replaceFile created a temp file, which was guaranteed to not overlap with
another temp file. However, makeAnnexLink then deleted that file, in
preparation for making the symlink in its place. This caused a race, since
some other replaceFile could create a temp file, using the same name!
I was able to reproduce the race easily running git-annex add -J10 in a
directory with 100 files (all with different contents). Some files would
get ingested into the annex, but their annex links would fail to be added.
There could be other situations where this same problem could occur.
Perhaps when the assistant is adding a file, if the user manually also ran
git-annex add. Perhaps in cases not involving adding a file.
The new replaceFile makes a temprary directory, which is guaranteed to be
unique, and doesn't make a temp file in there. makeAnnexLink can thus
create the symlink without problem and the race is avoided.
Audited all calls to replaceFile to make sure that the old behavior of
providing an empty temp file was not relied on.
The general problem of asking for a temp file and deleting it as part of
the process of using it could reach beyond replaceFile. Did some quick
audits and didn't find other cases of it. Probably only symlink creation
stuff would tend to make that mistake, mostly.
* Fix failure to build with aws-0.13.0.
* When built with aws-0.13.0, the S3 special remote can be used to create
google nearline buckets, by setting storageclass=NEARLINE.
Fixes a recent-ish build warning on about 64 bit vs non.
This is the method used by the disk-free-space library, and I tested it to
yield the same results on even 10 tb drives on OSX -- so it's getting 64
bit values.
Instead, only display transport error if the configlist output doesn't
include an annex.uuid line, even an empty one.
A recent change made git-annex init try to get all the remote uuids, and so
the transport error would be displayed by it. It was also displayed when
eg, copying files to a remote that had no uuid yet.
sync, merge, assistant: When git merge failed for a reason other than a
conflicted merge, such as a crippled filesystem not allowing particular
characters in filenames, git-annex would make a merge commit that could
omit such files or otherwise be bad. Fixed by aborting the whole merge
process when git merge fails for any reason other than a merge conflict.
Implemented with no additional overhead of compares etc.
This is safe to do for presence logs because of their locality of change;
a given repo's presence logs are only ever changed in that repo, or in a
repo that has just been actively changing the content of that repo.
So, we don't need to worry about a split-brain situation where there'd
be disagreement about the location of a key in a repo. And so, it's ok to
not update the timestamp when that's the only change that would be made
due to logging presence info.
/dev/null stderr; ssh is still able to display a password prompt
despite this
Show some messages so the user knows it's locking a remote, and
knows if that locking failed.
On Solaris, using f_bsize provided a value that is apparently much larger
than the real block size. The solaris docs for statvfs say
f_bsize is the "preferred" file system block size, and I guess the
filesystem prefers larger blocks, but uses smaller ones or something.
The docs also say that f_frsize is the "fundamental" block size.
Switched to using f_frsize on Linux and kFreeBSD too, since I guess
f_bsize could in theory vary the same way there too. Assuming that Solaris
is not violating the posix spec, I guess the linux man page for statvfs
is not as well written and I misunderstood it.
* When annex objects are received into git repositories, their checksums are
verified then too.
* To get the old, faster, behavior of not verifying checksums, set
annex.verify=false, or remote.<name>.annex-verify=false.
* setkey, rekey: These commands also now verify that the provided file
matches the key, unless annex.verify=false.
* reinject: Already verified content; this can now be disabled by
setting annex.verify=false.
recvkey and reinject already did verification, so removed now duplicate
code from them. fsck still does its own verification, which is ok since it
does not use getViaTmp, so verification doesn't happen twice when using fsck
--from.
Update install instructions.
Note that xmpp is disabled in the stack.yaml, because that needs C
libraries that would take extra work for the user to install.
xmpp is moving toward deprecation anyway.
Seems easy, but git ls-files can't list the right subset of files.
So, I wrote a whole new parser for git status output, and converted the
status command to use that.
There are a few other small behavior changes. The order changed. Unlocked
files show as T. In indirect mode, deleted files were not shown before, and
that's fixed. Regular files checked directly into git and modified
were not shown before, and are now.
I couldn't find a good way to make an *empty* index file (zero byte file
won't do), so I punted and just don't make index.lock when there's no index
yet. This means some other git process could race and write an index file
at the same time as the merge is ongoing, in theory. Only happens in new
repos though.
