingredients (tests False mempty)
test: Added --keep-failures option.
2016-01-06 17:44:12 +00:00
<*> switch
( long "keep-failures"
<> help "preserve repositories on test failure"
test suite infra for testing mocked ssh remotes This commit was supported by the NSF-funded DataLad project.
2017-03-17 23:13:52 +00:00
)
<*> switch
( long "fakessh"
<> internal
)
<*> cmdParams "non-options are for internal use only"
wire tasty's option parser into the main program option parser This makes bash completion work for git-annex test, and is generally cleaner.
2015-07-13 17:19:20 +00:00
remove leftovers from removed TestSuite build flag Test suite is always built, so this can be simplified.
2018-11-19 16:39:16 +00:00
runner :: TestOptions -> IO ()
runner opts
| fakeSsh opts = runFakeSsh (internalData opts)
| otherwise = runsubprocesstests =<< Utility.Env.getEnv subenv
test suite infra for testing mocked ssh remotes This commit was supported by the NSF-funded DataLad project.
2017-03-17 23:13:52 +00:00
where
test: Avoid most situations involving failure to delete test directories By forking a worker process and only deleting the test directory once it exits. This way, if a test leaves files open, they'll get closed when the worker exits, so avoiding failure to delete open files on Windows, and failure to delete directories due to NFS lock files. If a test leaves a git worker process running, the closed pipes should cause the worker to exit too, also avoiding the problem there. The 10 second sleep ought to give plenty of time for such worker processes to exit, although this is of course a race. Finally, even if test directory fails to be deleted still, it won't appear as if the last test in the test suite failed; the error will be displayed at the very end. This commit was supported by the NSF-funded DataLad project.
2017-08-14 20:27:35 +00:00
-- Run git-annex test in a subprocess, so that any files
-- it may open will be closed before running finalCleanup.
-- This should prevent most failures to clean up after the test
-- suite.
subenv = "GIT_ANNEX_TEST_SUBPROCESS"
remove leftovers from removed TestSuite build flag Test suite is always built, so this can be simplified.
2018-11-19 16:39:16 +00:00
runsubprocesstests Nothing = do
test: Avoid most situations involving failure to delete test directories By forking a worker process and only deleting the test directory once it exits. This way, if a test leaves files open, they'll get closed when the worker exits, so avoiding failure to delete open files on Windows, and failure to delete directories due to NFS lock files. If a test leaves a git worker process running, the closed pipes should cause the worker to exit too, also avoiding the problem there. The 10 second sleep ought to give plenty of time for such worker processes to exit, although this is of course a race. Finally, even if test directory fails to be deleted still, it won't appear as if the last test in the test suite failed; the error will be displayed at the very end. This commit was supported by the NSF-funded DataLad project.
2017-08-14 20:27:35 +00:00
pp <- Annex.Path.programPath
finally really add back custom-setup stanza Fourth or fifth try at this and finally found a way to make it work. Absurd amount of busy-work forced on me by change in cabal's behavior. Split up Utility modules that need posix stuff out of ones used by Setup. Various other hacks around inability for Setup to use anything that ifdefs a use of unix. Probably lost a full day of my life to this. This is how build systems make their users hate them. Just saying.
2017-12-31 20:08:31 +00:00
Utility.Env.Set.setEnv subenv "1" True
test: Avoid most situations involving failure to delete test directories By forking a worker process and only deleting the test directory once it exits. This way, if a test leaves files open, they'll get closed when the worker exits, so avoiding failure to delete open files on Windows, and failure to delete directories due to NFS lock files. If a test leaves a git worker process running, the closed pipes should cause the worker to exit too, also avoiding the problem there. The 10 second sleep ought to give plenty of time for such worker processes to exit, although this is of course a race. Finally, even if test directory fails to be deleted still, it won't appear as if the last test in the test suite failed; the error will be displayed at the very end. This commit was supported by the NSF-funded DataLad project.
2017-08-14 20:27:35 +00:00
ps <- getArgs
remove leftovers from removed TestSuite build flag Test suite is always built, so this can be simplified.
2018-11-19 16:39:16 +00:00
(Nothing, Nothing, Nothing, pid) <- createProcess (proc pp ps)
test: Avoid most situations involving failure to delete test directories By forking a worker process and only deleting the test directory once it exits. This way, if a test leaves files open, they'll get closed when the worker exits, so avoiding failure to delete open files on Windows, and failure to delete directories due to NFS lock files. If a test leaves a git worker process running, the closed pipes should cause the worker to exit too, also avoiding the problem there. The 10 second sleep ought to give plenty of time for such worker processes to exit, although this is of course a race. Finally, even if test directory fails to be deleted still, it won't appear as if the last test in the test suite failed; the error will be displayed at the very end. This commit was supported by the NSF-funded DataLad project.
2017-08-14 20:27:35 +00:00
exitcode <- waitForProcess pid
unless (keepFailuresOption opts) finalCleanup
exitWith exitcode
remove leftovers from removed TestSuite build flag Test suite is always built, so this can be simplified.
2018-11-19 16:39:16 +00:00
runsubprocesstests (Just _) = isolateGitConfig $ do
test suite infra for testing mocked ssh remotes This commit was supported by the NSF-funded DataLad project.
2017-03-17 23:13:52 +00:00
ensuretmpdir
test: Fix reversion that made it only run inside a git repository. Using annexeval to run probeCrippledFileSystem' caused Git.CurrentRepo.get to be run. Fixed easily since probeCrippledFileSystem' had no need to use the Annex monad. This commit was sponsored by Ethan Aubin.
2017-09-29 18:58:23 +00:00
crippledfilesystem <- fst <$> Annex.Init.probeCrippledFileSystem' tmpdir
test suite infra for testing mocked ssh remotes This commit was supported by the NSF-funded DataLad project.
2017-03-17 23:13:52 +00:00
case tryIngredients ingredients (tastyOptionSet opts) (tests crippledfilesystem opts) of
Nothing -> error "No tests found!?"
Just act -> ifM act
( exitSuccess
, do
test: Avoid most situations involving failure to delete test directories By forking a worker process and only deleting the test directory once it exits. This way, if a test leaves files open, they'll get closed when the worker exits, so avoiding failure to delete open files on Windows, and failure to delete directories due to NFS lock files. If a test leaves a git worker process running, the closed pipes should cause the worker to exit too, also avoiding the problem there. The 10 second sleep ought to give plenty of time for such worker processes to exit, although this is of course a race. Finally, even if test directory fails to be deleted still, it won't appear as if the last test in the test suite failed; the error will be displayed at the very end. This commit was supported by the NSF-funded DataLad project.
2017-08-14 20:27:35 +00:00
putStrLn " (Failures above could be due to a bug in git-annex, or an incompatibility"
test suite infra for testing mocked ssh remotes This commit was supported by the NSF-funded DataLad project.
2017-03-17 23:13:52 +00:00
putStrLn " with utilities, such as git, installed on this system.)"
exitFailure
)
expose tasty test suite's option parser
2014-01-21 04:08:43 +00:00
ingredients :: [Ingredient]
ingredients =
test: Fix --list-tests Seems list order matters, I don't know why, but have reported it on #tasty
2015-03-31 18:47:27 +00:00
[ listingTests
, rerunningTests [consoleTestReporter]
expose tasty test suite's option parser
2014-01-21 04:08:43 +00:00
]
test suite partially converted to use tasty test framework This is a win. Will need to wait for tasty getting into Debian, and do a trivial conversion of the remainder of the hunit tests.
2013-08-06 04:12:06 +00:00
fix numerous problem with test suite on crippled filesystems etc
2016-02-16 19:30:59 +00:00
tests :: Bool -> TestOptions -> TestTree
tests crippledfilesystem opts = testGroup "Tests" $ properties :
Split Test.hs and avoid optimising it much, to need less memory to compile. The ghc options were found by Sean Whitton; the debian arm autobuilders need those to build w/o OOM, and it seems to involve llvm using too much memory to optimize Test. This commit was sponsored by Boyd Stephen Smith Jr. on Patreon.
2018-02-18 15:48:48 +00:00
map (\(d, te) -> withTestMode te initTests (unitTests d)) testmodes
starting to work on test suite for v6
2015-12-15 21:19:26 +00:00
where
fix numerous problem with test suite on crippled filesystems etc
2016-02-16 19:30:59 +00:00
testmodes = catMaybes
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
[ Just ("v7 unlocked", (testMode opts (RepoVersion 7)) { unlockedFiles = True })
, unlesscrippled ("v5", testMode opts (RepoVersion 5))
, unlesscrippled ("v7 locked", testMode opts (RepoVersion 7))
, Just ("v5 direct", (testMode opts (RepoVersion 5)) { forceDirect = True })
fix build on windows
2015-12-28 16:57:42 +00:00
]
100% pass on FAT
2016-02-16 20:02:54 +00:00
unlesscrippled v
| crippledfilesystem = Nothing
| otherwise = Just v
refactor
2015-07-08 04:03:28 +00:00
test suite partially converted to use tasty test framework This is a win. Will need to wait for tasty getting into Debian, and do a trivial conversion of the remainder of the hunit tests.
2013-08-06 04:12:06 +00:00
properties :: TestTree
use tasty resources
2014-01-21 18:04:50 +00:00
properties = localOption (QuickCheckTests 1000) $ testGroup "QuickCheck"
fix failing quickcheck properties QuickCheck 2.10 found a counterexample eg "\929184" broke the property. As far as I can tell, Git.Filename is matching how git handles encoding of strange high unicode characters in filenames for display. Git does not display high unicode characters, and instead displays the C-style escaped form of each byte. This is ambiguous, but since git is not unicode aware, it doesn't need to roundtrip parse it. So, making Git.FileName's roundtrip test only chars < 256 seems fine. Utility.Format.format uses encode_c, in order to mimic git, so that's ok. Utility.Format.gen uses decode_c, but only so that stuff like "\n" in the format string is handled. If the format string contains C-style octal escapes, they will be converted to ascii characters, and not combined into unicode characters, but that should not be a problem. If the user wants unicode characters, they can include them in the format string, without escaping them. Finally, decode_c is used by Utility.Gpg.secretKeys, because gpg --with-colons hex-escapes some characters in particular ':' and '\\'. gpg passes unicode through, so this use of decode_c is not a problem. This commit was sponsored by Henrik Riomar on Patreon.
2017-06-17 20:17:09 +00:00
[ testProperty "prop_encode_decode_roundtrip" Git.Filename.prop_encode_decode_roundtrip
, testProperty "prop_encode_c_decode_c_roundtrip" Utility.Format.prop_encode_c_decode_c_roundtrip
remove 163 lines of code without changing anything except imports
2016-01-20 20:36:33 +00:00
, testProperty "prop_isomorphic_fileKey" Annex.Locations.prop_isomorphic_fileKey
factor non-type stuff out of Key
2017-02-24 17:42:30 +00:00
, testProperty "prop_isomorphic_key_encode" Key.prop_isomorphic_key_encode
, testProperty "prop_isomorphic_key_decode" Key.prop_isomorphic_key_decode
fix use of hifalutin terminology
2015-11-16 18:37:31 +00:00
, testProperty "prop_isomorphic_shellEscape" Utility.SafeCommand.prop_isomorphic_shellEscape
, testProperty "prop_isomorphic_shellEscape_multiword" Utility.SafeCommand.prop_isomorphic_shellEscape_multiword
, testProperty "prop_isomorphic_configEscape" Logs.Remote.prop_isomorphic_configEscape
test suite partially converted to use tasty test framework This is a win. Will need to wait for tasty getting into Debian, and do a trivial conversion of the remainder of the hunit tests.
2013-08-06 04:12:06 +00:00
, testProperty "prop_parse_show_Config" Logs.Remote.prop_parse_show_Config
convert parentDir to be based on takeDirectory, but fixed for trailing /
2015-01-09 18:26:52 +00:00
, testProperty "prop_upFrom_basics" Utility.Path.prop_upFrom_basics
test suite partially converted to use tasty test framework This is a win. Will need to wait for tasty getting into Debian, and do a trivial conversion of the remainder of the hunit tests.
2013-08-06 04:12:06 +00:00
, testProperty "prop_relPathDirToFile_basics" Utility.Path.prop_relPathDirToFile_basics
, testProperty "prop_relPathDirToFile_regressionTest" Utility.Path.prop_relPathDirToFile_regressionTest
, testProperty "prop_cost_sane" Config.Cost.prop_cost_sane
, testProperty "prop_matcher_sane" Utility.Matcher.prop_matcher_sane
, testProperty "prop_HmacSha1WithCipher_sane" Crypto.prop_HmacSha1WithCipher_sane
Added GIT_ANNEX_VECTOR_CLOCK environment variable Can be used to override the default timestamps used in log files in the git-annex branch. This is a dangerous environment variable; use with caution. Note that this only affects writing to the logs on the git-annex branch. It is not used for metadata in git commits (other env vars can be set for that). There are many other places where timestamps are still used, that don't get committed to git, but do touch disk. Including regular timestamps of files, and timestamps embedded in some files in .git/annex/, including the last fsck timestamp and timestamps in transfer log files. A good way to find such things in git-annex is to get for getPOSIXTime and getCurrentTime, although some of the results are of course false positives that never hit disk (unless git-annex gets swapped out..) So this commit does NOT necessarily make git-annex comply with some HIPPA privacy regulations; it's up to the user to determine if they can use it in a way compliant with such regulations. Benchmarking: It takes 0.00114 milliseconds to call getEnv "GIT_ANNEX_VECTOR_CLOCK" when that env var is not set. So, 100 thousand log files can be written with an added overhead of only 0.114 seconds. That should be by far swamped by the actual overhead of writing the log files and making the commit containing them. This commit was supported by the NSF-funded DataLad project.
2017-08-14 17:55:38 +00:00
, testProperty "prop_VectorClock_sane" Annex.VectorClock.prop_VectorClock_sane
factored out a generic MapLog from uuid-based logs UUIDBased is just a MapLog with a UUID for the field.
2014-03-15 17:44:31 +00:00
, testProperty "prop_addMapLog_sane" Logs.MapLog.prop_addMapLog_sane
test suite partially converted to use tasty test framework This is a win. Will need to wait for tasty getting into Debian, and do a trivial conversion of the remainder of the hunit tests.
2013-08-06 04:12:06 +00:00
, testProperty "prop_verifiable_sane" Utility.Verifiable.prop_verifiable_sane
, testProperty "prop_segment_regressionTest" Utility.Misc.prop_segment_regressionTest
, testProperty "prop_read_write_transferinfo" Logs.Transfer.prop_read_write_transferinfo
, testProperty "prop_read_show_inodecache" Utility.InodeCache.prop_read_show_inodecache
convert git-annex branch access to ByteStrings and Builders Most of the individual logs are not converted yet, only presense logs have an efficient ByteString Builder implemented so far. The rest convert to and from String.
2019-01-03 17:21:48 +00:00
, testProperty "prop_parse_build_log" Logs.Presence.prop_parse_build_log
test suite partially converted to use tasty test framework This is a win. Will need to wait for tasty getting into Debian, and do a trivial conversion of the remainder of the hunit tests.
2013-08-06 04:12:06 +00:00
, testProperty "prop_read_show_TrustLevel" Types.TrustLevel.prop_read_show_TrustLevel
, testProperty "prop_parse_show_TrustLog" Logs.Trust.prop_parse_show_TrustLog
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
, testProperty "prop_hashes_stable" Utility.Hash.prop_hashes_stable
add test for stable macs
2015-04-19 14:52:49 +00:00
, testProperty "prop_mac_stable" Utility.Hash.prop_mac_stable
reorg quickcheck to a separate module
2015-11-17 19:49:22 +00:00
, testProperty "prop_schedule_roundtrips" Utility.Scheduled.QuickCheck.prop_schedule_roundtrips
add new test
2014-04-12 18:19:56 +00:00
, testProperty "prop_past_sane" Utility.Scheduled.prop_past_sane
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
, testProperty "prop_duration_roundtrips" Utility.HumanTime.prop_duration_roundtrips
fix metadata log simplifier and additional quickcheck tests
2014-02-13 02:27:55 +00:00
, testProperty "prop_metadata_sane" Types.MetaData.prop_metadata_sane
data types and serialization for metadata A very haskell commit! Just data types, instances to serialize the metadata to a nice format, and QuickCheck tests. This commit was sponsored by Andreas Leha.
2014-02-12 21:54:28 +00:00
, testProperty "prop_metadata_serialize" Types.MetaData.prop_metadata_serialize
add git annex view command (And a vpop command, which is still a bit buggy.) Still need to do vadd and vrm, though this also adds their documentation. Currently not very happy with the view log data serialization. I had to lose the TDFA regexps temporarily, so I can have Read/Show instances of View. I expect the view log format will change in some incompatable way later, probably adding last known refs for the parent branch to View or something like that. Anyway, it basically works, although it's a bit slow looking up the metadata. The actual git branch construction is about as fast as it can be using the current git plumbing. This commit was sponsored by Peter Hogg.
2014-02-18 21:38:23 +00:00
, testProperty "prop_branchView_legal" Logs.View.prop_branchView_legal
view: Avoid using cute unicode homomorphs for '/' and '\' and instead use ugly escaping, as the unicode method doesn't work on non-unicode supporting systems.
2016-01-08 16:45:32 +00:00
, testProperty "prop_viewPath_roundtrips" Annex.View.prop_viewPath_roundtrips
add another quickcheck property, and several edge cases handled
2014-02-17 01:00:12 +00:00
, testProperty "prop_view_roundtrips" Annex.View.prop_view_roundtrips
change directory encoding in ViewedFile such that the original directory can be extracted from it
2014-02-22 18:54:53 +00:00
, testProperty "prop_viewedFile_rountrips" Annex.View.ViewedFile.prop_viewedFile_roundtrips
metadata: Fix encoding problem that led to mojibake when storing metadata strings that contained both unicode characters and a space (or '!') character. The fix is to stop using w82s, which does not properly reconstitute unicode strings. Instrad, use utf8 bytestring to get the [Word8] to base64. This passes unicode through perfectly, including any invalid filesystem encoded characters. Note that toB64 / fromB64 are also used for creds and cipher embedding. It would be unfortunate if this change broke those uses. For cipher embedding, note that ciphers can contain arbitrary bytes (should really be using ByteString.Char8 there). Testing indicated it's not safe to use the new fromB64 there; I think that characters were incorrectly combined. For credpair embedding, the username or password could contain unicode. Before, that unicode would fail to round-trip through the b64. So, I guess this is not going to break any embedded creds that worked before. This bug may have affected some creds before, and if so, this change will not fix old ones, but should fix new ones at least.
2015-03-04 15:16:03 +00:00
, testProperty "prop_b64_roundtrips" Utility.Base64.prop_b64_roundtrips
add test case that all standard group preferred content expressions parse
2015-06-17 17:44:19 +00:00
, testProperty "prop_standardGroups_parse" Logs.PreferredContent.prop_standardGroups_parse
better test framework
2010-11-02 21:08:16 +00:00
]
tests should now all run independently, without needing any prior tests to have run Should be ready now for tasty-rerun.
