git-annex/Build/DistributionUpdate.hs

262 lines
8.3 KiB
Haskell
Raw Normal View History

{- Downloads git-annex autobuilds and installs them into the git-annex
- repository in ~/lib/downloads that is used to distribute git-annex
- releases.
-
- Generates info files, containing the version (of the corresponding file
- from the autobuild).
2014-04-21 15:24:34 +00:00
-
- Builds standalone rpms from the standalone tarballs, and populates
- a rpm package repository with them using the createrepo_c program.
-
2014-04-21 15:24:34 +00:00
- Also gpg signs the files.
-}
2013-11-22 16:21:53 +00:00
import Annex.Common
2013-11-22 16:21:53 +00:00
import Types.Distribution
2015-04-20 20:09:24 +00:00
import Build.Version (getChangelogVersion, Version)
2013-11-22 16:21:53 +00:00
import Utility.UserInfo
import Utility.Url
2018-01-15 16:18:00 +00:00
import Utility.Tmp.Dir
2018-04-10 00:22:46 +00:00
import Utility.Metered
2013-11-22 16:21:53 +00:00
import qualified Git.Construct
import qualified Annex
import Annex.Content
2016-01-14 19:55:37 +00:00
import Annex.WorkTree
2013-11-22 16:21:53 +00:00
import Git.Command
2020-11-27 17:02:43 +00:00
import qualified Utility.RawFilePath as R
import Annex.Action
2013-11-22 16:21:53 +00:00
import Data.Time.Clock
2014-06-18 20:11:20 +00:00
import Data.Char
2019-11-15 02:25:06 +00:00
import Data.Either
import System.Posix.Directory
2013-11-22 16:21:53 +00:00
2014-04-21 15:24:34 +00:00
-- git-annex distribution signing key (for Joey Hess)
signingKey :: String
signingKey = "89C809CB"
-- URL to an autobuilt git-annex file, and the place to install
-- it in the repository.
autobuilds :: [(URLString, FilePath)]
autobuilds =
(map linuxarch ["i386", "amd64", "armel", "arm64", "i386-ancient", "arm64-ancient"]) ++
2021-07-14 16:22:34 +00:00
[ (autobuild "x86_64-apple-catalina/git-annex.dmg", "git-annex/OSX/current/10.15_Catalina/git-annex.dmg")
, (autobuild "windows/git-annex-installer.exe", "git-annex/windows/current/git-annex-installer.exe")
]
where
linuxarch a =
( autobuild (a ++ "/git-annex-standalone-" ++ a ++ ".tar.gz")
, "git-annex/linux/current/git-annex-standalone-" ++ a ++ ".tar.gz"
)
2014-07-07 18:30:38 +00:00
autobuild f = "https://downloads.kitenet.net/git-annex/autobuild/" ++ f
-- Names of architectures in standalone tarballs and the corresponding
-- rpm architecture.
tarrpmarches :: [(String, String)]
tarrpmarches =
[ ("i386", "i386")
, ("amd64", "x86_64")
, ("arm64", "aarch64")
]
main :: IO ()
2013-11-22 16:21:53 +00:00
main = do
useFileSystemEncoding
version <- getChangelogVersion
repodir <- getRepoDir
topdir <- getCurrentDirectory
changeWorkingDirectory repodir
updated <- catMaybes <$> mapM (getbuild repodir) autobuilds
2020-11-16 13:56:03 +00:00
state <- Annex.new =<< Git.Construct.fromPath (toRawFilePath ".")