It's confusing, easy to select by accident and get into a situation the
webapp offers no easy recovery from, and pausing syncing works just as well
in most situations.
Added support for storageclass=STANDARD_IA to use Amazon's
new Infrequently Accessed storage.
Also allows using storageclass=NEARLINE to use Google's NearLine storage.
The necessary changes to aws to support this are in
https://github.com/aristidb/aws/pull/176
This is pretty complicated, but I have both "git-annex" and "git annex"
working both in the git bash shell even with git not added to path.
And, when git's added to path, both work from MS-DOS prompt window too.
I think that the webapp startup does still need git in path, so
instructions will keep saying to do that. But, users often disregard them,
and hopefully this will reduce support traffic.
Also, switched the wget from the cygwin one to the msys2 one, avoiding the
complication of needing to bundle any cygwin dlls.
Using msysgit with git-annex is no longer supported.
At the same time, I'm updating the rsync.exe in my downloads repository
with the one from msys2.
Note that rsync is currently still being ldded and installed in Git/cmd/
like the other cygwin programs. The ldd fails and this failure is ignored.
It would be better to special case it to go in Git/usr/bin/, so that the
user can't run rsync in a dos prompt window, which doesn't work, as it needs
additional libs. However, as far as git-annex running rsync running ssh,
it works ok in this location.
Removed the ssh.cmd and ssh-keygen.cmd; these are not needed with git for
windows. Keeping them would let ssh be run manually from a dos prompt
window, but that's not really a goal.
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.
It was failing at link time, some problem with terminatePID.
Re-implemented that to not use a C wrapper function, which cleared up the
problem. Removed old EvilLinker hack with must have been related to the
same problem.
Note that I have not tested this with older ghc's. In
f11f7520b5 I mention having tried this
approach before, and getting segfaults.. So, who knows. It seems to work
fine with ghc 7.10 at least.
Ben Boeckel had a patch, but..
Actually, that was not the only place that used ScheduleIncremental when
built w/o database. Since the data type doesn't need database stuff,
I've instead fixed this build problem by exposing the
ScheduleIncremental constructor to database-less builds.
As a result of the Makefile changes, the Debian package is built
with various hardening options. Although their benefit to a largely
haskell program is unknown.
Oh boy, not again. So, another place that the filesystem encoding needs to
be applied. Yay.
In passing, I changed decodeBS so if a NUL is embedded in the input, the
resulting FilePath doesn't get truncated at that NUL. This was needed to
make prop_b64_roundtrips pass, and on reviewing the callers of decodeBS, I
didn't see any where this wouldn't make sense. When a FilePath is used to
operate on the filesystem, it'll get truncated at a NUL anyway, whereas if
a String is being used for something else, it might conceivably have a NUL
in it, and we wouldn't want it to get truncated when going through
decodeBS.
(NB: There may be a speed impact from this change.)
While cryptohash has SHA3 support, it has not been updated for the final
version of the spec. Note that cryptonite has not been ported to all arches
that cryptohash builds on yet.
This was potentially a hole in the readonly mode armor even before my last
commit. If the user could push a git-annex branch to a repo, they could get
git-annex-shell to initialize the repo. After my last commit, the user
didn't even need to be allowed to push a branch to init the repo, so
this hole certianly needs to be closed now.
Now it suffices to run git remote add, followed by git-annex sync. Now the
remote is automatically initialized for use by git-annex, where before the
git-annex branch had to manually be pushed before using git-annex sync.
Note that this involved changes to git-annex-shell, so if the remote is
using an old version, the manual push is still needed.
Implementation required git-annex-shell be changed, so configlist can
autoinit a repository even when no git-annex branch has been pushed yet.
Unfortunate because we'll have to wait for it to get deployed to servers
before being able to rely on this change in the documentation.
Did consider making git-annex sync push the git-annex branch to repos that
didn't have a uuid, but this seemed difficult to do without complicating it
in messy ways.
It would be cleaner to split a command out from configlist to handle
the initialization. But this is difficult without sacrificing backwards
compatability, for users of old git-annex versions which would not use the
new command.