2014-01-21 18:24:30 +00:00
{- These tests set up the test environment, but also test some basic parts
- of git-annex. They are always run before the unitTests. -}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
initTests :: TestTree
initTests = testGroup "Init Tests"
[ testCase "init" test_init
, testCase "add" test_add
tests should now all run independently, without needing any prior tests to have run Should be ready now for tasty-rerun.
2014-01-21 18:24:30 +00:00
]
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
unitTests :: String -> TestTree
unitTests note = testGroup ("Unit Tests " ++ note)
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
[ testCase "add dup" test_add_dup
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
, testCase "add extras" test_add_extras
init: Fix reversion in detection of repo made with git clone --shared
2015-09-09 17:56:37 +00:00
, testCase "shared clone" test_shared_clone
log: Fix reversion introduced in version 5.20150528 that broke this command.
2015-07-02 16:53:38 +00:00
, testCase "log" test_log
fairly complete tests for git-annex import
2015-06-03 18:39:47 +00:00
, testCase "import" test_import
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
, testCase "reinject" test_reinject
, testCase "unannex (no copy)" test_unannex_nocopy
, testCase "unannex (with copy)" test_unannex_withcopy
, testCase "drop (no remote)" test_drop_noremote
, testCase "drop (with remote)" test_drop_withremote
, testCase "drop (untrusted remote)" test_drop_untrustedremote
, testCase "get" test_get
test move with ssh remote
2017-03-17 23:18:45 +00:00
, testCase "get (ssh remote)" test_get_ssh_remote
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
, testCase "move" test_move
test move with ssh remote
2017-03-17 23:18:45 +00:00
, testCase "move (ssh remote)" test_move_ssh_remote
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
, testCase "copy" test_copy
, testCase "lock" test_lock
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
, testCase "lock (v7 --force)" test_lock_v7_force
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
, testCase "edit (no pre-commit)" test_edit
, testCase "edit (pre-commit)" test_edit_precommit
, testCase "partial commit" test_partial_commit
, testCase "fix" test_fix
add a test case for switching to direct mode and back to indirect
2015-02-09 17:41:29 +00:00
, testCase "direct" test_direct
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
, testCase "trust" test_trust
, testCase "fsck (basics)" test_fsck_basic
, testCase "fsck (bare)" test_fsck_bare
, testCase "fsck (local untrusted)" test_fsck_localuntrusted
, testCase "fsck (remote untrusted)" test_fsck_remoteuntrusted
Fix fsck --from a git remote in a local directory, and from a directory special remote. This was a reversion caused by the relative path changes in 5.20150113. The directory special remote was not affected in its normal configuration, since annex-directory is an absolute path normally. But it could fail when a relative path was used. The git remote was affected even when an absolute path to it was used in .git/config, since git-annex now converts all such paths to relative.
2015-04-18 17:36:12 +00:00
, testCase "fsck --from remote" test_fsck_fromremote
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
, testCase "migrate" test_migrate
, testCase "migrate (via gitattributes)" test_migrate_via_gitattributes
, testCase "unused" test_unused
, testCase "describe" test_describe
, testCase "find" test_find
, testCase "merge" test_merge
, testCase "info" test_info
, testCase "version" test_version
, testCase "sync" test_sync
, testCase "union merge regression" test_union_merge_regression
add test case for http://git-annex.branchable.com/bugs/Assistant_keeps_deleting_all_the_files_in_my_repo/
2016-06-13 16:56:31 +00:00
, testCase "adjusted branch merge regression" test_adjusted_branch_merge_regression
add test case for sync_in_adjusted_branch_deleted_recently_added_files This commit was sponsored by Denis Dzyubenko on Patreon.
2016-10-11 18:22:49 +00:00
, testCase "adjusted branch subtree regression" test_adjusted_branch_subtree_regression
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
, testCase "conflict resolution" test_conflict_resolution
add simple test for conflict resolution in adjusted branch This is not really extensive enough, but a start..
2016-04-12 17:01:31 +00:00
, testCase "conflict resolution (adjusted branch)" test_conflict_resolution_adjusted_branch
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
, testCase "conflict resolution movein regression" test_conflict_resolution_movein_regression
, testCase "conflict resolution (mixed directory and file)" test_mixed_conflict_resolution
, testCase "conflict resolution symlink bit" test_conflict_resolution_symlink_bit
, testCase "conflict resolution (uncommitted local file)" test_uncommitted_conflict_resolution
, testCase "conflict resolution (removed file)" test_remove_conflict_resolution
, testCase "conflict resolution (nonannexed file)" test_nonannexed_file_conflict_resolution
, testCase "conflict resolution (nonannexed symlink)" test_nonannexed_symlink_conflict_resolution
add test: conflict resolution (mixed locked and unlocked file)
2015-12-30 20:36:39 +00:00
, testCase "conflict resolution (mixed locked and unlocked file)" test_mixed_lock_conflict_resolution
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
, testCase "map" test_map
, testCase "uninit" test_uninit
, testCase "uninit (in git-annex branch)" test_uninit_inbranch
, testCase "upgrade" test_upgrade
, testCase "whereis" test_whereis
, testCase "hook remote" test_hook_remote
, testCase "directory remote" test_directory_remote
, testCase "rsync remote" test_rsync_remote
, testCase "bup remote" test_bup_remote
, testCase "crypto" test_crypto
, testCase "preferred content" test_preferred_content
, testCase "add subdirs" test_add_subdirs
test case for addurl file://
2015-02-10 19:20:07 +00:00
, testCase "addurl" test_addurl
added some toplevel git-annex subcommand tests note that test coverage doesn't work for those yet
2011-01-07 00:09:49 +00:00
]
tests should now all run independently, without needing any prior tests to have run Should be ready now for tasty-rerun.
2014-01-21 18:24:30 +00:00
-- this test case create the main repo
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_init :: Assertion
test_init = innewrepo $ do
pass --version to init when needed
2015-12-16 17:14:18 +00:00
ver <- annexVersion <$> getTestMode
honor init --version=5 on crippled filesystem init: When --version=5 is passed on a crippled filesystem, use a v5 direct mode repo as requested, rather than upgrading to v7 adjusted unlocked. Fixed test suite on crippled filesystems, making it request --version=5 to test direct mode.
2018-12-19 17:07:19 +00:00
git_annex "init" [reponame, "--version", show (fromRepoVersion ver)] @? "init failed"
starting to work on test suite for v6
2015-12-15 21:19:26 +00:00
setupTestMode
where indenting
2012-11-11 04:51:07 +00:00
where
reponame = "test repo"
added some toplevel git-annex subcommand tests note that test coverage doesn't work for those yet
2011-01-07 00:09:49 +00:00
test suite partially converted to use tasty test framework This is a win. Will need to wait for tasty getting into Debian, and do a trivial conversion of the remainder of the hunit tests.
2013-08-06 04:12:06 +00:00
-- this test case runs in the main repo, to set up a basic
-- annexed file that later tests will use
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_add :: Assertion
test_add = inmainrepo $ do
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent annexedfile $ content annexedfile
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
add_annex annexedfile @? "add failed"
test suite partially converted to use tasty test framework This is a win. Will need to wait for tasty getting into Debian, and do a trivial conversion of the remainder of the hunit tests.
2013-08-06 04:12:06 +00:00
annexed_present annexedfile
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent sha1annexedfile $ content sha1annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "add" [sha1annexedfile, "--backend=SHA1"] @? "add with SHA1 failed"
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
whenM (unlockedFiles <$> getTestMode) $
git_annex "unlock" [sha1annexedfile] @? "unlock failed"
test suite partially converted to use tasty test framework This is a win. Will need to wait for tasty getting into Debian, and do a trivial conversion of the remainder of the hunit tests.
2013-08-06 04:12:06 +00:00
annexed_present sha1annexedfile
checkbackend sha1annexedfile backendSHA1
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
ifM (annexeval Config.isDirect)
( do
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent ingitfile $ content ingitfile
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
not <$> boolSystem "git" [Param "add", File ingitfile] @? "git add failed to fail in direct mode"
use nukefile rather than shelling to rm
2014-02-10 22:29:15 +00:00
nukeFile ingitfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
, do
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent ingitfile $ content ingitfile
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
boolSystem "git" [Param "add", File ingitfile] @? "git add failed"
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "commit", Param "-q", Param "-m", Param "commit"] @? "git commit failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "add" [ingitfile] @? "add ingitfile should be no-op"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unannexed ingitfile
)
test suite partially converted to use tasty test framework This is a win. Will need to wait for tasty getting into Debian, and do a trivial conversion of the remainder of the hunit tests.
2013-08-06 04:12:06 +00:00
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
test_add_dup :: Assertion
test_add_dup = intmpclonerepo $ do
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent annexedfiledup $ content annexedfiledup
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
add_annex annexedfiledup @? "add of second file with same content failed"
annexed_present annexedfiledup
annexed_present annexedfile
test suite partially converted to use tasty test framework This is a win. Will need to wait for tasty getting into Debian, and do a trivial conversion of the remainder of the hunit tests.
2013-08-06 04:12:06 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_add_extras :: Assertion
test_add_extras = intmpclonerepo $ do
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent wormannexedfile $ content wormannexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "add" [wormannexedfile, "--backend=WORM"] @? "add with WORM failed"
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
whenM (unlockedFiles <$> getTestMode) $
git_annex "unlock" [wormannexedfile] @? "unlock failed"
tests should now all run independently, without needing any prior tests to have run Should be ready now for tasty-rerun.
2014-01-21 18:24:30 +00:00
annexed_present wormannexedfile
checkbackend wormannexedfile backendWORM
init: Fix reversion in detection of repo made with git clone --shared
2015-09-09 17:56:37 +00:00
test_shared_clone :: Assertion
test_shared_clone = intmpsharedclonerepo $ do
v <- catchMaybeIO $ Utility.Process.readProcess "git"
[ "config"
, "--bool"
, "--get"
, "annex.hardlink"
]
v == Just "true\n"
@? "shared clone of repo did not get annex.hardlink set"
log: Fix reversion introduced in version 5.20150528 that broke this command.
2015-07-02 16:53:38 +00:00
test_log :: Assertion
test_log = intmpclonerepo $ do
git_annex "log" [annexedfile] @? "log failed"
fairly complete tests for git-annex import
2015-06-03 18:39:47 +00:00
test_import :: Assertion
finally really add back custom-setup stanza Fourth or fifth try at this and finally found a way to make it work. Absurd amount of busy-work forced on me by change in cabal's behavior. Split up Utility modules that need posix stuff out of ones used by Setup. Various other hacks around inability for Setup to use anything that ifdefs a use of unix. Probably lost a full day of my life to this. This is how build systems make their users hate them. Just saying.
2017-12-31 20:08:31 +00:00
test_import = intmpclonerepo $ Utility.Tmp.Dir.withTmpDir "importtest" $ \importdir -> do
fairly complete tests for git-annex import
2015-06-03 18:39:47 +00:00
(toimport1, importf1, imported1) <- mktoimport importdir "import1"
git_annex "import" [toimport1] @? "import failed"
fix numerous problem with test suite on crippled filesystems etc
2016-02-16 19:30:59 +00:00
annexed_present_imported imported1
fairly complete tests for git-annex import
2015-06-03 18:39:47 +00:00
checkdoesnotexist importf1
(toimport2, importf2, imported2) <- mktoimport importdir "import2"
git_annex "import" [toimport2] @? "import of duplicate failed"
fix numerous problem with test suite on crippled filesystems etc
2016-02-16 19:30:59 +00:00
annexed_present_imported imported2
fairly complete tests for git-annex import
2015-06-03 18:39:47 +00:00
checkdoesnotexist importf2
(toimport3, importf3, imported3) <- mktoimport importdir "import3"
git_annex "import" ["--skip-duplicates", toimport3]
@? "import of duplicate with --skip-duplicates failed"
checkdoesnotexist imported3
checkexists importf3
git_annex "import" ["--clean-duplicates", toimport3]
@? "import of duplicate with --clean-duplicates failed"
checkdoesnotexist imported3
checkdoesnotexist importf3
(toimport4, importf4, imported4) <- mktoimport importdir "import4"
git_annex "import" ["--deduplicate", toimport4] @? "import --deduplicate failed"
checkdoesnotexist imported4
checkdoesnotexist importf4
(toimport5, importf5, imported5) <- mktoimport importdir "import5"
git_annex "import" ["--duplicate", toimport5] @? "import --duplicate failed"
fix numerous problem with test suite on crippled filesystems etc
2016-02-16 19:30:59 +00:00
annexed_present_imported imported5
fairly complete tests for git-annex import
2015-06-03 18:39:47 +00:00
checkexists importf5
git_annex "drop" ["--force", imported1, imported2, imported5] @? "drop failed"
fix numerous problem with test suite on crippled filesystems etc
2016-02-16 19:30:59 +00:00
annexed_notpresent_imported imported2
fairly complete tests for git-annex import
2015-06-03 18:39:47 +00:00
(toimportdup, importfdup, importeddup) <- mktoimport importdir "importdup"
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "import" ["--clean-duplicates", toimportdup]
import: Changed how --deduplicate, --skip-duplicates, and --clean-duplicates determine if a file is a duplicate Before, only content known to be present somewhere was considered a duplicate. Now, any content that has been annexed before will be considered a duplicate, even if all annexed copies of the data have been lost. Note that --clean-duplicates and --deduplicate still check numcopies, so won't delete duplicate files unless there's an annexed copy. This makes import use the same method as reinject --known. The man page already said that duplicate meant "its content is either present in the local repository already, or git-annex knows of another repository that contains it, or it was present in the annex before but has been removed now". So, this is really only bringing the implementation into line with the man page. This commit was sponsored by Jochen Bartl on Patreon.
2017-02-07 21:35:51 +00:00
@? "import of missing duplicate with --clean-duplicates failed to fail"
fairly complete tests for git-annex import
2015-06-03 18:39:47 +00:00
checkdoesnotexist importeddup
checkexists importfdup
where
mktoimport importdir subdir = do
createDirectory (importdir </> subdir)
let importf = subdir </> "f"
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent (importdir </> importf) (content importf)
fairly complete tests for git-annex import
2015-06-03 18:39:47 +00:00
return (importdir </> subdir, importdir </> importf, importf)
fix numerous problem with test suite on crippled filesystems etc
2016-02-16 19:30:59 +00:00
annexed_present_imported f = ifM (annexeval Config.crippledFileSystem)
( annexed_present_unlocked f
, annexed_present_locked f
)
annexed_notpresent_imported f = ifM (annexeval Config.crippledFileSystem)
( annexed_notpresent_unlocked f
, annexed_notpresent_locked f
)
fairly complete tests for git-annex import
2015-06-03 18:39:47 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_reinject :: Assertion
test_reinject = intmpclonerepoInDirect $ do
git_annex "drop" ["--force", sha1annexedfile] @? "drop failed"
fix some tests of v6 unlocked files
2016-01-05 21:19:30 +00:00
annexed_notpresent sha1annexedfile
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent tmp $ content sha1annexedfile
factor non-type stuff out of Key
2017-02-24 17:42:30 +00:00
key <- Key.key2file <$> getKey backendSHA1 tmp
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "reinject" [tmp, sha1annexedfile] @? "reinject failed"
fix some tests of v6 unlocked files
2016-01-05 21:19:30 +00:00
annexed_present sha1annexedfile
avoid fromkey test when on crippled fs
2016-02-15 20:54:44 +00:00
-- fromkey can't be used on a crippled filesystem, since it makes a
-- symlink
unlessM (annexeval Config.crippledFileSystem) $ do
git_annex "fromkey" [key, sha1annexedfiledup] @? "fromkey failed for dup"
annexed_present_locked sha1annexedfiledup
where indenting
2012-11-11 04:51:07 +00:00
where
tmp = "tmpfile"
add tests of setkey/fromkey
2011-01-11 23:59:11 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_unannex_nocopy :: Assertion
test_unannex_nocopy = intmpclonerepo $ do
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "unannex" [annexedfile] @? "unannex failed with no copy"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_unannex_withcopy :: Assertion
test_unannex_withcopy = intmpclonerepo $ do
git_annex "get" [annexedfile] @? "get failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "unannex" [annexedfile, sha1annexedfile] @? "unannex failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unannexed annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "unannex" [annexedfile] @? "unannex failed on non-annexed file"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unannexed annexedfile
unlessM (annexeval Config.isDirect) $ do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "unannex" [ingitfile] @? "unannex ingitfile should be no-op"
where indenting
2012-11-11 04:51:07 +00:00
unannexed ingitfile
more tests
2011-01-07 01:39:26 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_drop_noremote :: Assertion
test_drop_noremote = intmpclonerepo $ do
git_annex "get" [annexedfile] @? "get failed"
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "remote", Param "rm", Param "origin"]
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
@? "git remote rm origin failed"
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "drop" [annexedfile] @? "drop wrongly succeeded with no known copy of file"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "drop" ["--force", annexedfile] @? "drop --force failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "drop" [annexedfile] @? "drop of dropped file failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unlessM (annexeval Config.isDirect) $ do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "drop" [ingitfile] @? "drop ingitfile should be no-op"
where indenting
2012-11-11 04:51:07 +00:00
unannexed ingitfile
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_drop_withremote :: Assertion
test_drop_withremote = intmpclonerepo $ do
git_annex "get" [annexedfile] @? "get failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "numcopies" ["2"] @? "numcopies config failed"
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "drop" [annexedfile] @? "drop succeeded although numcopies is not satisfied"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "numcopies" ["1"] @? "numcopies config failed"
git_annex "drop" [annexedfile] @? "drop failed though origin has copy"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_notpresent annexedfile
Windows: Fix bug in dropping an annexed file, which caused a symlink to be staged that contained backslashes.
2015-02-09 19:24:33 +00:00
-- make sure that the correct symlink is staged for the file
-- after drop
git_annex_expectoutput "status" [] []
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ annexed_present annexedfile
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_drop_untrustedremote :: Assertion
test_drop_untrustedremote = intmpclonerepo $ do
git_annex "untrust" ["origin"] @? "untrust of origin failed"
git_annex "get" [annexedfile] @? "get failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_present annexedfile
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "drop" [annexedfile] @? "drop wrongly succeeded with only an untrusted copy of the file"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ annexed_present annexedfile
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_get :: Assertion
test suite infra for testing mocked ssh remotes This commit was supported by the NSF-funded DataLad project.