2020-03-30 19:13:16 +00:00
ood <- Annex.eval state $ do
buildrpms topdir updated
is <- makeinfos updated version
avoid flushing keys db queue after each Annex action The flush was only done Annex.run' to make sure that the queue was flushed before git-annex exits. But, doing it there means that as soon as one change gets queued, it gets flushed soon after, which contributes to excessive writes to the database, slowing git-annex down. (This does not yet speed git-annex up, but it is a stepping stone to doing so.) Database queues do not autoflush when garbage collected, so have to be flushed explicitly. I don't think it's possible to make them autoflush (except perhaps if git-annex sqitched to using ResourceT..). The comment in Database.Keys.closeDb used to be accurate, since the automatic flushing did mean that all writes reached the database even when closeDb was not called. But now, closeDb or flushDb needs to be called before stopping using an Annex state. So, removed that comment. In Remote.Git, change to using quiesce everywhere that it used to use stopCoProcesses. This means that uses on onLocal in there are just as slow as before. I considered only calling closeDb on the local git remotes when git-annex exits. But, the reason that Remote.Git calls stopCoProcesses in each onLocal is so as not to leave git processes running that have files open on the remote repo, when it's on removable media. So, it seemed to make sense to also closeDb after each one, since sqlite may also keep files open. Although that has not seemed to cause problems with removable media so far. It was also just easier to quiesce in each onLocal than once at the end. This does likely leave performance on the floor, so could be revisited. In Annex.Content.saveState, there was no reason to close the db, flushing it is enough. The rest of the changes are from auditing for Annex.new, and making sure that quiesce is called, after any action that might possibly need it. After that audit, I'm pretty sure that the change to Annex.run' is safe. The only concern might be that this does let more changes get queued for write to the db, and if git-annex is interrupted, those will be lost. But interrupting git-annex can obviously already prevent it from writing the most recent change to the db, so it must recover from such lost data... right? Sponsored-by: Dartmouth College's Datalad project
2022-10-12 17:50:46 +00:00
quiesce False
return is
syncToArchiveOrg
2020-03-30 19:13:16 +00:00
unless (null ood) $
error $ "Some info files are out of date: " ++ show (map fst ood)
-- Download a build from the autobuilder, virus check it, and return its
-- version.
-- It's very important that the version matches the build, otherwise
-- auto-upgrades can loop reatedly. So, check build-version before
-- and after downloading the file.
getbuild :: FilePath -> (URLString, FilePath) -> IO (Maybe (FilePath, Version))
getbuild repodir (url, f) = do
bv1 <- getbv
let dest = repodir </> f
let tmp = dest ++ ".tmp"
removeWhenExistsWith removeFile tmp
2020-11-16 13:56:03 +00:00
createDirectoryIfMissing True (fromRawFilePath (parentDir (toRawFilePath dest)))
2014-06-18 19:44:16 +00:00
let oops s = do
removeWhenExistsWith removeFile tmp
2014-06-18 19:44:16 +00:00
putStrLn $ "*** " ++ s
return Nothing
2018-04-10 00:22:46 +00:00
uo <- defUrlOptions
2021-09-03 16:36:42 +00:00
ifM (isRight <$> download nullMeterUpdate Nothing url tmp uo)
( ifM (liftIO $ virusFree tmp)
( do
bv2 <- getbv
case bv2 of
Nothing -> oops $ "no build-version file for " ++ url
(Just v)
| bv2 == bv1 -> do
removeWhenExistsWith removeFile dest
renameFile tmp dest
-- remove git rev part of version
let v' = takeWhile (/= '-') v
return $ Just (f, v')
| otherwise -> oops $ "build version changed while downloading " ++ url ++ " " ++ show (bv1, bv2)
, oops $ "VIRUS detected in " ++ url
)
2014-06-18 19:44:16 +00:00
, oops $ "failed to download " ++ url
)
where
2014-06-18 20:24:46 +00:00
bvurl = takeDirectory url ++ "/build-version"
getbv = do
2014-06-18 20:24:46 +00:00
bv <- catchDefaultIO "" $ readProcess "curl" ["--silent", bvurl]
2014-06-18 20:11:20 +00:00
return $ if null bv || any (not . versionchar) bv then Nothing else Just bv
versionchar c = isAlphaNum c || c == '.' || c == '-'
2013-11-22 16:21:53 +00:00
2020-05-01 23:05:35 +00:00
makeinfos :: [(FilePath, Version)] -> Version -> Annex [([Char], Maybe GitAnnexDistribution)]
makeinfos updated changelogversion = do
2014-11-11 20:49:24 +00:00
mapM_ (\f -> inRepo $ runBool [Param "annex", Param "add", File f]) (map fst updated)
void $ inRepo $ runBool