I think that the problem was caused by windows not having a concept of an
env var that is set, but to the empty string. So, GIT_ANNEX_SSHOPTION
got set to "" and was not seen as set at all.
Easy fix, which also makes git-annex sync a little faster is to not set
GIT_SSH, when GIT_ANNEX_SSHOPTION has no options. Might as well let git use
ssh per usual in this case, no need to run git-annex as the proxy ssh
command..
This fixes a reversion introduced by relative path changes back last winter.
The root cause is simplifyPath "../foo" was incorrectly yielding "foo".
absPathFrom seems quite horrible. Probably most things that use it should
use </> instead.
Git.Ref.headSha doesn't really work in direct mode as there's not a head,
so it was actually diffing against the empty tree and so not removing any
deleted files. Get the sha of the current branch instead, which is the same
thing Command.Sync does.
* proxy: Fix proxy git commit of non-annexed files in direct mode.
* proxy: If a non-proxied git command, such as git revert
would normally fail because of unstaged files in the work tree,
make the proxied command fail the same way.
* Perform a clean shutdown when --time-limit is reached.
This includes running queued git commands, and cleanup actions normally
run when a command is finished.
* fsck: Commit incremental fsck database when --time-limit is reached.
Previously, some of the last files fscked did not make it into the
database when using --time-limit.
Note that this changes Annex.addCleanup hooks, to run after --time-limit
expires. Fsck was using such a hook to clean up after a
--incremental-schedule, and that shouldn't run when --time-limit exipires
it. So, instead, moved that cleanup code to be run by cleanupIncremental.
Resulted in some data type juggling.
I've seen rss feeds that have no permalinks, only guids (which are
sometimes in the form of permalinks, argh/sigh).
I had previously avoided trusting guids to be globally unique, because my
survey of rss feeds that I subscribe to shows a lot of pretty bad
"guids" like "2 at http://serialpodcast.org" or even worse "oth20150401-hq".
Worry was that two podcasts that are generating guids so badly, that
there's no guarantee they're actually globally unique.
But, I'm seeing too many url changes that result in redundant files, so
let's try this. If feeds are so broken that guids overlap, they could just
as well incorrectly call them permalinks too.
"checkPresent baser" was wrong; the baser has a dummy checkPresent action
not the real one. So, to fix this, we need to call preparecheckpresent to
get a checkpresent action that can be used to check if chunks are present.
Note that, for remotes like S3, this means that the preparer is run,
which opens a S3 handle, that will be used for each checkpresent of a
chunk. That's a good thing; if we're resuming an upload that's already many
chunks in, it'll reuse that same http connection for each chunk it checks.
Still, it's not a perfectly ideal thing, since this is a different http
connection that the one that will be used to upload chunks. It would be
nice to improve the API so that both use the same http connection.
The branch needs to be created when merging from the remote in sync,
since we diff between it and the remote's sync branch. But git annex merge
should not be creating sync branches.
This was a reversion caused by the relative path changes in 5.20150113.
Other uses of addAuthorizedKeys seem to be ok. If the user enters a
directory like ~/annex, it writes GIT_ANNEX_SHELL_DIRECTORY=annex, and
git-annex-shell assumes that's relative to HOME.
This makes git annex unused use around 48 mb more memory than it did before,
but the massive increase in accuracy makes this worthwhile for all but the
smallest systems.
Also, I want to use the bloom filter for sync --all --content, to avoid
dropping files that the preferred content doesn't want, and 1/1000
false positives would be far too many in that use case, even if it were
acceptable for unused.
Actual memory use numbers:
1000: 21.06user 3.42system 0:26.40elapsed 92%CPU (0avgtext+0avgdata 501552maxresident)k
1000000: 21.41user 3.55system 0:26.84elapsed 93%CPU (0avgtext+0avgdata 549496maxresident)k
10000000: 21.84user 3.52system 0:27.89elapsed 90%CPU (0avgtext+0avgdata 549920maxresident)k
Based on these numbers, 10 million seemed a better pick than 1 million.
backup: Use new "anything" terminal. This means that content that
is not unused, but has no associated file will be wanted by backup repos.
unwanted: "not anything" will result in any and all content moving
off of these repos.
incremental backup: Remove the "(include=* or unused)",
so it matches content that has no associated files
but is not unused.
client: Add a include=* to the expression. This limits it to matching
only files in the work tree. Without this change, sync --all --content
would match a key against the expression, and since it matches
exclude=archive/*, the client repo would have wanted the file content.