2017-03-17 23:13:52 +00:00
test_get = test_get' intmpclonerepo
test_get_ssh_remote :: Assertion
test_get_ssh_remote = test_get' (with_ssh_origin intmpclonerepo)
test_get' :: (Assertion -> Assertion) -> Assertion
test_get' setup = setup $ do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ annexed_present annexedfile
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [annexedfile] @? "get of file failed"
inmainrepo $ annexed_present annexedfile
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [annexedfile] @? "get of file already here failed"
inmainrepo $ annexed_present annexedfile
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_present annexedfile
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unlessM (annexeval Config.isDirect) $ do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ unannexed ingitfile
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unannexed ingitfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [ingitfile] @? "get ingitfile should be no-op"
inmainrepo $ unannexed ingitfile
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unannexed ingitfile
more tests
2011-01-07 05:02:06 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_move :: Assertion
test move with ssh remote
2017-03-17 23:18:45 +00:00
test_move = test_move' intmpclonerepo
test_move_ssh_remote :: Assertion
test_move_ssh_remote = test_move' (with_ssh_origin intmpclonerepo)
test_move' :: (Assertion -> Assertion) -> Assertion
test_move' setup = setup $ do
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ annexed_present annexedfile
git_annex "move" ["--from", "origin", annexedfile] @? "move --from of file failed"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ annexed_notpresent annexedfile
git_annex "move" ["--from", "origin", annexedfile] @? "move --from of file already here failed"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ annexed_notpresent annexedfile
git_annex "move" ["--to", "origin", annexedfile] @? "move --to of file failed"
inmainrepo $ annexed_present annexedfile
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "move" ["--to", "origin", annexedfile] @? "move --to of file already there failed"
inmainrepo $ annexed_present annexedfile
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_notpresent annexedfile
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unlessM (annexeval Config.isDirect) $ do
unannexed ingitfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ unannexed ingitfile
git_annex "move" ["--to", "origin", ingitfile] @? "move of ingitfile should be no-op"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unannexed ingitfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ unannexed ingitfile
git_annex "move" ["--from", "origin", ingitfile] @? "move of ingitfile should be no-op"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unannexed ingitfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ unannexed ingitfile
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_copy :: Assertion
test_copy = intmpclonerepo $ do
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ annexed_present annexedfile
git_annex "copy" ["--from", "origin", annexedfile] @? "copy --from of file failed"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ annexed_present annexedfile
git_annex "copy" ["--from", "origin", annexedfile] @? "copy --from of file already here failed"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ annexed_present annexedfile
git_annex "copy" ["--to", "origin", annexedfile] @? "copy --to of file already there failed"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ annexed_present annexedfile
git_annex "move" ["--to", "origin", annexedfile] @? "move --to of file already there failed"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ annexed_present annexedfile
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unlessM (annexeval Config.isDirect) $ do
unannexed ingitfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ unannexed ingitfile
git_annex "copy" ["--to", "origin", ingitfile] @? "copy of ingitfile should be no-op"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unannexed ingitfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
inmainrepo $ unannexed ingitfile
git_annex "copy" ["--from", "origin", ingitfile] @? "copy of ingitfile should be no-op"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
checkregularfile ingitfile
checkcontent ingitfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_preferred_content :: Assertion
test_preferred_content = intmpclonerepo $ do
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
annexed_notpresent annexedfile
add test cases for copy --auto (Only --from for now, should add --to sometime but I'm on low sleep so this is the mininal test to catch the reversion just fixed, bad me for doing it backwards and incompletely.)
2015-10-06 21:35:29 +00:00
-- get/copy --auto looks only at numcopies when preferred content is not
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
-- set, and with 1 copy existing, does not get the file.
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" ["--auto", annexedfile] @? "get --auto of file failed with default preferred content"
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
annexed_notpresent annexedfile
add test cases for copy --auto (Only --from for now, should add --to sometime but I'm on low sleep so this is the mininal test to catch the reversion just fixed, bad me for doing it backwards and incompletely.)
2015-10-06 21:35:29 +00:00
git_annex "copy" ["--from", "origin", "--auto", annexedfile] @? "copy --auto --from of file failed with default preferred content"
annexed_notpresent annexedfile
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "wanted" [".", "standard"] @? "set expression to standard failed"
git_annex "group" [".", "client"] @? "set group to standard failed"
git_annex "get" ["--auto", annexedfile] @? "get --auto of file failed for client"
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
annexed_present annexedfile
add test cases for copy --auto (Only --from for now, should add --to sometime but I'm on low sleep so this is the mininal test to catch the reversion just fixed, bad me for doing it backwards and incompletely.)
2015-10-06 21:35:29 +00:00
git_annex "drop" [annexedfile] @? "drop of file failed"
git_annex "copy" ["--from", "origin", "--auto", annexedfile] @? "copy --auto --from of file failed for client"
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "ungroup" [".", "client"] @? "ungroup failed"
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "wanted" [".", "standard"] @? "set expression to standard failed"
git_annex "group" [".", "manual"] @? "set group to manual failed"
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
-- drop --auto with manual leaves the file where it is
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "drop" ["--auto", annexedfile] @? "drop --auto of file failed with manual preferred content"
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "drop" [annexedfile] @? "drop of file failed"
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
annexed_notpresent annexedfile
add test cases for copy --auto (Only --from for now, should add --to sometime but I'm on low sleep so this is the mininal test to catch the reversion just fixed, bad me for doing it backwards and incompletely.)
2015-10-06 21:35:29 +00:00
-- copy/get --auto with manual does not get the file
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" ["--auto", annexedfile] @? "get --auto of file failed with manual preferred content"
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
annexed_notpresent annexedfile
add test cases for copy --auto (Only --from for now, should add --to sometime but I'm on low sleep so this is the mininal test to catch the reversion just fixed, bad me for doing it backwards and incompletely.)
2015-10-06 21:35:29 +00:00
git_annex "copy" ["--from", "origin", "--auto", annexedfile] @? "copy --auto --from of file failed with manual preferred content"
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "ungroup" [".", "client"] @? "ungroup failed"
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "wanted" [".", "exclude=*"] @? "set expression to exclude=* failed"
git_annex "get" [annexedfile] @? "get of file failed"
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "drop" ["--auto", annexedfile] @? "drop --auto of file failed with exclude=*"
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" ["--auto", annexedfile] @? "get --auto of file failed with exclude=*"
some basic tests of preferred content expressions and standard groups
2013-05-25 17:51:36 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_lock :: Assertion
test_lock = intmpclonerepoInDirect $ do
unlock: Fix behavior when file content is not present.
2011-01-14 03:46:02 +00:00
annexed_notpresent annexedfile
update test suite for lock/unlock with missing file content change in v6
2016-06-09 20:16:39 +00:00
unlessM (annexeval Annex.Version.versionSupportsUnlockedPointers) $
ifM (unlockedFiles <$> getTestMode)
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
( git_annex_shouldfail "lock" [annexedfile] @? "lock failed to fail with not present file"
, git_annex_shouldfail "unlock" [annexedfile] @? "unlock failed to fail with not present file"
update test suite for lock/unlock with missing file content change in v6
2016-06-09 20:16:39 +00:00
)
unlock: Fix behavior when file content is not present.
2011-01-14 03:46:02 +00:00
annexed_notpresent annexedfile
regression test for fd55563b02dd8fedbb2420f440603be7192d94dc
2015-01-02 17:59:25 +00:00
-- regression test: unlock of newly added, not committed file
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
-- should fail in v5 mode. In v7 mode, this is allowed.
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent "newfile" "foo"
regression test for fd55563b02dd8fedbb2420f440603be7192d94dc
2015-01-02 17:59:25 +00:00
git_annex "add" ["newfile"] @? "add new file failed"
test suite down to only 4 fails with v6
2015-12-16 20:56:27 +00:00
ifM (annexeval Annex.Version.versionSupportsUnlockedPointers)
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
( git_annex "unlock" ["newfile"] @? "unlock failed on newly added, never committed file in v7 repository"
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
, git_annex_shouldfail "unlock" ["newfile"] @? "unlock failed to fail on newly added, never committed file in v5 repository"
test suite down to only 4 fails with v6
2015-12-16 20:56:27 +00:00
)
regression test for fd55563b02dd8fedbb2420f440603be7192d94dc
2015-01-02 17:59:25 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [annexedfile] @? "get of file failed"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "unlock" [annexedfile] @? "unlock failed"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
unannexed annexedfile
-- write different content, to verify that lock
-- throws it away
changecontent annexedfile
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent annexedfile $ content annexedfile ++ "foo"
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "lock" [annexedfile] @? "lock failed to fail without --force"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "lock" ["--force", annexedfile] @? "lock --force failed"
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
-- In v7 mode, the original content of the file is not always
test suite down to only 4 fails with v6
2015-12-16 20:56:27 +00:00
-- preserved after modification, so re-get it.
git_annex "get" [annexedfile] @? "get of file failed after lock --force"
fix test file is locked here, so use right test.
2016-01-06 20:12:00 +00:00
annexed_present_locked annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "unlock" [annexedfile] @? "unlock failed"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
unannexed annexedfile
changecontent annexedfile
test suite down to only 4 fails with v6
2015-12-16 20:56:27 +00:00
ifM (annexeval Annex.Version.versionSupportsUnlockedPointers)
( do
boolSystem "git" [Param "add", Param annexedfile] @? "add of modified file failed"
runchecks [checkregularfile, checkwritable] annexedfile
, do
git_annex "add" [annexedfile] @? "add of modified file failed"
runchecks [checklink, checkunwritable] annexedfile
)
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
c <- readFile annexedfile
hlint tweaks Did all sources except Remotes/* and Command/*
2011-07-15 07:12:05 +00:00
assertEqual "content of modified file" c (changedcontent annexedfile)
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
r' <- git_annex "drop" [annexedfile]
unlock: Fix behavior when file content is not present.
2011-01-14 03:46:02 +00:00
not r' @? "drop wrongly succeeded with no known copy of modified file"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
lock: Fix edge cases where data loss could occur in v6 mode. In the case where the pointer file is in place, and not the content of the object, lock's performNew was called with filemodified=True, which caused it to try to repopulate the object from an unmodified associated file, of which there were none. So, the content of the object got thrown away incorrectly. This was the cause (although not the root cause) of data loss in https://github.com/datalad/datalad/issues/1020 The same problem could also occur when the work tree file is modified, but the object is not, and lock is called with --force. Added a test case for this, since it's excercising the same code path and is easier to set up than the problem above. Note that this only occurred when the keys database did not have an inode cache recorded for the annex object. Normally, the annex object would be in there, but there are of course circumstances where the inode cache is out of sync with reality, since it's only a cache. Fixed by checking if the object is unmodified; if so we don't need to try to repopulate it. This does add an additional checksum to the unlock path, but it's already checksumming the worktree file in another case, so it doesn't slow it down overall. Further investigation found a similar problem occurred when smudge --clean is called on a file and the inode cache is not populated. cleanOldKeys deleted the unmodified old object file in this case. This was also fixed by checking if the object is unmodified. In general, use of getInodeCaches and sameInodeCache is potentially dangerous if the inode cache has not gotten populated for some reason. Better to use isUnmodified. I breifly auited other places that check the inode cache, and did not see any immediate problems, but it would be easy to miss this kind of problem.
2016-10-17 16:56:26 +00:00
-- Regression test: lock --force when work tree file
-- was modified lost the (unmodified) annex object.
-- (Only occurred when the keys database was out of sync.)
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
test_lock_v7_force :: Assertion
test_lock_v7_force = intmpclonerepoInDirect $ do
lock: Fix edge cases where data loss could occur in v6 mode. In the case where the pointer file is in place, and not the content of the object, lock's performNew was called with filemodified=True, which caused it to try to repopulate the object from an unmodified associated file, of which there were none. So, the content of the object got thrown away incorrectly. This was the cause (although not the root cause) of data loss in https://github.com/datalad/datalad/issues/1020 The same problem could also occur when the work tree file is modified, but the object is not, and lock is called with --force. Added a test case for this, since it's excercising the same code path and is easier to set up than the problem above. Note that this only occurred when the keys database did not have an inode cache recorded for the annex object. Normally, the annex object would be in there, but there are of course circumstances where the inode cache is out of sync with reality, since it's only a cache. Fixed by checking if the object is unmodified; if so we don't need to try to repopulate it. This does add an additional checksum to the unlock path, but it's already checksumming the worktree file in another case, so it doesn't slow it down overall. Further investigation found a similar problem occurred when smudge --clean is called on a file and the inode cache is not populated. cleanOldKeys deleted the unmodified old object file in this case. This was also fixed by checking if the object is unmodified. In general, use of getInodeCaches and sameInodeCache is potentially dangerous if the inode cache has not gotten populated for some reason. Better to use isUnmodified. I breifly auited other places that check the inode cache, and did not see any immediate problems, but it would be easy to miss this kind of problem.
2016-10-17 16:56:26 +00:00
git_annex "upgrade" [] @? "upgrade failed"
whenM (annexeval Annex.Version.versionSupportsUnlockedPointers) $ do
git_annex "get" [annexedfile] @? "get of file failed"
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
git_annex "unlock" [annexedfile] @? "unlock failed in v7 mode"
lock: Fix edge cases where data loss could occur in v6 mode. In the case where the pointer file is in place, and not the content of the object, lock's performNew was called with filemodified=True, which caused it to try to repopulate the object from an unmodified associated file, of which there were none. So, the content of the object got thrown away incorrectly. This was the cause (although not the root cause) of data loss in https://github.com/datalad/datalad/issues/1020 The same problem could also occur when the work tree file is modified, but the object is not, and lock is called with --force. Added a test case for this, since it's excercising the same code path and is easier to set up than the problem above. Note that this only occurred when the keys database did not have an inode cache recorded for the annex object. Normally, the annex object would be in there, but there are of course circumstances where the inode cache is out of sync with reality, since it's only a cache. Fixed by checking if the object is unmodified; if so we don't need to try to repopulate it. This does add an additional checksum to the unlock path, but it's already checksumming the worktree file in another case, so it doesn't slow it down overall. Further investigation found a similar problem occurred when smudge --clean is called on a file and the inode cache is not populated. cleanOldKeys deleted the unmodified old object file in this case. This was also fixed by checking if the object is unmodified. In general, use of getInodeCaches and sameInodeCache is potentially dangerous if the inode cache has not gotten populated for some reason. Better to use isUnmodified. I breifly auited other places that check the inode cache, and did not see any immediate problems, but it would be easy to miss this kind of problem.
2016-10-17 16:56:26 +00:00
annexeval $ do
don't move keys db out from under sqlite Instead remove enough data from it that this regression test tests what it needs to. Moving the database was the last thing that made the test suite unstable, including sometimes hanging completely. It seems that, despite closeDb being called before the move, sqlite was not quite done with it, and somehow this causes other sqlite handles to become unstable. Not good. With this change, the test suite has successfully run 100+ times without any issues.
2018-10-31 12:04:16 +00:00
Just k <- Annex.WorkTree.lookupFile annexedfile
Database.Keys.removeInodeCaches k
Fix a git-annex test failure when run on NFS due to NFS lock files preventing directory removal. Should fix this: lock (v6 --force): FAIL Exception: .git/annex/keys: removeDirectoryRecursive: unsatisfied constraints (Directory not empty) Verified that the test case still catches the regression it's meant to. This commit was supported by the NSF-funded DataLad project.
2017-08-14 19:11:42 +00:00
Database.Keys.closeDb
don't move keys db out from under sqlite Instead remove enough data from it that this regression test tests what it needs to. Moving the database was the last thing that made the test suite unstable, including sometimes hanging completely. It seems that, despite closeDb being called before the move, sqlite was not quite done with it, and somehow this causes other sqlite handles to become unstable. Not good. With this change, the test suite has successfully run 100+ times without any issues.
2018-10-31 12:04:16 +00:00
liftIO . nukeFile =<< Annex.fromRepo Annex.Locations.gitAnnexKeysDbIndexCache
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent annexedfile "test_lock_v7_force content"
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "lock" [annexedfile] @? "lock of modified file failed to fail in v7 mode"
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
git_annex "lock" ["--force", annexedfile] @? "lock --force of modified file failed in v7 mode"
lock: Fix edge cases where data loss could occur in v6 mode. In the case where the pointer file is in place, and not the content of the object, lock's performNew was called with filemodified=True, which caused it to try to repopulate the object from an unmodified associated file, of which there were none. So, the content of the object got thrown away incorrectly. This was the cause (although not the root cause) of data loss in https://github.com/datalad/datalad/issues/1020 The same problem could also occur when the work tree file is modified, but the object is not, and lock is called with --force. Added a test case for this, since it's excercising the same code path and is easier to set up than the problem above. Note that this only occurred when the keys database did not have an inode cache recorded for the annex object. Normally, the annex object would be in there, but there are of course circumstances where the inode cache is out of sync with reality, since it's only a cache. Fixed by checking if the object is unmodified; if so we don't need to try to repopulate it. This does add an additional checksum to the unlock path, but it's already checksumming the worktree file in another case, so it doesn't slow it down overall. Further investigation found a similar problem occurred when smudge --clean is called on a file and the inode cache is not populated. cleanOldKeys deleted the unmodified old object file in this case. This was also fixed by checking if the object is unmodified. In general, use of getInodeCaches and sameInodeCache is potentially dangerous if the inode cache has not gotten populated for some reason. Better to use isUnmodified. I breifly auited other places that check the inode cache, and did not see any immediate problems, but it would be easy to miss this kind of problem.
2016-10-17 16:56:26 +00:00
annexed_present_locked annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_edit :: Assertion
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
test_edit = test_edit' False
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_edit_precommit :: Assertion
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
test_edit_precommit = test_edit' True
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_edit' :: Bool -> Assertion
test_edit' precommit = intmpclonerepoInDirect $ do
git_annex "get" [annexedfile] @? "get of file failed"
where indenting
2012-11-11 04:51:07 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "edit" [annexedfile] @? "edit failed"
where indenting
2012-11-11 04:51:07 +00:00
unannexed annexedfile
changecontent annexedfile
fix test
2013-05-16 02:10:49 +00:00
boolSystem "git" [Param "add", File annexedfile]
@? "git add of edited file failed"
where indenting
2012-11-11 04:51:07 +00:00
if precommit
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
then git_annex "pre-commit" []
fix test
2013-05-16 02:10:49 +00:00
@? "pre-commit failed"
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
else boolSystem "git" [Param "commit", Param "-q", Param "-m", Param "contentchanged"]
fix test
2013-05-16 02:10:49 +00:00
@? "git commit of edited file failed"
test suite down to only 4 fails with v6
2015-12-16 20:56:27 +00:00
ifM (annexeval Annex.Version.versionSupportsUnlockedPointers)
( runchecks [checkregularfile, checkwritable] annexedfile
, runchecks [checklink, checkunwritable] annexedfile
)
where indenting
2012-11-11 04:51:07 +00:00
c <- readFile annexedfile
assertEqual "content of modified file" c (changedcontent annexedfile)
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "drop" [annexedfile] @? "drop wrongly succeeded with no known copy of modified file"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_partial_commit :: Assertion
test_partial_commit = intmpclonerepoInDirect $ do
git_annex "get" [annexedfile] @? "get of file failed"
pre-commit: Block partial commit of unlocked annexed file, since that left a typechange staged in index I had hoped that the git devs could change git's handling of partial commits to not use a false index file, but seems not. So, this relies on some git internals to detect that case. The test suite has a test case added to catch it if changes to git break it. This commit was sponsored by Paul Tagliamonte.