[ Param "commit"
2014-02-27 16:20:53 +00:00
, Param "-a"
2015-04-06 22:56:38 +00:00
, Param ("-S" ++ signingKey)
, Param "-m"
, Param $ "publishing git-annex " ++ descversion
]
2013-11-22 16:21:53 +00:00
now <- liftIO getCurrentTime
liftIO $ putStrLn $ "building info files"
forM_ updated $ \(f, bv) -> do
2020-07-10 18:17:35 +00:00
v <- lookupKey (toRawFilePath f)
2013-11-22 16:21:53 +00:00
case v of
Nothing -> noop
Just k -> whenM (inAnnex k) $ do
2013-11-22 16:21:53 +00:00
liftIO $ putStrLn f
let infofile = f ++ ".info"
let d = GitAnnexDistribution
{ distributionUrl = mkUrl f
2019-12-18 18:37:59 +00:00
, distributionKey = fromKey id k
, distributionVersion = bv
2013-11-22 16:21:53 +00:00
, distributionReleasedate = now
, distributionUrgentUpgrade = Just "6.20180626"
2013-11-22 16:21:53 +00:00
}
liftIO $ writeFile infofile $ formatInfoFile d
2014-04-21 15:24:34 +00:00
void $ inRepo $ runBool [Param "add", File infofile]
signFile infofile
signFile f
2013-11-22 16:21:53 +00:00
void $ inRepo $ runBool
[ Param "commit"
2015-04-06 22:56:38 +00:00
, Param ("-S" ++ signingKey)
2015-04-06 22:38:34 +00:00
, Param "-m"
, Param $ "updated info files for git-annex " ++ descversion
2013-11-22 16:21:53 +00:00
]
2013-11-22 19:02:31 +00:00
void $ inRepo $ runBool
2013-11-25 18:14:45 +00:00
[ Param "annex"
, Param "move"
, Param "--to"
, Param "website"
, Param "--force"
2013-11-22 19:02:31 +00:00
]
void $ inRepo $ runBool
2013-11-25 18:14:45 +00:00
[ Param "annex"
, Param "sync"
2013-11-22 19:02:31 +00:00
]
-- Check for out of date info files.
2014-02-10 19:33:37 +00:00
infos <- liftIO $ filter (".info" `isSuffixOf`)
fix empty tree import when directory does not exist Fix behavior when importing a tree from a directory remote when the directory does not exist. An empty tree was imported, rather than the import failing. Merging that tree would delete every file in the branch, if those files had been exported to the directory before. The problem was that dirContentsRecursive returned [] when the directory did not exist. Better for it to throw an exception. But in commit 74f0d67aa3988a71f3a53b88de4344272d924b95 back in 2012, I made it never theow exceptions, because exceptions throw inside unsafeInterleaveIO become untrappable when the list is being traversed. So, changed it to list the contents of the directory before entering unsafeInterleaveIO. So exceptions are thrown for the directory. But still not if it's unable to list the contents of a subdirectory. That's less of a problem, because the subdirectory does exist (or if not, it got removed after being listed, and it's ok to not include it in the list). A subdirectory that has permissions that don't allow listing it will have its contents omitted from the list still. (Might be better to have it return a type that includes indications of errors listing contents of subdirectories?) The rest of the changes are making callers of dirContentsRecursive use emptyWhenDoesNotExist when they relied on the behavior of it not throwing an exception when the directory does not exist. Note that it's possible some callers of dirContentsRecursive that used to ignore permissions problems listing a directory will now start throwing exceptions on them. The fix to the directory special remote consisted of not making its call in listImportableContentsM use emptyWhenDoesNotExist. So it will throw an exception as desired. Sponsored-by: Joshua Antonishen on Patreon
2023-08-15 16:57:41 +00:00
<$> emptyWhenDoesNotExist (dirContentsRecursive "git-annex")
ds <- liftIO $ forM infos (readish <$$> readFile)
let dis = zip infos ds
let ood = filter outofdate dis
2020-03-30 19:13:16 +00:00
return ood
where
outofdate (_, md) = case md of
Nothing -> True
Just d -> distributionVersion d /= changelogversion
descversion = unwords (nub (map snd updated))
2013-11-22 16:21:53 +00:00
getRepoDir :: IO FilePath
getRepoDir = do
home <- liftIO myHomeDir
return $ home </> "lib" </> "downloads"
mkUrl :: FilePath -> String
mkUrl f = "https://downloads.kitenet.net/" ++ f
2014-04-21 15:24:34 +00:00
signFile :: FilePath -> Annex ()
signFile f = do
void $ liftIO $ boolSystem "gpg"
[ Param "-a"
, Param $ "--default-key=" ++ signingKey
2014-04-21 15:56:06 +00:00
, Param "--detach-sign"
2014-04-21 15:24:34 +00:00
, File f
]
2022-07-25 18:10:30 +00:00
liftIO $ R.rename (toRawFilePath (f ++ ".asc")) (toRawFilePath (f ++ ".sig"))
2014-04-21 15:24:34 +00:00
void $ inRepo $ runBool [Param "add", File (f ++ ".sig")]
-- clamscan should handle unpacking archives, but did not in my
-- testing, so do it manually.