The "and not unused" would have kept unused objects out, but not
objects that were not known to be unused, or objects that another branch
referred to. In practice, everything would have flooded into client repos
without this change.
In my tests, this has to be set when uploading a file to the bucket
and then the file can be accessed using the bucketname.s3.amazonaws.com
url.
Setting it when creating the bucket didn't seem to make the whole bucket
public, or allow accessing files stored in it. But I have gone ahead and
also sent it when creating the bucket just in case that is needed in some
case.
This reverts commit cf650eaa99.
It's too early to do this; the linux and android autobuilder will need to
be updated to use the new version of shakespeare, and that will require a
complete refresh of them. In the meantime, this has knocked the webapp out
of the autobuilders.
This is especially useful because the caller doesn't need to generate valid
url keys, which involves some escaping of characters, and may involve
taking a md5sum of the url if it's too long.
The one exception is in Utility.Daemon. As long as a process only
daemonizes once, which seems reasonable, and as long as it avoids calling
checkDaemon once it's already running as a daemon, the fcntl locking
gotchas won't be a problem there.
Annex.LockFile has it's own separate lock pool layer, which has been
renamed to LockCache. This is a persistent cache of locks that persist
until closed.
This is not quite done; lockContent stil needs to be converted.
Only the assistant uses these, and only the assistant cleans them up, so
make only git annex transferkeys write them,
There is one behavior change from this. If glacier is being used, and a
manual git annex get --from glacier fails because the file isn't available
yet, the assistant will no longer later see that failed transfer file and
retry the get. Hope no-one depended on that old behavior.
I've tested all the dataenc to sandi conversions except Assistant.XMPP,
and all have unchanged behavior, including behavior on large unicode code
points.
For example, it failed to get files from a bucket named S3.
Also fixes `git annex initremote UPPERCASE type=S3`, which failed with the
new aws library, with a signing error message.
The setDifferences that got added to initialize turns out to make a git
commit, and before ensureCommit has been used. Thus, repo init can fail
when the system has a broken hostname etc.
Move the ensureCommit to the very first thing to avoid this kind of breakage.
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.
Since we started using this for git repos, when a remote was on another
drive, it resulted in a bogus relative path to it being used by git-annex,
which didn't work.
This is a nearly free feature; it piggybacks on the location log lookups
done for the numcopies stats. So, the only extra overhead is updating
the map of repository sizes.
However, I had to switch to Data.Map.Strict, which needs containers 0.5.
If backporting to wheezy, will probably need to revert this commit.
This works, and seems fairly robust. Clean get of 20 files at -J3. At -J10,
there are some messages about ssh multiplexing, probably due to a race
spinning up the ssh connection cacher. But, it manages to get all the files
ok regardless.
The progress bars are a scrambled mess though, due to bugs in
ascii-progress, which I've already filed. Particularly this one:
https://github.com/yamadapc/haskell-ascii-progress/issues/8
webapp: When adding another local repository, and combining it with the
current repository, the new repository's remote path was set to "." rather
than the path to the current repository. This was a reversion caused by the
relative path changes in 5.20150113.
git-checkignore refuses to work if any pathspec options are set. Urgh.
I audited the rest of git, and no other commands used by git-annex have
such limitations. Indeed, AFAICS, *all* other commands support
--literal-pathspecs. So, worked around this where git-checkignore is
called.
I don't quite understand the cause of the deadlock. It only occurred
when git-annex-shell transferinfo was being spawned over ssh to feed
download transfer progress back. And if I removed this line from
feedprogressback, the deadlock didn't occur:
bytes <- readSV v
The problem was not a leaked FD, as far as I could see. So what was it?
I don't know.
Anyway, this is a nice clean implementation, that avoids the deadlock.
Just fork off the async threads to handle filtering the stdout and stderr,
and let them clean up their handles whenever they decide to exit.
I've verified that the handles do get promptly closed, although a little
later than I would expect. Presumably that "little later" is what
was making waiting on the threads deadlock.