2014-11-10 19:36:24 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "unlock" [annexedfile] @? "unlock failed"
fix test have to change the content of unlocked file before committing otherwise git commit will fail in v6 mode when the file was already unlocked, because no changes have been made
2016-01-06 20:14:51 +00:00
changecontent annexedfile
test suite down to only 4 fails with v6
2015-12-16 20:56:27 +00:00
ifM (annexeval Annex.Version.versionSupportsUnlockedPointers)
( boolSystem "git" [Param "commit", Param "-q", Param "-m", Param "test", File annexedfile]
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
@? "partial commit of unlocked file should be allowed in v7 repository"
test suite down to only 4 fails with v6
2015-12-16 20:56:27 +00:00
, not <$> boolSystem "git" [Param "commit", Param "-q", Param "-m", Param "test", File annexedfile]
@? "partial commit of unlocked file not blocked by pre-commit hook"
)
pre-commit: Block partial commit of unlocked annexed file, since that left a typechange staged in index I had hoped that the git devs could change git's handling of partial commits to not use a false index file, but seems not. So, this relies on some git internals to detect that case. The test suite has a test case added to catch it if changes to git break it. This commit was sponsored by Paul Tagliamonte.
2014-11-10 19:36:24 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_fix :: Assertion
skip fix test in unlocked v6 pass fix is not relevant for unlocked files
2016-01-06 20:16:19 +00:00
test_fix = intmpclonerepoInDirect $ unlessM (unlockedFiles <$> getTestMode) $ do
more
2011-01-07 18:08:43 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "fix" [annexedfile] @? "fix of not present failed"
more
2011-01-07 18:08:43 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [annexedfile] @? "get of file failed"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "fix" [annexedfile] @? "fix of present file failed"
more
2011-01-07 18:08:43 +00:00
annexed_present annexedfile
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
createDirectory subdir
split groups of related functions out of Utility
2011-08-22 20:14:12 +00:00
boolSystem "git" [Param "mv", File annexedfile, File subdir]
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
@? "git mv failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "fix" [newfile] @? "fix of moved file failed"
Test suite improvements. Current top-level test coverage: 65%
2011-01-07 18:06:32 +00:00
runchecks [checklink, checkunwritable] newfile
c <- readFile newfile
hlint tweaks Did all sources except Remotes/* and Command/*
2011-07-15 07:12:05 +00:00
assertEqual "content of moved file" c (content annexedfile)
where indenting
2012-11-11 04:51:07 +00:00
where
subdir = "s"
newfile = subdir ++ "/" ++ annexedfile
more tests
2011-01-07 01:39:26 +00:00
add a test case for switching to direct mode and back to indirect
2015-02-09 17:41:29 +00:00
test_direct :: Assertion
test_direct = intmpclonerepoInDirect $ do
git_annex "get" [annexedfile] @? "get of file failed"
annexed_present annexedfile
test suite down to only 4 fails with v6
2015-12-16 20:56:27 +00:00
ifM (annexeval Annex.Version.versionSupportsUnlockedPointers)
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
( git_annex_shouldfail "direct" [] @? "switch to direct mode failed to fail in v7 repository"
test suite down to only 4 fails with v6
2015-12-16 20:56:27 +00:00
, do
git_annex "direct" [] @? "switch to direct mode failed"
annexed_present annexedfile
git_annex "indirect" [] @? "switch to indirect mode failed"
)
add a test case for switching to direct mode and back to indirect
2015-02-09 17:41:29 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_trust :: Assertion
test_trust = intmpclonerepo $ do
git_annex "trust" [repo] @? "trust failed"
reorganize log modules no code changes
2011-10-15 20:21:08 +00:00
trustcheck Logs.Trust.Trusted "trusted 1"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "trust" [repo] @? "trust of trusted failed"
reorganize log modules no code changes
2011-10-15 20:21:08 +00:00
trustcheck Logs.Trust.Trusted "trusted 2"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "untrust" [repo] @? "untrust failed"
reorganize log modules no code changes
2011-10-15 20:21:08 +00:00
trustcheck Logs.Trust.UnTrusted "untrusted 1"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "untrust" [repo] @? "untrust of untrusted failed"
reorganize log modules no code changes
2011-10-15 20:21:08 +00:00
trustcheck Logs.Trust.UnTrusted "untrusted 2"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "dead" [repo] @? "dead failed"
add basic tests for some of the newer commands For most this just makes sure they run successfully, and improves test coverage statistics; it's hard to test the output of eg, find and status.
2011-12-20 20:26:39 +00:00
trustcheck Logs.Trust.DeadTrusted "deadtrusted 1"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "dead" [repo] @? "dead of dead failed"
add basic tests for some of the newer commands For most this just makes sure they run successfully, and improves test coverage statistics; it's hard to test the output of eg, find and status.
2011-12-20 20:26:39 +00:00
trustcheck Logs.Trust.DeadTrusted "deadtrusted 2"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "semitrust" [repo] @? "semitrust failed"
reorganize log modules no code changes
2011-10-15 20:21:08 +00:00
trustcheck Logs.Trust.SemiTrusted "semitrusted 1"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "semitrust" [repo] @? "semitrust of semitrusted failed"
reorganize log modules no code changes
2011-10-15 20:21:08 +00:00
trustcheck Logs.Trust.SemiTrusted "semitrusted 2"
where indenting
2012-11-11 04:51:07 +00:00
where
repo = "origin"
trustcheck expected msg = do
present <- annexeval $ do
l <- Logs.Trust.trustGet expected
u <- Remote.nameToUUID repo
return $ u `elem` l
assertBool msg present
add tests for trust/untrust
2011-01-11 22:50:18 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_fsck_basic :: Assertion
test_fsck_basic = intmpclonerepo $ do
git_annex "fsck" [] @? "fsck failed"
git_annex "numcopies" ["2"] @? "numcopies config failed"
fsck_should_fail "numcopies unsatisfied"
git_annex "numcopies" ["1"] @? "numcopies config failed"
tasty fully working
2013-11-15 00:09:47 +00:00
corrupt annexedfile
corrupt sha1annexedfile
where indenting
2012-11-11 04:51:07 +00:00
where
corrupt f = do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [f] @? "get of file failed"
where indenting
2012-11-11 04:51:07 +00:00
Utility.FileMode.allowWrite f
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent f (changedcontent f)
unused: Bug fix when a new file was added to the annex, and then removed (but not git rmed). git still has the add staged in this case, so the content should not be unused and was wrongly treated as such. So, we need to look at both the file on disk to see if it's a annex link, and the file in the index too. lookupFile doesn't look in the index if the file is not present on disk.
2016-01-06 20:30:49 +00:00
ifM (annexeval Config.isDirect <||> unlockedFiles <$> getTestMode)
( git_annex "fsck" [] @? "fsck failed on unlocked file with changed file content"
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
, git_annex_shouldfail "fsck" [] @? "fsck failed to fail with corrupted file content"
fsck behavior varies in direct mode when the file is modified
2013-05-17 19:20:21 +00:00
)
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "fsck" [] @? "fsck unexpectedly failed again; previous one did not fix problem with " ++ f
test_fsck_bare :: Assertion
test_fsck_bare = intmpbareclonerepo $
git_annex "fsck" [] @? "fsck failed"
test_fsck_localuntrusted :: Assertion
test_fsck_localuntrusted = intmpclonerepo $ do
git_annex "get" [annexedfile] @? "get failed"
git_annex "untrust" ["origin"] @? "untrust of origin repo failed"
git_annex "untrust" ["."] @? "untrust of current repo failed"
fsck_should_fail "content only available in untrusted (current) repository"
git_annex "trust" ["."] @? "trust of current repo failed"
git_annex "fsck" [annexedfile] @? "fsck failed on file present in trusted repo"
test_fsck_remoteuntrusted :: Assertion
test_fsck_remoteuntrusted = intmpclonerepo $ do
git_annex "numcopies" ["2"] @? "numcopies config failed"
git_annex "get" [annexedfile] @? "get failed"
git_annex "get" [sha1annexedfile] @? "get failed"
git_annex "fsck" [] @? "fsck failed with numcopies=2 and 2 copies"
git_annex "untrust" ["origin"] @? "untrust of origin failed"
fsck_should_fail "content not replicated to enough non-untrusted repositories"
Fix fsck --from a git remote in a local directory, and from a directory special remote. This was a reversion caused by the relative path changes in 5.20150113. The directory special remote was not affected in its normal configuration, since annex-directory is an absolute path normally. But it could fail when a relative path was used. The git remote was affected even when an absolute path to it was used in .git/config, since git-annex now converts all such paths to relative.
2015-04-18 17:36:12 +00:00
test_fsck_fromremote :: Assertion
test_fsck_fromremote = intmpclonerepo $ do
git_annex "fsck" ["--from", "origin"] @? "fsck --from origin failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
fsck_should_fail :: String -> Assertion
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
fsck_should_fail m = git_annex_shouldfail "fsck" []
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
@? "fsck failed to fail with " ++ m
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_migrate :: Assertion
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
test_migrate = test_migrate' False
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_migrate_via_gitattributes :: Assertion
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
test_migrate_via_gitattributes = test_migrate' True
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_migrate' :: Bool -> Assertion
test_migrate' usegitattributes = intmpclonerepoInDirect $ do
where indenting
2012-11-11 04:51:07 +00:00
annexed_notpresent annexedfile
annexed_notpresent sha1annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "migrate" [annexedfile] @? "migrate of not present failed"
git_annex "migrate" [sha1annexedfile] @? "migrate of not present failed"
git_annex "get" [annexedfile] @? "get of file failed"
git_annex "get" [sha1annexedfile] @? "get of file failed"
where indenting
2012-11-11 04:51:07 +00:00
annexed_present annexedfile
annexed_present sha1annexedfile
if usegitattributes
then do
more hlint
2014-02-11 05:35:11 +00:00
writeFile ".gitattributes" "* annex.backend=SHA1"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "migrate" [sha1annexedfile]
where indenting
2012-11-11 04:51:07 +00:00
@? "migrate sha1annexedfile failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "migrate" [annexedfile]
where indenting
2012-11-11 04:51:07 +00:00
@? "migrate annexedfile failed"
else do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "migrate" [sha1annexedfile, "--backend", "SHA1"]
where indenting
2012-11-11 04:51:07 +00:00
@? "migrate sha1annexedfile failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "migrate" [annexedfile, "--backend", "SHA1"]
where indenting
2012-11-11 04:51:07 +00:00
@? "migrate annexedfile failed"
annexed_present annexedfile
annexed_present sha1annexedfile
checkbackend annexedfile backendSHA1
checkbackend sha1annexedfile backendSHA1
-- check that reversing a migration works
more hlint
2014-02-11 05:35:11 +00:00
writeFile ".gitattributes" "* annex.backend=SHA256"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "migrate" [sha1annexedfile]
where indenting
2012-11-11 04:51:07 +00:00
@? "migrate sha1annexedfile failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "migrate" [annexedfile]
where indenting
2012-11-11 04:51:07 +00:00
@? "migrate annexedfile failed"
annexed_present annexedfile
annexed_present sha1annexedfile
checkbackend annexedfile backendSHA256
checkbackend sha1annexedfile backendSHA256
test migrate; 75%
2011-01-12 01:11:32 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_unused :: Assertion
disable unused test in direct mode This test is actually pointing at a problem, with getKeysPresent in direct mode. When a file is git rm -f'd, this removes its content, but it's still listed by getKeysPresent. However, unused is fairly useless in direct mode, and the test needs to be updated to handle other differences, so disable it for now
2013-05-17 19:16:28 +00:00
-- This test is broken in direct mode
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_unused = intmpclonerepoInDirect $ do
set annex.largefiles earlier in test suite The ingitfile was having git run it through the clean filter in some cases.
2015-12-30 17:29:00 +00:00
checkunused [] "in new clone"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [annexedfile] @? "get of file failed"
git_annex "get" [sha1annexedfile] @? "get of file failed"
unused: deal with v6 unlocked file that is implicitly ingested by git diff etc
2016-01-07 02:11:21 +00:00
annexedfilekey <- getKey backendSHA256E annexedfile
sha1annexedfilekey <- getKey backendSHA1 sha1annexedfile
fix breakage caught by test suite Can't use show-ref --tags --branches, as that omits remote branches. Instead, filter out the synced refs directly.
2012-09-21 04:49:48 +00:00
checkunused [] "after get"
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "rm", Param "-fq", File annexedfile] @? "git rm failed"
fix breakage caught by test suite Can't use show-ref --tags --branches, as that omits remote branches. Instead, filter out the synced refs directly.
2012-09-21 04:49:48 +00:00
checkunused [] "after rm"
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "commit", Param "-q", Param "-m", Param "foo"] @? "git commit failed"
fix breakage caught by test suite Can't use show-ref --tags --branches, as that omits remote branches. Instead, filter out the synced refs directly.
2012-09-21 04:49:48 +00:00
checkunused [] "after commit"
update test suite for smarter unused
2011-09-28 21:48:11 +00:00
-- unused checks origin/master; once it's gone it is really unused
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "remote", Param "rm", Param "origin"] @? "git remote rm origin failed"
fix breakage caught by test suite Can't use show-ref --tags --branches, as that omits remote branches. Instead, filter out the synced refs directly.
2012-09-21 04:49:48 +00:00
checkunused [annexedfilekey] "after origin branches are gone"
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "rm", Param "-fq", File sha1annexedfile] @? "git rm failed"
boolSystem "git" [Param "commit", Param "-q", Param "-m", Param "foo"] @? "git commit failed"
fix breakage caught by test suite Can't use show-ref --tags --branches, as that omits remote branches. Instead, filter out the synced refs directly.
2012-09-21 04:49:48 +00:00
checkunused [annexedfilekey, sha1annexedfilekey] "after rm sha1annexedfile"
test unused et al
2011-01-12 05:58:23 +00:00
-- good opportunity to test dropkey also
factor non-type stuff out of Key
2017-02-24 17:42:30 +00:00
git_annex "dropkey" ["--force", Key.key2file annexedfilekey]
test unused et al
2011-01-12 05:58:23 +00:00
@? "dropkey failed"
factor non-type stuff out of Key
2017-02-24 17:42:30 +00:00
checkunused [sha1annexedfilekey] ("after dropkey --force " ++ Key.key2file annexedfilekey)
test unused et al
2011-01-12 05:58:23 +00:00
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "dropunused" ["1"] @? "dropunused failed to fail without --force"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "dropunused" ["--force", "1"] @? "dropunused failed"
fix breakage caught by test suite Can't use show-ref --tags --branches, as that omits remote branches. Instead, filter out the synced refs directly.
2012-09-21 04:49:48 +00:00
checkunused [] "after dropunused"
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "dropunused" ["--force", "10", "501"] @? "dropunused failed to fail on bogus numbers"
test unused et al
2011-01-12 05:58:23 +00:00
100% pass on FAT
2016-02-16 20:02:54 +00:00
-- Unused used to miss renamed symlinks that were not staged
-- and pointed at annexed content, and think that content was unused.
-- This is only relevant when using locked files; if the file is
-- unlocked, the work tree file has the content, and there's no way
-- to associate it with the key.
unlessM (unlockedFiles <$> getTestMode) $ do
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent "unusedfile" "unusedcontent"
100% pass on FAT
2016-02-16 20:02:54 +00:00
git_annex "add" ["unusedfile"] @? "add of unusedfile failed"
unusedfilekey <- getKey backendSHA256E "unusedfile"
renameFile "unusedfile" "unusedunstagedfile"
boolSystem "git" [Param "rm", Param "-qf", File "unusedfile"] @? "git rm failed"
checkunused [] "with unstaged link"
removeFile "unusedunstagedfile"
checkunused [unusedfilekey] "with renamed link deleted"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
-- unused used to miss symlinks that were deleted or modified
unused: Bug fix when a new file was added to the annex, and then removed (but not git rmed). git still has the add staged in this case, so the content should not be unused and was wrongly treated as such. So, we need to look at both the file on disk to see if it's a annex link, and the file in the index too. lookupFile doesn't look in the index if the file is not present on disk.
2016-01-06 20:30:49 +00:00
-- manually
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent "unusedfile" "unusedcontent"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "add" ["unusedfile"] @? "add of unusedfile failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
boolSystem "git" [Param "add", File "unusedfile"] @? "git add failed"
unused: deal with v6 unlocked file that is implicitly ingested by git diff etc
2016-01-07 02:11:21 +00:00
unusedfilekey' <- getKey backendSHA256E "unusedfile"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
checkunused [] "with staged deleted link"
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "rm", Param "-qf", File "unusedfile"] @? "git rm failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
checkunused [unusedfilekey'] "with staged link deleted"
unused: Bug fix when a new file was added to the annex, and then removed (but not git rmed). git still has the add staged in this case, so the content should not be unused and was wrongly treated as such. So, we need to look at both the file on disk to see if it's a annex link, and the file in the index too. lookupFile doesn't look in the index if the file is not present on disk.
2016-01-06 20:30:49 +00:00
-- unused used to false positive on symlinks that were
-- deleted or modified manually, but not staged as such
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent "unusedfile" "unusedcontent"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "add" ["unusedfile"] @? "add of unusedfile failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
boolSystem "git" [Param "add", File "unusedfile"] @? "git add failed"
unused: Bug fix when a new file was added to the annex, and then removed (but not git rmed). git still has the add staged in this case, so the content should not be unused and was wrongly treated as such. So, we need to look at both the file on disk to see if it's a annex link, and the file in the index too. lookupFile doesn't look in the index if the file is not present on disk.
2016-01-06 20:30:49 +00:00
checkunused [] "with staged file"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
removeFile "unusedfile"
unused: Bug fix when a new file was added to the annex, and then removed (but not git rmed). git still has the add staged in this case, so the content should not be unused and was wrongly treated as such. So, we need to look at both the file on disk to see if it's a annex link, and the file in the index too. lookupFile doesn't look in the index if the file is not present on disk.