virusFree :: FilePath -> IO Bool
virusFree f
| ".tar.gz" `isSuffixOf` f = unpack $ \tmpdir ->
boolSystem "tar" [ Param "xf", File f, Param "-C", File tmpdir ]
| ".dmg" `isSuffixOf` f = unpack $ \tmpdir -> do
-- 7z can extract partitions from a dmg, and then
-- run on partitions can extract their files
unhfs tmpdir f
parts <- filter (".hfs" `isSuffixOf`) <$> getDirectoryContents tmpdir
forM_ parts $ unhfs tmpdir
return True
| otherwise = clamscan f
where
clamscan f' = boolSystem "clamscan"
[ Param "--no-summary"
, Param "-r"
, Param f'
]
unpack unpacker = withTmpDir "clamscan" $ \tmpdir -> do
unlessM (unpacker tmpdir) $
error $ "Failed to unpack " ++ f ++ " for virus scan"
clamscan tmpdir
unhfs dest f' = unlessM (boolSystem "7z" [ Param "x", Param ("-o" ++ dest), File f' ]) $
error $ "Failed extracting hfs " ++ f'
2019-09-13 16:00:16 +00:00
buildrpms :: FilePath -> [(FilePath, Version)] -> Annex ()
buildrpms topdir l = do
liftIO $ createDirectoryIfMissing True rpmrepo
oldrpms <- map (rpmrepo </>) . filter (".rpm" `isSuffixOf`)
2019-12-18 18:36:19 +00:00
<$> liftIO (getDirectoryContents rpmrepo)
2019-09-13 16:00:16 +00:00
forM_ tarrpmarches $ \(tararch, rpmarch) ->
forM_ (filter (isstandalonetarball tararch . fst) l) $ \(tarball, v) -> do
2020-11-27 17:02:43 +00:00
liftIO $ mapM_ (removeWhenExistsWith (R.removeLink . toRawFilePath))
(filter ((rpmarch ++ ".rpm") `isSuffixOf`) oldrpms)
2019-09-13 16:00:16 +00:00
void $ liftIO $ boolSystem script
[ Param rpmarch
, File tarball
, Param v
, File rpmrepo
]
void $ inRepo $ runBool [Param "annex", Param "get", File rpmrepo]
void $ liftIO $ boolSystem "createrepo_c" [File rpmrepo]
2019-09-13 16:00:16 +00:00
void $ inRepo $ runBool [Param "annex", Param "add", File rpmrepo]
where
isstandalonetarball tararch f =
("git-annex-standalone-" ++ tararch ++ ".tar.gz") `isSuffixOf` f
script = topdir </> "standalone" </> "rpm" </> "rpmbuild-from-standalone-tarball"
rpmrepo = "git-annex/linux/current/rpms"
-- My .mrconfig is configured to copy new files to archive.org,
-- and moves old versions of content to archive.org to free up space on my
-- server.
syncToArchiveOrg :: IO ()
syncToArchiveOrg = void $ boolSystem "mr"
[ Param "-d"
, File "/srv/web/downloads.kitenet.net"
, Param "update"
]