Despite the late exit, the last line of stdout and stderr appears where
I'd want it to, so I guess this is ok..
Stderr reader blocks waiting for all stderr, and so blocks the process ever
exiting.
I tried several ways to get around this, but no success yet. For now,
disable the stderr reader entirely.
New approach is to do it the expensive way for the first 100 paths
on the command line, but then assume the user doesn't care about order too
much and fall back to the cheap way that does not preserve order.
This will only ever result in a few more git-ls-files being run than were run
before. (Only 1 more is really needed, but around 10 more are currently run
for a max length command line.)
So, no need to worry about the extra zombie, or lost laziness due to concat.
Note that previously, `git annex find *.jpg` would find eg, foo/bar.jpg.
That was never intended or documented behavior, so I'm going to change it.
But this is potentially a behavior change if someone discovered that
behavior and relied on it despite it being accidental. Oh well.. can't make
an omlette w/o breaking some eggs.
This was introduced by commit 450ee53ab6
However, the same problem could affect other calls to programPath,
specifically some on the assistant. So, I fixed it at a deeper level.
In this situation, curl -o exits successfully without creating the output
file.
There was already a workaround for curl file:/// but I did not realize this
also affected regular url downloads.
To fix it, pre-create the destination file before starting curl.
Since we cannot always know the size of an url before trying to download
it, let's always do this.
Note that since curl is told -C -, we have to consider if this
makes curl try to do a ranged download, which might fail on some servers
where a regular download would have succeeded. My testing indicates
this isn't a problem; since the file is empty, curl seems to not try to
do a ranged download.
Original report: https://github.com/datalad/datalad/issues/79
Curl bug report: https://github.com/bagder/curl/issues/183
Useful for things like ipfs that don't use regular urls.
An external special remote can add a regular url to a key, and then
git-annex get will download it from the web. But for ipfs, we want to
instead tell git-annex that the uri uses OtherDownloader. Before this
change, the external special remote protocol lacked a way to do that.
Seen for example, a newly checked out git submodule. In this case,
.git/HEAD is a raw sha, rather than the usual reference to a ref.
Removed currentSha in passing, since it was a more roundabout way of
doing what headSha does, and headSha is more robust.
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.
Seems to work, but still experimental until it's been tested more.
When repositories are on filesystems not supporting symlinks, the .git dir
symlink trick cannot be used. Since we're going to be in direct mode
anyway, the .git dir symlink is not strictly needed.
However, I have not fixed the code that creates new annex symlinks to
handle this case -- the committed symlinks will be wrong.
git annex sync happens to currently fail in a submodule using direct mode,
because there's no HEAD ref. That also needs to be dealt with to get
this fully working in crippled filesystems.
Leaving http://github.com/datalad/datalad/issues/44 open until these issues
are dealt with.
Most of the time, there will be no discreprancy between programPath and
readProgramFile.
But, the programFile might have been written by an old version of git-annex
that is still installed, while a newer one is currently running. In this
case, we want to run the same one that's currently running.
This is especially important for things like the GIT_SSH=git-annex used for
ssh connection caching.
The only code that still uses readProgramFile directly is the upgrade code,
which needs to know where the standalone git-annex was installed, in order to
upgrade it.
See my comment in the bug report for analysis; basically this is safe
because it's a non-forced push, so won't lose history. Even if it was a
forced push or somehow races, things will eventually become consistent and
no git-annex branch info will be lost.
(This used to be done, but it forgot to do it since version 4.20130909.)
* sync: Use the ssh-options git config when doing git pull and push.
* remotedaemon: Use the ssh-options git config.
Note that the rename env var means that if a new git-annex calls an old one
for git-annex ssh, or a new calls an old, nothing much will go wrong;
just ssh caching won't happen.
Note that while the assistant detects changes made to remote names, I left
the commit message fixed rather than calculating it after every commit. It
doesn't seem worth the CPU to do the latter.
hGetSomeString reads one byte at a time, so unicode bytes are not composed.
The problem comes when outputting that to the console with hPut; that
tried to apply the handle's encoding, and so we get mojibake.
Instead, use ByteStrings, and only convert it to a string for parsing, not
for display.
Note that there are a couple of other things that use hGetSomeString,
which I've left as-is for now.