2016-01-06 20:30:49 +00:00
checkunused [] "with staged deleted file"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
unused: deal with v6 unlocked file that is implicitly ingested by git diff etc
2016-01-07 02:11:21 +00:00
-- When an unlocked file is modified, git diff will cause git-annex
-- to add its content to the repository. Make sure that's not
-- found as unused.
whenM (unlockedFiles <$> getTestMode) $ do
let f = "unlockedfile"
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent f "unlockedcontent1"
unused: deal with v6 unlocked file that is implicitly ingested by git diff etc
2016-01-07 02:11:21 +00:00
boolSystem "git" [Param "add", File "unlockedfile"] @? "git add failed"
checkunused [] "with unlocked file before modification"
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent f "unlockedcontent2"
unused: deal with v6 unlocked file that is implicitly ingested by git diff etc
2016-01-07 02:11:21 +00:00
checkunused [] "with unlocked file after modification"
not <$> boolSystem "git" [Param "diff", Param "--quiet", File f] @? "git diff did not show changes to unlocked file"
-- still nothing unused because one version is in the index
-- and the other is in the work tree
checkunused [] "with unlocked file after git diff"
where indenting
2012-11-11 04:51:07 +00:00
where
checkunused expectedkeys desc = do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "unused" [] @? "unused failed"
add timestamps to unused log files This will be used in expiring old unused objects. The timestamp is when it was first noticed it was unused. Backwards compatability: It supports reading old format unused log files. The old version of git-annex will ignore lines in log files written by the new version, so the worst interop problem would be git annex dropunused not knowing some numbers that git-annex unused reported.
2014-01-22 19:33:02 +00:00
unusedmap <- annexeval $ Logs.Unused.readUnusedMap ""
where indenting
2012-11-11 04:51:07 +00:00
let unusedkeys = M.elems unusedmap
assertEqual ("unused keys differ " ++ desc)
(sort expectedkeys) (sort unusedkeys)
test unused et al
2011-01-12 05:58:23 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_describe :: Assertion
test_describe = intmpclonerepo $ do
git_annex "describe" [".", "this repo"] @? "describe 1 failed"
git_annex "describe" ["origin", "origin repo"] @? "describe 2 failed"
add basic tests for some of the newer commands For most this just makes sure they run successfully, and improves test coverage statistics; it's hard to test the output of eg, find and status.
2011-12-20 20:26:39 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_find :: Assertion
test_find = intmpclonerepo $ do
add basic tests for some of the newer commands For most this just makes sure they run successfully, and improves test coverage statistics; it's hard to test the output of eg, find and status.
2011-12-20 20:26:39 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex_expectoutput "find" [] []
git_annex "get" [annexedfile] @? "get failed"
test coverage improvements
2011-12-21 06:32:40 +00:00
annexed_present annexedfile
annexed_notpresent sha1annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex_expectoutput "find" [] [annexedfile]
git_annex_expectoutput "find" ["--exclude", annexedfile, "--and", "--exclude", sha1annexedfile] []
git_annex_expectoutput "find" ["--include", annexedfile] [annexedfile]
git_annex_expectoutput "find" ["--not", "--in", "origin"] []
git_annex_expectoutput "find" ["--copies", "1", "--and", "--not", "--copies", "2"] [sha1annexedfile]
git_annex_expectoutput "find" ["--inbackend", "SHA1"] [sha1annexedfile]
git_annex_expectoutput "find" ["--inbackend", "WORM"] []
add basic tests for some of the newer commands For most this just makes sure they run successfully, and improves test coverage statistics; it's hard to test the output of eg, find and status.
2011-12-20 20:26:39 +00:00
fix * glob matching files in subdirectories
2013-02-26 17:49:22 +00:00
{- --include=* should match files in subdirectories too,
- and --exclude=* should exclude them. -}
createDirectory "dir"
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent "dir/subfile" "subfile"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "add" ["dir"] @? "add of subdir failed"
git_annex_expectoutput "find" ["--include", "*", "--exclude", annexedfile, "--exclude", sha1annexedfile] ["dir/subfile"]
git_annex_expectoutput "find" ["--exclude", "*"] []
fix * glob matching files in subdirectories
2013-02-26 17:49:22 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_merge :: Assertion
test_merge = intmpclonerepo $
git_annex "merge" [] @? "merge failed"
add basic tests for some of the newer commands For most this just makes sure they run successfully, and improves test coverage statistics; it's hard to test the output of eg, find and status.
2011-12-20 20:26:39 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_info :: Assertion
test_info = intmpclonerepo $ do
Removed dependency on json library; all JSON is now handled by aeson. I've eyeballed all --json commands, and the only difference should be that some fields are re-ordered.
2016-07-26 23:15:34 +00:00
json <- BU8.fromString <$> git_annex_output "info" ["--json"]
Fix mangling of --json output of utf-8 characters when not running in a utf-8 locale As long as all code imports Utility.Aeson rather than Data.Aeson, and no Strings that may contain utf-8 characters are used for eg, object keys via T.pack, this is guaranteed to fix the problem everywhere that git-annex generates json. It's kind of annoying to need to wrap ToJSON with a ToJSON', especially since every data type that has a ToJSON instance has to be ported over. However, that only took 50 lines of code, which is worth it to ensure full coverage. I initially tried an alternative approach of a newtype FileEncoded, which had to be used everywhere a String was fed into aeson, and chasing down all the sites would have been far too hard. Did consider creating an intentionally overlapping instance ToJSON String, and letting ghc fail to build anything that passed in a String, but am not sure that wouldn't pollute some library that git-annex depends on that happens to use ToJSON String internally. This commit was supported by the NSF-funded DataLad project.
2018-04-16 19:42:45 +00:00
case Utility.Aeson.eitherDecode json :: Either String Utility.Aeson.Value of
Removed dependency on json library; all JSON is now handled by aeson. I've eyeballed all --json commands, and the only difference should be that some fields are re-ordered.
2016-07-26 23:15:34 +00:00
Right _ -> return ()
Left e -> assertFailure e
add basic tests for some of the newer commands For most this just makes sure they run successfully, and improves test coverage statistics; it's hard to test the output of eg, find and status.
2011-12-20 20:26:39 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_version :: Assertion
test_version = intmpclonerepo $
git_annex "version" [] @? "version failed"
Test suite improvements. Current top-level test coverage: 68% Been higher before, but a lot of new code has been added.
2011-12-20 21:31:25 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_sync :: Assertion
test_sync = intmpclonerepo $ do
git_annex "sync" [] @? "sync failed"
regression test for 7b0970b340d7faeb745c666146c7f701ec71808f
2013-07-30 21:03:50 +00:00
{- Regression test for bug fixed in
- 7b0970b340d7faeb745c666146c7f701ec71808f, where in direct mode
- sync committed the symlink standin file to the annex. -}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex_expectoutput "find" ["--in", "."] []
test case for recent sync --content reversion
2016-01-26 21:33:16 +00:00
{- Regression test for bug fixed in
- 039e83ed5d1a11fd562cce55b8429c840d72443e, where a present
- wanted file was dropped. -}
git_annex "get" [annexedfile] @? "get failed"
git_annex_expectoutput "find" ["--in", "."] [annexedfile]
git_annex "wanted" [".", "present"] @? "wanted failed"
git_annex "sync" ["--content"] @? "sync failed"
git_annex_expectoutput "find" ["--in", "."] [annexedfile]
git_annex "drop" [annexedfile] @? "drop failed"
git_annex_expectoutput "find" ["--in", "."] []
git_annex "sync" ["--content"] @? "sync failed"
git_annex_expectoutput "find" ["--in", "."] []
add basic tests for some of the newer commands For most this just makes sure they run successfully, and improves test coverage statistics; it's hard to test the output of eg, find and status.
2011-12-20 20:26:39 +00:00
add a test case for merge conflict automatic resolution
2013-05-20 19:51:40 +00:00
{- Regression test for union merge bug fixed in
Regression test for union merge bug Fixed in 0214e0fb175a608a49b812d81b4632c081f63027
2013-01-17 20:02:58 +00:00
- 0214e0fb175a608a49b812d81b4632c081f63027 -}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_union_merge_regression :: Assertion
test_union_merge_regression =
Regression test for union merge bug Fixed in 0214e0fb175a608a49b812d81b4632c081f63027
2013-01-17 20:02:58 +00:00
{- We need 3 repos to see this bug. -}
init: Fix reversion in detection of repo made with git clone --shared
2015-09-09 17:56:37 +00:00
withtmpclonerepo $ \r1 ->
withtmpclonerepo $ \r2 ->
withtmpclonerepo $ \r3 -> do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
forM_ [r1, r2, r3] $ \r -> indir r $ do
Regression test for union merge bug Fixed in 0214e0fb175a608a49b812d81b4632c081f63027
2013-01-17 20:02:58 +00:00
when (r /= r1) $
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "remote", Param "add", Param "r1", File ("../../" ++ r1)] @? "remote add"
Regression test for union merge bug Fixed in 0214e0fb175a608a49b812d81b4632c081f63027
2013-01-17 20:02:58 +00:00
when (r /= r2) $
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "remote", Param "add", Param "r2", File ("../../" ++ r2)] @? "remote add"
Regression test for union merge bug Fixed in 0214e0fb175a608a49b812d81b4632c081f63027
2013-01-17 20:02:58 +00:00
when (r /= r3) $
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "remote", Param "add", Param "r3", File ("../../" ++ r3)] @? "remote add"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [annexedfile] @? "get failed"
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "remote", Param "rm", Param "origin"] @? "remote rm"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
forM_ [r3, r2, r1] $ \r -> indir r $
better failure diagnosis
2016-06-03 16:51:27 +00:00
git_annex "sync" [] @? ("sync failed in " ++ r)
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
forM_ [r3, r2] $ \r -> indir r $
better failure diagnosis
2016-06-03 16:51:27 +00:00
git_annex "drop" ["--force", annexedfile] @? ("drop failed in " ++ r)
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir r1 $ do
git_annex "sync" [] @? "sync failed in r1"
git_annex_expectoutput "find" ["--in", "r3"] []
Regression test for union merge bug Fixed in 0214e0fb175a608a49b812d81b4632c081f63027
2013-01-17 20:02:58 +00:00
{- This was the bug. The sync
- mangled location log data and it
- thought the file was still in r2 -}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex_expectoutput "find" ["--in", "r2"] []
Regression test for union merge bug Fixed in 0214e0fb175a608a49b812d81b4632c081f63027
2013-01-17 20:02:58 +00:00
add a test case for merge conflict automatic resolution
2013-05-20 19:51:40 +00:00
{- Regression test for the automatic conflict resolution bug fixed
- in f4ba19f2b8a76a1676da7bb5850baa40d9c388e2. -}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_conflict_resolution_movein_regression :: Assertion
init: Fix reversion in detection of repo made with git clone --shared
2015-09-09 17:56:37 +00:00
test_conflict_resolution_movein_regression = withtmpclonerepo $ \r1 ->
withtmpclonerepo $ \r2 -> do
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
let rname r = if r == r1 then "r1" else "r2"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
forM_ [r1, r2] $ \r -> indir r $ do
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
{- Get all files, see check below. -}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [] @? "get failed"
fix test suite to pass on windows On windows, the sync of the second cloned repo to origin failed, because synced/master was a non-fast-forward. This may be a bug of its own, but it's not the issue that this test was intended to test, so disconnect the repos from origin before syncing.
2014-02-03 16:03:10 +00:00
disconnectOrigin
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
pair r1 r2
forM_ [r1, r2] $ \r -> indir r $ do
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
{- Set up a conflict. -}
let newcontent = content annexedfile ++ rname r
ifM (annexeval Config.isDirect)
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
( writecontent annexedfile newcontent
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
, do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "unlock" [annexedfile] @? "unlock failed"
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent annexedfile newcontent
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
)
{- Sync twice in r1 so it gets the conflict resolution
- update from r2 -}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
forM_ [r1, r2, r1] $ \r -> indir r $
git_annex "sync" ["--force"] @? "sync failed in " ++ rname r
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
{- After the sync, it should be possible to get all
- files. This includes both sides of the conflict,
- although the filenames are not easily predictable.
-
- The bug caused, in direct mode, one repo to
- be missing the content of the file that had
- been put in it. -}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
forM_ [r1, r2] $ \r -> indir r $ do
git_annex "get" [] @? "unable to get all files after merge conflict resolution in " ++ rname r
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
more conflict resolution tests
2014-03-05 03:12:15 +00:00
{- Simple case of conflict resolution; 2 different versions of annexed
- file. -}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_conflict_resolution :: Assertion
test_conflict_resolution =
init: Fix reversion in detection of repo made with git clone --shared
2015-09-09 17:56:37 +00:00
withtmpclonerepo $ \r1 ->
withtmpclonerepo $ \r2 -> do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir r1 $ do
more conflict resolution tests
2014-03-05 03:12:15 +00:00
disconnectOrigin
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor1"
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
add_annex conflictor @? "add conflicter failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed in r1"
indir r2 $ do
more conflict resolution tests
2014-03-05 03:12:15 +00:00
disconnectOrigin
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor2"
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
add_annex conflictor @? "add conflicter failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed in r2"
pair r1 r2
forM_ [r1,r2,r1] $ \r -> indir r $
git_annex "sync" [] @? "sync failed"
more conflict resolution tests
2014-03-05 03:12:15 +00:00
checkmerge "r1" r1
checkmerge "r2" r2
where
conflictor = "conflictor"
variantprefix = conflictor ++ ".variant"
checkmerge what d = do
l <- getDirectoryContents d
let v = filter (variantprefix `isPrefixOf`) l
length v == 2
@? (what ++ " not exactly 2 variant files in: " ++ show l)
Deal with change in git 2.0 that made indirect mode merge conflict resolution leave behind old files. I think this is a git behavior change, but have not checked to be sure. Conflict cruft used to look like $foo~HEAD, but now just $foo is left behind as conflict cruft. With test case.
2014-07-11 20:56:19 +00:00
conflictor `notElem` l @? ("conflictor still present after conflict resolution")
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir d $ do
git_annex "get" v @? "get failed"
git_annex_expectoutput "find" v v
more conflict resolution tests
2014-03-05 03:12:15 +00:00
add simple test for conflict resolution in adjusted branch This is not really extensive enough, but a start..
2016-04-12 17:01:31 +00:00
{- Conflict resolution while in an adjusted branch. -}
test_conflict_resolution_adjusted_branch :: Assertion
Fix test suite failure when git-annex test is not run inside a git repository Not the first time this kind of test suite breakage has happened.. It would be good to avoid somehow it looking up from .t and finding a git repo. But just running the test suite from time to time outside of git-annex would also let me notice these before the distribution packagers do. This commit was sponsored by mo on Patreon.
2018-11-05 17:27:16 +00:00
test_conflict_resolution_adjusted_branch =
add simple test for conflict resolution in adjusted branch This is not really extensive enough, but a start..
2016-04-12 17:01:31 +00:00
withtmpclonerepo $ \r1 ->
Fix test suite failure when git-annex test is not run inside a git repository Not the first time this kind of test suite breakage has happened.. It would be good to avoid somehow it looking up from .t and finding a git repo. But just running the test suite from time to time outside of git-annex would also let me notice these before the distribution packagers do. This commit was sponsored by mo on Patreon.
2018-11-05 17:27:16 +00:00
withtmpclonerepo $ \r2 -> whenM (adjustedbranchsupported r2) $ do
add simple test for conflict resolution in adjusted branch This is not really extensive enough, but a start..
2016-04-12 17:01:31 +00:00
indir r1 $ do
disconnectOrigin
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor1"
add simple test for conflict resolution in adjusted branch This is not really extensive enough, but a start..
2016-04-12 17:01:31 +00:00
add_annex conflictor @? "add conflicter failed"
git_annex "sync" [] @? "sync failed in r1"
indir r2 $ do
disconnectOrigin
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor2"
add simple test for conflict resolution in adjusted branch This is not really extensive enough, but a start..
2016-04-12 17:01:31 +00:00
add_annex conflictor @? "add conflicter failed"
git_annex "sync" [] @? "sync failed in r2"
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
-- need v7 to use adjust
fix recent test suite reversion git annex adjust --force will overwrite any current adjusted branch. I didn't document this because for the user, deleting the branch is just as good.
2016-05-23 15:23:30 +00:00
git_annex "upgrade" [] @? "upgrade failed"
-- We might be in an adjusted branch
-- already, when eg on a crippled
-- filesystem. So, --force it.
git_annex "adjust" ["--unlock", "--force"] @? "adjust failed"
add simple test for conflict resolution in adjusted branch This is not really extensive enough, but a start..
2016-04-12 17:01:31 +00:00
pair r1 r2
forM_ [r1,r2,r1] $ \r -> indir r $
git_annex "sync" [] @? "sync failed"
checkmerge "r1" r1
checkmerge "r2" r2
where
conflictor = "conflictor"
variantprefix = conflictor ++ ".variant"
checkmerge what d = do
l <- getDirectoryContents d
let v = filter (variantprefix `isPrefixOf`) l
length v == 2
@? (what ++ " not exactly 2 variant files in: " ++ show l)
conflictor `notElem` l @? ("conflictor still present after conflict resolution")
indir d $ do
git_annex "get" v @? "get failed"
git_annex_expectoutput "find" v v
more conflict resolution tests
2014-03-05 03:12:15 +00:00
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
{- Check merge conflict resolution when one side is an annexed
- file, and the other is a directory. -}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_mixed_conflict_resolution :: Assertion
test_mixed_conflict_resolution = do
added test for conflict resolution with a file that is both deleted and modified
2014-03-04 21:12:26 +00:00
check True
check False
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
where
init: Fix reversion in detection of repo made with git clone --shared
2015-09-09 17:56:37 +00:00
check inr1 = withtmpclonerepo $ \r1 ->
withtmpclonerepo $ \r2 -> do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir r1 $ do
fix test suite to pass on windows On windows, the sync of the second cloned repo to origin failed, because synced/master was a non-fast-forward. This may be a bug of its own, but it's not the issue that this test was intended to test, so disconnect the repos from origin before syncing.
2014-02-03 16:03:10 +00:00
disconnectOrigin
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor"
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
add_annex conflictor @? "add conflicter failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed in r1"
indir r2 $ do
fix test suite to pass on windows On windows, the sync of the second cloned repo to origin failed, because synced/master was a non-fast-forward. This may be a bug of its own, but it's not the issue that this test was intended to test, so disconnect the repos from origin before syncing.
2014-02-03 16:03:10 +00:00
disconnectOrigin
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
createDirectory conflictor
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent subfile "subfile"
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
add_annex conflictor @? "add conflicter failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed in r2"
pair r1 r2
fix broken test 1c997fd08c04b27c9acb561a33c6fb2829089be7 broke the mixed conflict resolution test. Since origin is no longer involved, both sides have to sync after pairing. (The sync before pairing is just to get a commit.)
2014-02-03 18:32:07 +00:00
let l = if inr1 then [r1, r2] else [r2, r1]
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
forM_ l $ \r -> indir r $
git_annex "sync" [] @? "sync failed in mixed conflict"
fix broken test 1c997fd08c04b27c9acb561a33c6fb2829089be7 broke the mixed conflict resolution test. Since origin is no longer involved, both sides have to sync after pairing. (The sync before pairing is just to get a commit.)
2014-02-03 18:32:07 +00:00
checkmerge "r1" r1
added test for conflict resolution with a file that is both deleted and modified
2014-03-04 21:12:26 +00:00
checkmerge "r2" r2
narrowing test case
2014-02-04 19:14:16 +00:00
conflictor = "conflictor"
improved test for FAT symlink regressions The test case that was supposed to cover this did not.. It is for some other bug, which is apparently a bit of a heisenbug.
2014-03-04 17:23:47 +00:00
subfile = conflictor </> "subfile"
narrowing test case
2014-02-04 19:14:16 +00:00
variantprefix = conflictor ++ ".variant"
checkmerge what d = do
doesDirectoryExist (d </> conflictor) @? (d ++ " conflictor directory missing")
l <- getDirectoryContents d
improved test for FAT symlink regressions The test case that was supposed to cover this did not.. It is for some other bug, which is apparently a bit of a heisenbug.
2014-03-04 17:23:47 +00:00
let v = filter (variantprefix `isPrefixOf`) l
not (null v)
more conflict resolution tests
2014-03-05 03:12:15 +00:00
@? (what ++ " conflictor variant file missing in: " ++ show l )
length v == 1
@? (what ++ " too many variant files in: " ++ show v)
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir d $ do
git_annex "get" (conflictor:v) @? ("get failed in " ++ what)
git_annex_expectoutput "find" [conflictor] [Git.FilePath.toInternalGitPath subfile]
git_annex_expectoutput "find" v v
narrowing test case
2014-02-04 19:14:16 +00:00
add test case for merge conflict resolution between annexed and in-git file
2014-03-04 21:37:36 +00:00
{- Check merge conflict resolution when both repos start with an annexed
added test for conflict resolution with a file that is both deleted and modified
2014-03-04 21:12:26 +00:00
- file; one modifies it, and the other deletes it. -}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_remove_conflict_resolution :: Assertion
test_remove_conflict_resolution = do
added test for conflict resolution with a file that is both deleted and modified
2014-03-04 21:12:26 +00:00
check True
check False
where
init: Fix reversion in detection of repo made with git clone --shared
2015-09-09 17:56:37 +00:00
check inr1 = withtmpclonerepo $ \r1 ->
withtmpclonerepo $ \r2 -> do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir r1 $ do
added test for conflict resolution with a file that is both deleted and modified
2014-03-04 21:12:26 +00:00
disconnectOrigin
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor"
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
add_annex conflictor @? "add conflicter failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed in r1"
indir r2 $
added test for conflict resolution with a file that is both deleted and modified
2014-03-04 21:12:26 +00:00
disconnectOrigin
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
pair r1 r2
indir r2 $ do
git_annex "sync" [] @? "sync failed in r2"
git_annex "get" [conflictor]
added test for conflict resolution with a file that is both deleted and modified
2014-03-04 21:12:26 +00:00
@? "get conflictor failed"
unlessM (annexeval Config.isDirect) $ do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "unlock" [conflictor]
added test for conflict resolution with a file that is both deleted and modified
2014-03-04 21:12:26 +00:00
@? "unlock conflictor failed"
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "newconflictor"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir r1 $
added test for conflict resolution with a file that is both deleted and modified
2014-03-04 21:12:26 +00:00
nukeFile conflictor
let l = if inr1 then [r1, r2, r1] else [r2, r1, r2]
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
forM_ l $ \r -> indir r $
git_annex "sync" [] @? "sync failed"
added test for conflict resolution with a file that is both deleted and modified
2014-03-04 21:12:26 +00:00
checkmerge "r1" r1
checkmerge "r2" r2
conflictor = "conflictor"
variantprefix = conflictor ++ ".variant"
checkmerge what d = do
l <- getDirectoryContents d
let v = filter (variantprefix `isPrefixOf`) l
not (null v)
more conflict resolution tests
2014-03-05 03:12:15 +00:00
@? (what ++ " conflictor variant file missing in: " ++ show l )
length v == 1
@? (what ++ " too many variant files in: " ++ show v)
added test for conflict resolution with a file that is both deleted and modified
2014-03-04 21:12:26 +00:00
indent with tabs not spaces Found these with: git grep "^ " $(find -type f -name \*.hs) |grep -v ': where' Unfortunately there is some inline hamlet that cannot use tabs for indentation. Also, Assistant/WebApp/Bootstrap3.hs is a copy of a module and so I'm leaving it as-is.
2014-10-09 19:09:26 +00:00
{- Check merge confalict resolution when a file is annexed in one repo,
- and checked directly into git in the other repo.
-
- This test requires indirect mode to set it up, but tests both direct and
- indirect mode.
-}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_nonannexed_file_conflict_resolution :: Assertion
test_nonannexed_file_conflict_resolution = do
add test case for merge conflict resolution between annexed and in-git file
2014-03-04 21:37:36 +00:00
check True False
check False False
move check for if direct mode is supported to the right place Need to run that check inside a annex repo.
2015-12-29 21:48:25 +00:00
check True True
check False True
add test case for merge conflict resolution between annexed and in-git file
2014-03-04 21:37:36 +00:00
where
init: Fix reversion in detection of repo made with git clone --shared
2015-09-09 17:56:37 +00:00
check inr1 switchdirect = withtmpclonerepo $ \r1 ->
withtmpclonerepo $ \r2 ->
add test case for merge conflict resolution between annexed and in-git file
2014-03-04 21:37:36 +00:00
whenM (isInDirect r1 <&&> isInDirect r2) $ do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir r1 $ do
add test case for merge conflict resolution between annexed and in-git file
2014-03-04 21:37:36 +00:00
disconnectOrigin
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor"
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
add_annex conflictor @? "add conflicter failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed in r1"
indir r2 $ do
add test case for merge conflict resolution between annexed and in-git file
2014-03-04 21:37:36 +00:00
disconnectOrigin
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor nonannexed_content
test suite 100% pass in v6, finally! Set annex.largefiles when adding the conflicting non-annexed file, otherwise it would be added as an annexed file.
2015-12-30 19:12:27 +00:00
boolSystem "git"
[ Param "config"
, Param "annex.largefiles"
, Param ("exclude=" ++ ingitfile ++ " and exclude=" ++ conflictor)
] @? "git config annex.largefiles failed"
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "add", File conflictor] @? "git add conflictor failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed in r2"
pair r1 r2
add test case for merge conflict resolution between annexed and in-git file
2014-03-04 21:37:36 +00:00
let l = if inr1 then [r1, r2] else [r2, r1]
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
forM_ l $ \r -> indir r $ do
add test case for merge conflict resolution between annexed and in-git file
2014-03-04 21:37:36 +00:00
when switchdirect $
move check for if direct mode is supported to the right place Need to run that check inside a annex repo.
2015-12-29 21:48:25 +00:00
whenM (annexeval Annex.Version.versionSupportsDirectMode) $
git_annex "direct" [] @? "failed switching to direct mode"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed"
sync: Automatically resolve merge conflict between and annexed file and a regular git file. This is a new feature, it was not handled before, since it's a bit of an edge case. However, it can be handled exactly the same as a file/dir conflict, just leave the non-annexed item alone. While implementing this, the core resolveMerge' function got a lot simpler and clearer. Note especially that where before there was an asymetric call to stagefromdirectmergedir, now graftin is called symmetrically in both cases. And, in order to add that `graftin us`, the current branch needed to be known (if there is no current branch, there cannot be a merge conflict). This led to some cleanups of how autoMergeFrom behaved when there is no current branch. This commit was sponsored by Philippe Gauthier.
2014-03-04 21:45:11 +00:00
checkmerge ("r1" ++ show switchdirect) r1
checkmerge ("r2" ++ show switchdirect) r2
add test case for merge conflict resolution between annexed and in-git file
2014-03-04 21:37:36 +00:00
conflictor = "conflictor"
nonannexed_content = "nonannexed"
variantprefix = conflictor ++ ".variant"
checkmerge what d = do
l <- getDirectoryContents d
let v = filter (variantprefix `isPrefixOf`) l
not (null v)
@? (what ++ " conflictor variant file missing in: " ++ show l )
more conflict resolution tests
2014-03-05 03:12:15 +00:00
length v == 1
@? (what ++ " too many variant files in: " ++ show v)
add test case for merge conflict resolution between annexed and in-git file
2014-03-04 21:37:36 +00:00
conflictor `elem` l @? (what ++ " conflictor file missing in: " ++ show l)
sync: Automatically resolve merge conflict between and annexed file and a regular git file. This is a new feature, it was not handled before, since it's a bit of an edge case. However, it can be handled exactly the same as a file/dir conflict, just leave the non-annexed item alone. While implementing this, the core resolveMerge' function got a lot simpler and clearer. Note especially that where before there was an asymetric call to stagefromdirectmergedir, now graftin is called symmetrically in both cases. And, in order to add that `graftin us`, the current branch needed to be known (if there is no current branch, there cannot be a merge conflict). This led to some cleanups of how autoMergeFrom behaved when there is no current branch. This commit was sponsored by Philippe Gauthier.
2014-03-04 21:45:11 +00:00
s <- catchMaybeIO (readFile (d </> conflictor))
s == Just nonannexed_content
@? (what ++ " wrong content for nonannexed file: " ++ show s)
add test case for merge conflict resolution between annexed and in-git file
2014-03-04 21:37:36 +00:00
add regression test case for ba42b67c70715282fc38953e029f0a172fc00c5d better late than never
2014-07-10 17:34:53 +00:00
{- Check merge confalict resolution when a file is annexed in one repo,
- and is a non-git-annex symlink in the other repo.
-
- Test can only run when coreSymlinks is supported, because git needs to
- be able to check out the non-git-annex symlink.
-}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_nonannexed_symlink_conflict_resolution :: Assertion
test_nonannexed_symlink_conflict_resolution = do
add regression test case for ba42b67c70715282fc38953e029f0a172fc00c5d better late than never
2014-07-10 17:34:53 +00:00
check True False
check False False
move check for if direct mode is supported to the right place Need to run that check inside a annex repo.
2015-12-29 21:48:25 +00:00
check True True
check False True
add regression test case for ba42b67c70715282fc38953e029f0a172fc00c5d better late than never
2014-07-10 17:34:53 +00:00
where
init: Fix reversion in detection of repo made with git clone --shared
2015-09-09 17:56:37 +00:00
check inr1 switchdirect = withtmpclonerepo $ \r1 ->
withtmpclonerepo $ \r2 ->
add regression test case for ba42b67c70715282fc38953e029f0a172fc00c5d better late than never
2014-07-10 17:34:53 +00:00
whenM (checkRepo (Types.coreSymlinks <$> Annex.getGitConfig) r1
<&&> isInDirect r1 <&&> isInDirect r2) $ do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir r1 $ do
add regression test case for ba42b67c70715282fc38953e029f0a172fc00c5d better late than never
2014-07-10 17:34:53 +00:00
disconnectOrigin
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor"
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
add_annex conflictor @? "add conflicter failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed in r1"
indir r2 $ do
add regression test case for ba42b67c70715282fc38953e029f0a172fc00c5d better late than never
2014-07-10 17:34:53 +00:00
disconnectOrigin
createSymbolicLink symlinktarget "conflictor"
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "add", File conflictor] @? "git add conflictor failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed in r2"
pair r1 r2
add regression test case for ba42b67c70715282fc38953e029f0a172fc00c5d better late than never
2014-07-10 17:34:53 +00:00
let l = if inr1 then [r1, r2] else [r2, r1]
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
forM_ l $ \r -> indir r $ do
add regression test case for ba42b67c70715282fc38953e029f0a172fc00c5d better late than never
2014-07-10 17:34:53 +00:00
when switchdirect $
move check for if direct mode is supported to the right place Need to run that check inside a annex repo.
2015-12-29 21:48:25 +00:00
whenM (annexeval Annex.Version.versionSupportsDirectMode) $ do
git_annex "direct" [] @? "failed switching to direct mode"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed"
add regression test case for ba42b67c70715282fc38953e029f0a172fc00c5d better late than never
2014-07-10 17:34:53 +00:00
checkmerge ("r1" ++ show switchdirect) r1
checkmerge ("r2" ++ show switchdirect) r2
conflictor = "conflictor"
symlinktarget = "dummy-target"
variantprefix = conflictor ++ ".variant"
checkmerge what d = do
l <- getDirectoryContents d
let v = filter (variantprefix `isPrefixOf`) l
not (null v)
@? (what ++ " conflictor variant file missing in: " ++ show l )
length v == 1
@? (what ++ " too many variant files in: " ++ show v)
conflictor `elem` l @? (what ++ " conflictor file missing in: " ++ show l)
s <- catchMaybeIO (readSymbolicLink (d </> conflictor))
s == Just symlinktarget
@? (what ++ " wrong target for nonannexed symlink: " ++ show s)
add test cases for the new direct mode merge bug
2014-03-04 05:41:00 +00:00
{- Check merge conflict resolution when there is a local file,
- that is not staged or committed, that conflicts with what's being added
- from the remmote.
-
- Case 1: Remote adds file named conflictor; local has a file named
- conflictor.
-
- Case 2: Remote adds conflictor/file; local has a file named conflictor.
-}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_uncommitted_conflict_resolution :: Assertion
test_uncommitted_conflict_resolution = do
add test cases for the new direct mode merge bug
2014-03-04 05:41:00 +00:00
check conflictor
check (conflictor </> "file")
where
init: Fix reversion in detection of repo made with git clone --shared
2015-09-09 17:56:37 +00:00
check remoteconflictor = withtmpclonerepo $ \r1 ->
withtmpclonerepo $ \r2 -> do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir r1 $ do
add test cases for the new direct mode merge bug
2014-03-04 05:41:00 +00:00
disconnectOrigin
revert parentDir change Reverts 965e106f242b34a8269c72e136f60fc0c36ae460 Unfortunately, this caused breakage on Windows, and possibly elsewhere, because parentDir and takeDirectory do not behave the same when there is a trailing directory separator.
2015-01-09 17:11:56 +00:00
createDirectoryIfMissing True (parentDir remoteconflictor)
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent remoteconflictor annexedcontent
started working on testing v6 unlocked files Many failures.
2016-01-01 17:59:57 +00:00
add_annex conflictor @? "add remoteconflicter failed"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed in r1"
indir r2 $ do
add test cases for the new direct mode merge bug
2014-03-04 05:41:00 +00:00
disconnectOrigin
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor localcontent
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
pair r1 r2
indir r2 $ ifM (annexeval Config.isDirect)
add test cases for the new direct mode merge bug
2014-03-04 05:41:00 +00:00
( do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "sync" [] @? "sync failed"
add test cases for the new direct mode merge bug
2014-03-04 05:41:00 +00:00
let local = conflictor ++ localprefix
doesFileExist local @? (local ++ " missing after merge")
s <- readFile local
s == localcontent @? (local ++ " has wrong content: " ++ s)
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [conflictor] @? "get failed"
add test cases for the new direct mode merge bug
2014-03-04 05:41:00 +00:00
doesFileExist remoteconflictor @? (remoteconflictor ++ " missing after merge")
s' <- readFile remoteconflictor
s' == annexedcontent @? (remoteconflictor ++ " has wrong content: " ++ s)
-- this case is intentionally not handled
-- in indirect mode, since the user
-- can recover on their own easily
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
, git_annex_shouldfail "sync" [] @? "sync failed to fail"
add test cases for the new direct mode merge bug
2014-03-04 05:41:00 +00:00
)
conflictor = "conflictor"
localprefix = ".variant-local"
localcontent = "local"
annexedcontent = "annexed"
much improved test and real fix for FAT symlink loss on conflicted merge I think that 751f496c11b421f498d19d4ea939bfd84125d349 didn't quite manage to actually fix the bug, although I have not checked since its "fix" got redone. The test suite now actually checks the file staged in git is a symlink, rather than relying on the bug casing a later sync failure. This seems a more reliable way to detect it, and probably avoids a heisenbug in the test suite.
2014-03-04 18:31:26 +00:00
{- On Windows/FAT, repeated conflict resolution sometimes
- lost track of whether a file was a symlink.
add test case & fix conflict resolution bug on Windows & FAT Fix bug in automatic merge conflict resolution code when used on a filesystem not supporting symlinks, which resulted in it losing track of the symlink bit of annexed files. This was the underlying bug that was causing another test to fail, which got worked around in 1c997fd08c04b27c9acb561a33c6fb2829089be7. I've chosen to keep 2 separate test cases since the old test case only detected the problem accidentially. Test suite passes on FAT & in windows, as well as on proper unix systems. This commit was sponsored by Ellis Whitehead.
2014-02-04 19:31:53 +00:00
-}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_conflict_resolution_symlink_bit :: Assertion
fix numerous problem with test suite on crippled filesystems etc
2016-02-16 19:30:59 +00:00
test_conflict_resolution_symlink_bit = unlessM (unlockedFiles <$> getTestMode) $
init: Fix reversion in detection of repo made with git clone --shared
2015-09-09 17:56:37 +00:00
withtmpclonerepo $ \r1 ->
withtmpclonerepo $ \r2 ->
withtmpclonerepo $ \r3 -> do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir r1 $ do
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "add" [conflictor] @? "add conflicter failed"
git_annex "sync" [] @? "sync failed in r1"
much improved test and real fix for FAT symlink loss on conflicted merge I think that 751f496c11b421f498d19d4ea939bfd84125d349 didn't quite manage to actually fix the bug, although I have not checked since its "fix" got redone. The test suite now actually checks the file staged in git is a symlink, rather than relying on the bug casing a later sync failure. This seems a more reliable way to detect it, and probably avoids a heisenbug in the test suite.
2014-03-04 18:31:26 +00:00
check_is_link conflictor "r1"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir r2 $ do
much improved test and real fix for FAT symlink loss on conflicted merge I think that 751f496c11b421f498d19d4ea939bfd84125d349 didn't quite manage to actually fix the bug, although I have not checked since its "fix" got redone. The test suite now actually checks the file staged in git is a symlink, rather than relying on the bug casing a later sync failure. This seems a more reliable way to detect it, and probably avoids a heisenbug in the test suite.
2014-03-04 18:31:26 +00:00
createDirectory conflictor
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent (conflictor </> "subfile") "subfile"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "add" [conflictor] @? "add conflicter failed"
git_annex "sync" [] @? "sync failed in r2"
much improved test and real fix for FAT symlink loss on conflicted merge I think that 751f496c11b421f498d19d4ea939bfd84125d349 didn't quite manage to actually fix the bug, although I have not checked since its "fix" got redone. The test suite now actually checks the file staged in git is a symlink, rather than relying on the bug casing a later sync failure. This seems a more reliable way to detect it, and probably avoids a heisenbug in the test suite.
2014-03-04 18:31:26 +00:00
check_is_link (conflictor </> "subfile") "r2"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
indir r3 $ do
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "add" [conflictor] @? "add conflicter failed"
git_annex "sync" [] @? "sync failed in r1"
much improved test and real fix for FAT symlink loss on conflicted merge I think that 751f496c11b421f498d19d4ea939bfd84125d349 didn't quite manage to actually fix the bug, although I have not checked since its "fix" got redone. The test suite now actually checks the file staged in git is a symlink, rather than relying on the bug casing a later sync failure. This seems a more reliable way to detect it, and probably avoids a heisenbug in the test suite.
2014-03-04 18:31:26 +00:00
check_is_link (conflictor </> "subfile") "r3"
where
narrowing test case
2014-02-04 19:14:16 +00:00
conflictor = "conflictor"
much improved test and real fix for FAT symlink loss on conflicted merge I think that 751f496c11b421f498d19d4ea939bfd84125d349 didn't quite manage to actually fix the bug, although I have not checked since its "fix" got redone. The test suite now actually checks the file staged in git is a symlink, rather than relying on the bug casing a later sync failure. This seems a more reliable way to detect it, and probably avoids a heisenbug in the test suite.
2014-03-04 18:31:26 +00:00
check_is_link f what = do
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex_expectoutput "find" ["--include=*", f] [Git.FilePath.toInternalGitPath f]
much improved test and real fix for FAT symlink loss on conflicted merge I think that 751f496c11b421f498d19d4ea939bfd84125d349 didn't quite manage to actually fix the bug, although I have not checked since its "fix" got redone. The test suite now actually checks the file staged in git is a symlink, rather than relying on the bug casing a later sync failure. This seems a more reliable way to detect it, and probably avoids a heisenbug in the test suite.
2014-03-04 18:31:26 +00:00
l <- annexeval $ Annex.inRepo $ Git.LsTree.lsTreeFiles Git.Ref.headRef [f]
rename BlobType and add submodule to it This was badly named, it's a not a blob necessarily, but anything that a tree can refer to. Also removed the Show instance which was used for serialization to git format, instead use fmtTreeItemType. This commit was supported by the NSF-funded DataLad project.
2018-05-14 18:22:44 +00:00
all (\i -> Git.Types.toTreeItemType (Git.LsTree.mode i) == Just Git.Types.TreeSymlink) l
much improved test and real fix for FAT symlink loss on conflicted merge I think that 751f496c11b421f498d19d4ea939bfd84125d349 didn't quite manage to actually fix the bug, although I have not checked since its "fix" got redone. The test suite now actually checks the file staged in git is a symlink, rather than relying on the bug casing a later sync failure. This seems a more reliable way to detect it, and probably avoids a heisenbug in the test suite.
2014-03-04 18:31:26 +00:00
@? (what ++ " " ++ f ++ " lost symlink bit after merge: " ++ show l)
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
v7 Install new git hooks in this version. This does beg the question of what to do if git later gets eg a post-smudge hook, that could run git-annex smudge --update. I think the thing to do in that case would be to make git-annex smudge --update install the new hooks. That way, as the user uses git-annex, the hook would be created pretty quickly and without needing any extra syscalls except for when git-annex smudge --update is called. I considered doing something like that for installation of the post-checkout and post-merge hooks, which would have avoided the need for v7. But the only place it was cheap to do it would be in git-annex smudge which could cheaply notice that smudge.log didn't exist yet and so know the hooks needed to be installed. But since smudge used to populate pointer files, it would be quite surprising if a single git checkout/merge failed to update the work tree, and so that idea didn't work out. The other reason for v7 is psychological -- users don't need to worry about whether they might be running an old version of git-annex that doesn't support their v7 repository very well. And bug reports about "v6" have gotten a bit of a bad association in my head since they often hit one of the known limitations and didn't realize it was experimental. newtyped RepoVersion Int to avoid needing 2 comparisons in versionSupportsUnlockedPointers etc. Also it's just nicer. This commit was sponsored by John Pellman on Patreon.
2018-10-25 21:23:53 +00:00
{- A v7 unlocked file that conflicts with a locked file should be resolved
add test: conflict resolution (mixed locked and unlocked file)
2015-12-30 20:36:39 +00:00
- in favor of the unlocked file, with no variant files, as long as they
- both point to the same key. -}
test_mixed_lock_conflict_resolution :: Assertion
test_mixed_lock_conflict_resolution =
withtmpclonerepo $ \r1 ->
withtmpclonerepo $ \r2 -> do
indir r1 $ whenM shouldtest $ do
disconnectOrigin
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor"
add test: conflict resolution (mixed locked and unlocked file)
2015-12-30 20:36:39 +00:00
git_annex "add" [conflictor] @? "add conflicter failed"
git_annex "sync" [] @? "sync failed in r1"
indir r2 $ whenM shouldtest $ do
disconnectOrigin
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor"
add test: conflict resolution (mixed locked and unlocked file)
2015-12-30 20:36:39 +00:00
git_annex "add" [conflictor] @? "add conflicter failed"
git_annex "unlock" [conflictor] @? "unlock conflicter failed"
git_annex "sync" [] @? "sync failed in r2"
pair r1 r2
forM_ [r1,r2,r1] $ \r -> indir r $
git_annex "sync" [] @? "sync failed"
checkmerge "r1" r1
checkmerge "r2" r2
where
shouldtest = annexeval Annex.Version.versionSupportsUnlockedPointers
conflictor = "conflictor"
variantprefix = conflictor ++ ".variant"
checkmerge what d = indir d $ whenM shouldtest $ do
l <- getDirectoryContents "."
let v = filter (variantprefix `isPrefixOf`) l
length v == 0
@? (what ++ " not exactly 0 variant files in: " ++ show l)
conflictor `elem` l @? ("conflictor not present after conflict resolution")
git_annex "get" [conflictor] @? "get failed"
git_annex_expectoutput "find" [conflictor] [conflictor]
-- regular file because it's unlocked
checkregularfile conflictor
add test case for http://git-annex.branchable.com/bugs/Assistant_keeps_deleting_all_the_files_in_my_repo/
2016-06-13 16:56:31 +00:00
{- Regression test for a bad merge between two adjusted branch repos,
- where the same file is added to both independently. The bad merge
- emptied the whole tree. -}
test_adjusted_branch_merge_regression :: Assertion
Fix test suite failure when git-annex test is not run inside a git repository Not the first time this kind of test suite breakage has happened.. It would be good to avoid somehow it looking up from .t and finding a git repo. But just running the test suite from time to time outside of git-annex would also let me notice these before the distribution packagers do. This commit was sponsored by mo on Patreon.
2018-11-05 17:27:16 +00:00
test_adjusted_branch_merge_regression = do
add test case for http://git-annex.branchable.com/bugs/Assistant_keeps_deleting_all_the_files_in_my_repo/
2016-06-13 16:56:31 +00:00
withtmpclonerepo $ \r1 ->
Fix test suite failure when git-annex test is not run inside a git repository Not the first time this kind of test suite breakage has happened.. It would be good to avoid somehow it looking up from .t and finding a git repo. But just running the test suite from time to time outside of git-annex would also let me notice these before the distribution packagers do. This commit was sponsored by mo on Patreon.
2018-11-05 17:27:16 +00:00
withtmpclonerepo $ \r2 -> whenM (adjustedbranchsupported r1) $ do
add test case for http://git-annex.branchable.com/bugs/Assistant_keeps_deleting_all_the_files_in_my_repo/
2016-06-13 16:56:31 +00:00
pair r1 r2
setup r1
setup r2
checkmerge "r1" r1
checkmerge "r2" r2
where
conflictor = "conflictor"
avoid a FAT fail
2018-12-19 17:36:14 +00:00
setup r = indir r $ whensupported $ do
add test case for http://git-annex.branchable.com/bugs/Assistant_keeps_deleting_all_the_files_in_my_repo/
2016-06-13 16:56:31 +00:00
disconnectOrigin
git_annex "upgrade" [] @? "upgrade failed"
git_annex "adjust" ["--unlock", "--force"] @? "adjust failed"
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent conflictor "conflictor"
add test case for http://git-annex.branchable.com/bugs/Assistant_keeps_deleting_all_the_files_in_my_repo/
2016-06-13 16:56:31 +00:00
git_annex "add" [conflictor] @? "add conflicter failed"
git_annex "sync" [] @? "sync failed"
avoid a FAT fail
2018-12-19 17:36:14 +00:00
checkmerge what d = indir d $ whensupported $ do
add test case for http://git-annex.branchable.com/bugs/Assistant_keeps_deleting_all_the_files_in_my_repo/
2016-06-13 16:56:31 +00:00
git_annex "sync" [] @? ("sync failed in " ++ what)
l <- getDirectoryContents "."
conflictor `elem` l
@? ("conflictor not present after merge in " ++ what)
avoid a FAT fail
2018-12-19 17:36:14 +00:00
-- Currently this fails on FAT, for unknown reasons not to
-- do with what it's intended to test.
whensupported = unlessM (annexeval Config.crippledFileSystem)
add test case for http://git-annex.branchable.com/bugs/Assistant_keeps_deleting_all_the_files_in_my_repo/
2016-06-13 16:56:31 +00:00
add test case for sync_in_adjusted_branch_deleted_recently_added_files This commit was sponsored by Denis Dzyubenko on Patreon.
2016-10-11 18:22:49 +00:00
{- Regression test for a bug in adjusted branch syncing code, where adding
- a subtree to an existing tree lost files. -}
test_adjusted_branch_subtree_regression :: Assertion
test_adjusted_branch_subtree_regression =
Fix test suite failure when git-annex test is not run inside a git repository Not the first time this kind of test suite breakage has happened.. It would be good to avoid somehow it looking up from .t and finding a git repo. But just running the test suite from time to time outside of git-annex would also let me notice these before the distribution packagers do. This commit was sponsored by mo on Patreon.
2018-11-05 17:27:16 +00:00
withtmpclonerepo $ \r -> whenM (adjustedbranchsupported r) $ do
indir r $ do
disconnectOrigin
git_annex "upgrade" [] @? "upgrade failed"
git_annex "adjust" ["--unlock", "--force"] @? "adjust failed"
createDirectoryIfMissing True "a/b/c"
writecontent "a/b/c/d" "foo"
git_annex "add" ["a/b/c"] @? "add a/b/c failed"
git_annex "sync" [] @? "sync failed"
createDirectoryIfMissing True "a/b/x"
writecontent "a/b/x/y" "foo"
git_annex "add" ["a/b/x"] @? "add a/b/x failed"
git_annex "sync" [] @? "sync failed"
boolSystem "git" [Param "checkout", Param "master"] @? "git checkout master failed"
doesFileExist "a/b/x/y" @? ("a/b/x/y missing from master after adjusted branch sync")
add test case for sync_in_adjusted_branch_deleted_recently_added_files This commit was sponsored by Denis Dzyubenko on Patreon.
2016-10-11 18:22:49 +00:00
fix test suite to pass on windows On windows, the sync of the second cloned repo to origin failed, because synced/master was a non-fast-forward. This may be a bug of its own, but it's not the issue that this test was intended to test, so disconnect the repos from origin before syncing.
2014-02-03 16:03:10 +00:00
{- Set up repos as remotes of each other. -}
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
pair :: FilePath -> FilePath -> Assertion
pair r1 r2 = forM_ [r1, r2] $ \r -> indir r $ do
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
when (r /= r1) $
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "remote", Param "add", Param "r1", File ("../../" ++ r1)] @? "remote add"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
when (r /= r2) $
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "remote", Param "add", Param "r2", File ("../../" ++ r2)] @? "remote add"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_map :: Assertion
test_map = intmpclonerepo $ do
map: --fast disables use of dot to display map Generally useful, and allows the test suite to test it.
2011-12-20 20:31:59 +00:00
-- set descriptions, that will be looked for in the map
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "describe" [".", "this repo"] @? "describe 1 failed"
git_annex "describe" ["origin", "origin repo"] @? "describe 2 failed"
map: --fast disables use of dot to display map Generally useful, and allows the test suite to test it.
2011-12-20 20:31:59 +00:00
-- --fast avoids it running graphviz, not a build dependency
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "map" ["--fast"] @? "map failed"
Test suite improvements. Current top-level test coverage: 68% Been higher before, but a lot of new code has been added.
2011-12-20 21:31:25 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_uninit :: Assertion
test_uninit = intmpclonerepo $ do
git_annex "get" [] @? "get failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
_ <- git_annex "uninit" [] -- exit status not checked; does abnormal exit
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
checkregularfile annexedfile
doesDirectoryExist ".git" @? ".git vanished in uninit"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_uninit_inbranch :: Assertion
test_uninit_inbranch = intmpclonerepoInDirect $ do
remove Params constructor from Utility.SafeCommand This removes a bit of complexity, and should make things faster (avoids tokenizing Params string), and probably involve less garbage collection. In a few places, it was useful to use Params to avoid needing a list, but that is easily avoided. Problems noticed while doing this conversion: * Some uses of Params "oneword" which was entirely unnecessary overhead. * A few places that built up a list of parameters with ++ and then used Params to split it! Test suite passes.
2015-06-01 17:52:23 +00:00
boolSystem "git" [Param "checkout", Param "git-annex"] @? "git checkout git-annex"
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "uninit" [] @? "uninit failed to fail when git-annex branch was checked out"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_upgrade :: Assertion
test_upgrade = intmpclonerepo $
add simple test for conflict resolution in adjusted branch This is not really extensive enough, but a start..
2016-04-12 17:01:31 +00:00
git_annex "upgrade" [] @? "upgrade failed"
Test suite improvements. Current top-level test coverage: 68% Been higher before, but a lot of new code has been added.
2011-12-20 21:31:25 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_whereis :: Assertion
test_whereis = intmpclonerepo $ do
Test suite improvements. Current top-level test coverage: 68% Been higher before, but a lot of new code has been added.
2011-12-20 21:31:25 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "whereis" [annexedfile] @? "whereis on non-present file failed"
git_annex "untrust" ["origin"] @? "untrust failed"
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "whereis" [annexedfile] @? "whereis on non-present file only present in untrusted repo failed to fail"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [annexedfile] @? "get failed"
Test suite improvements. Current top-level test coverage: 68% Been higher before, but a lot of new code has been added.
2011-12-20 21:31:25 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "whereis" [annexedfile] @? "whereis on present file failed"
map: --fast disables use of dot to display map Generally useful, and allows the test suite to test it.
2011-12-20 20:31:59 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_hook_remote :: Assertion
test_hook_remote = intmpclonerepo $ do
missed some __WINDOWS__ defines
2013-08-04 17:07:55 +00:00
#ifndef mingw32_HOST_OS
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "initremote" (words "foo type=hook encryption=none hooktype=foo") @? "initremote failed"
added a test of the hook special remote
2011-12-20 19:13:36 +00:00
createDirectory dir
git_config "annex.foo-store-hook" $
"cp $ANNEX_FILE " ++ loc
git_config "annex.foo-retrieve-hook" $
"cp " ++ loc ++ " $ANNEX_FILE"
git_config "annex.foo-remove-hook" $
"rm -f " ++ loc
git_config "annex.foo-checkpresent-hook" $
"if [ -e " ++ loc ++ " ]; then echo $ANNEX_KEY; fi"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [annexedfile] @? "get of file failed"
added a test of the hook special remote
2011-12-20 19:13:36 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "copy" [annexedfile, "--to", "foo"] @? "copy --to hook remote failed"
added a test of the hook special remote
2011-12-20 19:13:36 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "drop" [annexedfile, "--numcopies=2"] @? "drop failed"
added a test of the hook special remote
2011-12-20 19:13:36 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "move" [annexedfile, "--from", "foo"] @? "move --from hook remote failed"
added a test of the hook special remote
2011-12-20 19:13:36 +00:00
annexed_present annexedfile
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "drop" [annexedfile, "--numcopies=2"] @? "drop failed to fail"
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
annexed_present annexedfile
where indenting
2012-11-11 04:51:07 +00:00
where
dir = "dir"
loc = dir ++ "/$ANNEX_KEY"
git_config k v = boolSystem "git" [Param "config", Param k, Param v]
@? "git config failed"
Windows: The test suite now passes on Windows (a few broken parts are disabled). Better to have a working test suite that doesn't test a few things than no working test suite. Most of the disabled stuff is because for some reason "git annex sync" doesn't work when run inside the test suite. Looks like PATH problems. The directory and rsync special remotes seem broken on Windows, or maybe the tests are. Pretty sure the hook special remote test is broken.
2013-06-18 20:12:43 +00:00
#else
-- this test doesn't work in Windows TODO
noop
#endif
added a test of the hook special remote
2011-12-20 19:13:36 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_directory_remote :: Assertion
test_directory_remote = intmpclonerepo $ do
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
createDirectory "dir"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "initremote" (words "foo type=directory encryption=none directory=dir") @? "initremote failed"
git_annex "get" [annexedfile] @? "get of file failed"
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "copy" [annexedfile, "--to", "foo"] @? "copy --to directory remote failed"
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "drop" [annexedfile, "--numcopies=2"] @? "drop failed"
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "move" [annexedfile, "--from", "foo"] @? "move --from directory remote failed"
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
annexed_present annexedfile
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "drop" [annexedfile, "--numcopies=2"] @? "drop failed to fail"
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_rsync_remote :: Assertion
test_rsync_remote = intmpclonerepo $ do
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
createDirectory "dir"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "initremote" (words "foo type=rsync encryption=none rsyncurl=dir") @? "initremote failed"
git_annex "get" [annexedfile] @? "get of file failed"
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "copy" [annexedfile, "--to", "foo"] @? "copy --to rsync remote failed"
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "drop" [annexedfile, "--numcopies=2"] @? "drop failed"
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "move" [annexedfile, "--from", "foo"] @? "move --from rsync remote failed"
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
annexed_present annexedfile
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "drop" [annexedfile, "--numcopies=2"] @? "drop failed to fail"
add tests of directory and rsync special remotes
2011-12-20 19:17:56 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_bup_remote :: Assertion
fold Build/SysConfig.hs into BuildInfo via include This avoids warnings from stack about the module not being listed in the cabal file. So, the generated file is also renamed to Build/SysConfig. Note that the setup program seems to be cached despite these changes; I had to cabal clean to get cabal to update it so that Build/SysConfig was written. This commit was sponsored by Jochen Bartl on Patreon.
2017-12-14 16:46:57 +00:00
test_bup_remote = intmpclonerepo $ when BuildInfo.bup $ do
test bup special remote (74% coverage)
2011-12-21 17:50:33 +00:00
dir <- absPath "dir" -- bup special remote needs an absolute path
createDirectory dir
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "initremote" (words $ "foo type=bup encryption=none buprepo="++dir) @? "initremote failed"
git_annex "get" [annexedfile] @? "get of file failed"
test bup special remote (74% coverage)
2011-12-21 17:50:33 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "copy" [annexedfile, "--to", "foo"] @? "copy --to bup remote failed"
test bup special remote (74% coverage)
2011-12-21 17:50:33 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "drop" [annexedfile, "--numcopies=2"] @? "drop failed"
test bup special remote (74% coverage)
2011-12-21 17:50:33 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "copy" [annexedfile, "--from", "foo"] @? "copy --from bup remote failed"
test bup special remote (74% coverage)
2011-12-21 17:50:33 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "move" [annexedfile, "--from", "foo"] @? "move --from bup remote failed"
test bup special remote (74% coverage)
2011-12-21 17:50:33 +00:00
annexed_present annexedfile
improve
2011-12-21 18:10:36 +00:00
-- gpg is not a build dependency, so only test when it's available
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_crypto :: Assertion
missed some __WINDOWS__ defines
2013-08-04 17:07:55 +00:00
#ifndef mingw32_HOST_OS
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_crypto = do
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
testscheme "shared"
testscheme "hybrid"
testscheme "pubkey"
where
support gpg.program When gpg.program is configured, it's used to get the command to run for gpg. Useful on systems that have only a gpg2 command or want to use it instead of the gpg command.
2015-09-09 22:06:49 +00:00
gpgcmd = Utility.Gpg.mkGpgCmd Nothing
testscheme scheme = intmpclonerepo $ whenM (Utility.Path.inPath (Utility.Gpg.unGpgCmd gpgcmd)) $ do
Make test suite work better when the temp directory is on NFS. Deleting directories is one of the great unsolved problems of CS, thanks to abominations like NFS lock files and Windows and races with other processes cleaning up after themselves in the background. The gpg test harness sometimes failed to delete its temp directory on NFS. Avoid the problem class by not deleting it at all, and putting it inside the tmp repo being tested. The test suite's more robust (and/or nonsensical) workarounds for deleting its test dir will thus be used, hopefully avoiding the problem until an OS finds a new way to violate POSIX and the laws of nature. Note that this means that the .gnupg directory will be on whatever filesystem the test suite is being run on, which may be a lesser quality filesystem than gpg is really expecting. Gpg does not seem to need to write sockets etc to there so this seems ok. The only known problem is that if the filesystem forces a directory mode like 777, gpg will warn about unsafe home directory perms, but it still works.
2018-12-19 15:56:39 +00:00
gpgtmpdir <- annexeval $ (</> "gpgtest")
<$> Annex.fromRepo Annex.Locations.gitAnnexTmpMiscDir
annexeval $ Annex.Perms.createAnnexDirectory gpgtmpdir
Utility.Gpg.testTestHarness gpgtmpdir gpgcmd
support gpg.program When gpg.program is configured, it's used to get the command to run for gpg. Useful on systems that have only a gpg2 command or want to use it instead of the gpg command.
2015-09-09 22:06:49 +00:00
@? "test harness self-test failed"
Make test suite work better when the temp directory is on NFS. Deleting directories is one of the great unsolved problems of CS, thanks to abominations like NFS lock files and Windows and races with other processes cleaning up after themselves in the background. The gpg test harness sometimes failed to delete its temp directory on NFS. Avoid the problem class by not deleting it at all, and putting it inside the tmp repo being tested. The test suite's more robust (and/or nonsensical) workarounds for deleting its test dir will thus be used, hopefully avoiding the problem until an OS finds a new way to violate POSIX and the laws of nature. Note that this means that the .gnupg directory will be on whatever filesystem the test suite is being run on, which may be a lesser quality filesystem than gpg is really expecting. Gpg does not seem to need to write sockets etc to there so this seems ok. The only known problem is that if the filesystem forces a directory mode like 777, gpg will warn about unsafe home directory perms, but it still works.
2018-12-19 15:56:39 +00:00
Utility.Gpg.testHarness gpgtmpdir gpgcmd $ do
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
createDirectory "dir"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
let a cmd = git_annex cmd $
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
[ "foo"
, "type=directory"
, "encryption=" ++ scheme
, "directory=dir"
, "highRandomQuality=false"
] ++ if scheme `elem` ["hybrid","pubkey"]
then ["keyid=" ++ Utility.Gpg.testKeyId]
else []
a "initremote" @? "initremote failed"
not <$> a "initremote" @? "initremote failed to fail when run twice in a row"
a "enableremote" @? "enableremote failed"
a "enableremote" @? "enableremote failed when run twice in a row"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "get" [annexedfile] @? "get of file failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "copy" [annexedfile, "--to", "foo"] @? "copy --to encrypted remote failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
(c,k) <- annexeval $ do
uuid <- Remote.nameToUUID "foo"
rs <- Logs.Remote.readRemoteLog
reorg
2015-12-15 19:34:28 +00:00
Just k <- Annex.WorkTree.lookupFile annexedfile
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
return (fromJust $ M.lookup uuid rs, k)
let key = if scheme `elem` ["hybrid","pubkey"]
then Just $ Utility.Gpg.KeyIds [Utility.Gpg.testKeyId]
else Nothing
testEncryptedRemote scheme key c [k] @? "invalid crypto setup"
annexed_present annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "drop" [annexedfile, "--numcopies=2"] @? "drop failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_notpresent annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "move" [annexedfile, "--from", "foo"] @? "move --from encrypted remote failed"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_present annexedfile
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "drop" [annexedfile, "--numcopies=2"] @? "drop failed to fail"
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
annexed_present annexedfile
{- Ensure the configuration complies with the encryption scheme, and
- that all keys are encrypted properly for the given directory remote. -}
testEncryptedRemote scheme ks c keys = case Remote.Helper.Encryptable.extractCipher c of
Just cip@Crypto.SharedCipher{} | scheme == "shared" && isNothing ks ->
checkKeys cip Nothing
Just cip@(Crypto.EncryptedCipher encipher v ks')
| checkScheme v && keysMatch ks' ->
checkKeys cip (Just v) <&&> checkCipher encipher ks'
_ -> return False
where
keysMatch (Utility.Gpg.KeyIds ks') =
maybe False (\(Utility.Gpg.KeyIds ks2) ->
sort (nub ks2) == sort (nub ks')) ks
support gpg.program When gpg.program is configured, it's used to get the command to run for gpg. Useful on systems that have only a gpg2 command or want to use it instead of the gpg command.
2015-09-09 22:06:49 +00:00
checkCipher encipher = Utility.Gpg.checkEncryptionStream gpgcmd encipher . Just
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
checkScheme Types.Crypto.Hybrid = scheme == "hybrid"
checkScheme Types.Crypto.PubKey = scheme == "pubkey"
checkKeys cip mvariant = do
add annex-ignore-command and annex-sync-command configs Added remote configuration settings annex-ignore-command and annex-sync-command, which are dynamic equivilants of the annex-ignore and annex-sync configurations. For this I needed a new DynamicConfig infrastructure. Its implementation should be as fast as before when there is no dynamic config, and it caches so shell commands are only run once. Note that annex-ignore-command exits nonzero when the remote should be ignored. While that may seem backwards, it allows using the same command for it as for annex-sync-command when you want to disable both. This commit was sponsored by Trenton Cronholm on Patreon.
2017-08-17 16:26:14 +00:00
dummycfg <- Types.GitConfig.dummyRemoteGitConfig
let encparams = (mempty :: Types.Remote.RemoteConfig, dummycfg)
plumb RemoteGitConfig through to decryptCipher
2016-05-23 21:27:15 +00:00
cipher <- Crypto.decryptCipher gpgcmd encparams cip
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
files <- filterM doesFileExist $
map ("dir" </>) $ concatMap (key2files cipher) keys
return (not $ null files) <&&> allM (checkFile mvariant) files
checkFile mvariant filename =
support gpg.program When gpg.program is configured, it's used to get the command to run for gpg. Useful on systems that have only a gpg2 command or want to use it instead of the gpg command.
2015-09-09 22:06:49 +00:00
Utility.Gpg.checkEncryptionFile gpgcmd filename $
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
if mvariant == Just Types.Crypto.PubKey then ks else Nothing
remove 163 lines of code without changing anything except imports
2016-01-20 20:36:33 +00:00
key2files cipher = Annex.Locations.keyPaths .
finished converting test suite for tasty Also merged in Test.hs from master.
2013-11-14 21:03:49 +00:00
Crypto.encryptKey Types.Crypto.HmacSha1 cipher
can't test gpg on Windows yet
2013-05-17 20:49:34 +00:00
#else
fix windows build
2014-12-30 21:02:51 +00:00
test_crypto = putStrLn "gpg testing not implemented on Windows"
can't test gpg on Windows yet
2013-05-17 20:49:34 +00:00
#endif
test encrypted special remote This involved adding a test harness to run gpg with a dummy key, and lots of fun.
2011-12-21 03:20:36 +00:00
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
test_add_subdirs :: Assertion
test_add_subdirs = intmpclonerepo $ do
add regression test for symlink calculation Note: Test reordered because running git-annex sync early broke the environment for some other tests.
2014-02-07 00:33:30 +00:00
createDirectory "dir"
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent ("dir" </> "foo") $ "dir/" ++ content annexedfile
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "add" ["dir"] @? "add of subdir failed"
add regression test for symlink calculation Note: Test reordered because running git-annex sync early broke the environment for some other tests.
2014-02-07 00:33:30 +00:00
{- Regression test for Windows bug where symlinks were not
- calculated correctly for files in subdirs. -}
fix numerous problem with test suite on crippled filesystems etc
2016-02-16 19:30:59 +00:00
unlessM (unlockedFiles <$> getTestMode) $ do
git_annex "sync" [] @? "sync failed"
strict bytestring encoders and decoders Only had lazy ones before. Already sped up a few parts of the code.
2019-01-01 18:54:06 +00:00
l <- annexeval $ Utility.FileSystemEncoding.decodeBL
Always use filesystem encoding for all file and handle reads and writes. This is a big scary change. I have convinced myself it should be safe. I hope!
2016-12-24 18:46:31 +00:00
<$> Annex.CatFile.catObject (Git.Types.Ref "HEAD:dir/foo")
fix numerous problem with test suite on crippled filesystems etc
2016-02-16 19:30:59 +00:00
"../.git/annex/" `isPrefixOf` l @? ("symlink from subdir to .git/annex is wrong: " ++ l)
add regression test for symlink calculation Note: Test reordered because running git-annex sync early broke the environment for some other tests.
2014-02-07 00:33:30 +00:00
createDirectory "dir2"
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent ("dir2" </> "foo") $ content annexedfile
add regression test for symlink calculation Note: Test reordered because running git-annex sync early broke the environment for some other tests.
2014-02-07 00:33:30 +00:00
setCurrentDirectory "dir"
remove testenv threading from test suite Simplify the test suite now that I have a way to setEnv on Windows by not threading an environement through it. Incidentially this fixed a failure on Windows, observed when running the test suite in cygwin terminal, where there is apparently an env var named "", and it tried to set that when propigating the environment, which failed. But mostly, it makes the code a lot less horrible. This commit was sponsored by Rémi Vanicat.
2014-12-30 20:58:02 +00:00
git_annex "add" [".." </> "dir2"] @? "add of ../subdir failed"
add regression test for symlink calculation Note: Test reordered because running git-annex sync early broke the environment for some other tests.
2014-02-07 00:33:30 +00:00
test case for addurl file://
2015-02-10 19:20:07 +00:00
test_addurl :: Assertion
test_addurl = intmpclonerepo $ do
-- file:// only; this test suite should not hit the network
limit url downloads to whitelisted schemes Security fix! Allowing any schemes, particularly file: and possibly others like scp: allowed file exfiltration by anyone who had write access to the git repository, since they could add an annexed file using such an url, or using an url that redirected to such an url, and wait for the victim to get it into their repository and send them a copy. * Added annex.security.allowed-url-schemes setting, which defaults to only allowing http and https URLs. Note especially that file:/ is no longer enabled by default. * Removed annex.web-download-command, since its interface does not allow supporting annex.security.allowed-url-schemes across redirects. If you used this setting, you may want to instead use annex.web-options to pass options to curl. With annex.web-download-command removed, nearly all url accesses in git-annex are made via Utility.Url via http-client or curl. http-client only supports http and https, so no problem there. (Disabling one and not the other is not implemented.) Used curl --proto to limit the allowed url schemes. Note that this will cause git annex fsck --from web to mark files using a disallowed url scheme as not being present in the web. That seems acceptable; fsck --from web also does that when a web server is not available. youtube-dl already disabled file: itself (probably for similar reasons). The scheme check was also added to youtube-dl urls for completeness, although that check won't catch any redirects it might follow. But youtube-dl goes off and does its own thing with other protocols anyway, so that's fine. Special remotes that support other domain-specific url schemes are not affected by this change. In the bittorrent remote, aria2c can still download magnet: links. The download of the .torrent file is otherwise now limited by annex.security.allowed-url-schemes. This does not address any external special remotes that might download an url themselves. Current thinking is all external special remotes will need to be audited for this problem, although many of them will use http libraries that only support http and not curl's menagarie. The related problem of accessing private localhost and LAN urls is not addressed by this commit. This commit was sponsored by Brett Eisenberg on Patreon.
2018-06-15 20:52:24 +00:00
let filecmd c ps = git_annex c ("-cannex.security.allowed-url-schemes=file" : ps)
test case for addurl file://
2015-02-10 19:20:07 +00:00
f <- absPath "myurl"
dropdrive drops the leading slash, so need fil:///
2015-02-11 16:22:43 +00:00
let url = replace "\\" "/" ("file:///" ++ dropDrive f)
add small delay to avoid problems on systems with low-resolution mtime I've seen intermittent failures of the test suite with v6 for a long time, it seems to have possibly gotten worse with the changes around v7. Or just being unlucky; all tests failed today. Seen on amd64 and i386 builders, repeatedly but intermittently: unused: FAIL (4.86s) Test.hs:928: git diff did not show changes to unlocked file And I think other such failures, all involving v7/v6 mode tests. I managed to reproduce the unused failure with --keep-failures, and inside the repo, git diff was indeed not showing any changes for the modified unlocked file. The two stats will be the same other than mtime; the old and new files have the same size and inode, since the test case writes to the file and then overwrites it. Indeed, notice the identical timestamps: builder@orca:~/gitbuilder/build/.t/tmprepo335$ echo 1 > foo; stat foo; echo 2 > foo; stat foo File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.894942036 +0000 Change: 2018-10-29 22:14:10.894942036 +0000 Birth: - File: foo Size: 2 Blocks: 8 IO Block: 4096 regular file Device: 801h/2049d Inode: 3546179 Links: 1 Access: (0644/-rw-r--r--) Uid: ( 1000/ builder) Gid: ( 1000/ builder) Access: 2018-10-29 22:14:10.894942036 +0000 Modify: 2018-10-29 22:14:10.898942036 +0000 Change: 2018-10-29 22:14:10.898942036 +0000 Birth: - I'm seeing this in Linux VMs; it doesn't happen on my laptop. I've also not experienced the intermittent test suite failures on my laptop. So, I hope that this small delay will avoid the problem. Update: I didn't, indeed I then reproduced the same failure on my laptop, so it must be due to something else. But keeping this change anyway since not needing to worry about lowish-resolution mtime in the test suite seems worthwhile.
2018-10-29 22:42:20 +00:00
writecontent f "foo"
test: display error messages from git-annex on unexpected failures .. but not on expected failures
2018-10-30 14:49:39 +00:00
git_annex_shouldfail "addurl" [url] @? "addurl failed to fail on file url"
limit url downloads to whitelisted schemes Security fix! Allowing any schemes, particularly file: and possibly others like scp: allowed file exfiltration by anyone who had write access to the git repository, since they could add an annexed file using such an url, or using an url that redirected to such an url, and wait for the victim to get it into their repository and send them a copy. * Added annex.security.allowed-url-schemes setting, which defaults to only allowing http and https URLs. Note especially that file:/ is no longer enabled by default. * Removed annex.web-download-command, since its interface does not allow supporting annex.security.allowed-url-schemes across redirects. If you used this setting, you may want to instead use annex.web-options to pass options to curl. With annex.web-download-command removed, nearly all url accesses in git-annex are made via Utility.Url via http-client or curl. http-client only supports http and https, so no problem there. (Disabling one and not the other is not implemented.) Used curl --proto to limit the allowed url schemes. Note that this will cause git annex fsck --from web to mark files using a disallowed url scheme as not being present in the web. That seems acceptable; fsck --from web also does that when a web server is not available. youtube-dl already disabled file: itself (probably for similar reasons). The scheme check was also added to youtube-dl urls for completeness, although that check won't catch any redirects it might follow. But youtube-dl goes off and does its own thing with other protocols anyway, so that's fine. Special remotes that support other domain-specific url schemes are not affected by this change. In the bittorrent remote, aria2c can still download magnet: links. The download of the .torrent file is otherwise now limited by annex.security.allowed-url-schemes. This does not address any external special remotes that might download an url themselves. Current thinking is all external special remotes will need to be audited for this problem, although many of them will use http libraries that only support http and not curl's menagarie. The related problem of accessing private localhost and LAN urls is not addressed by this commit. This commit was sponsored by Brett Eisenberg on Patreon.
2018-06-15 20:52:24 +00:00
filecmd "addurl" [url] @? ("addurl failed on " ++ url)
test case for addurl file://
2015-02-10 19:20:07 +00:00
let dest = "addurlurldest"
limit url downloads to whitelisted schemes Security fix! Allowing any schemes, particularly file: and possibly others like scp: allowed file exfiltration by anyone who had write access to the git repository, since they could add an annexed file using such an url, or using an url that redirected to such an url, and wait for the victim to get it into their repository and send them a copy. * Added annex.security.allowed-url-schemes setting, which defaults to only allowing http and https URLs. Note especially that file:/ is no longer enabled by default. * Removed annex.web-download-command, since its interface does not allow supporting annex.security.allowed-url-schemes across redirects. If you used this setting, you may want to instead use annex.web-options to pass options to curl. With annex.web-download-command removed, nearly all url accesses in git-annex are made via Utility.Url via http-client or curl. http-client only supports http and https, so no problem there. (Disabling one and not the other is not implemented.) Used curl --proto to limit the allowed url schemes. Note that this will cause git annex fsck --from web to mark files using a disallowed url scheme as not being present in the web. That seems acceptable; fsck --from web also does that when a web server is not available. youtube-dl already disabled file: itself (probably for similar reasons). The scheme check was also added to youtube-dl urls for completeness, although that check won't catch any redirects it might follow. But youtube-dl goes off and does its own thing with other protocols anyway, so that's fine. Special remotes that support other domain-specific url schemes are not affected by this change. In the bittorrent remote, aria2c can still download magnet: links. The download of the .torrent file is otherwise now limited by annex.security.allowed-url-schemes. This does not address any external special remotes that might download an url themselves. Current thinking is all external special remotes will need to be audited for this problem, although many of them will use http libraries that only support http and not curl's menagarie. The related problem of accessing private localhost and LAN urls is not addressed by this commit. This commit was sponsored by Brett Eisenberg on Patreon.
2018-06-15 20:52:24 +00:00
filecmd "addurl" ["--file", dest, url] @? ("addurl failed on " ++ url ++ " with --file")
test case for addurl file://
2015-02-10 19:20:07 +00:00
doesFileExist dest @? (dest ++ " missing after addurl --file")
Reference in a new issue Copy permalink