git-annex/Annex.hs

490 lines
14 KiB
Haskell
Raw Normal View History

2010-10-27 20:53:54 +00:00
{- git-annex monad
-
- Copyright 2010-2024 Joey Hess <id@joeyh.name>
2010-10-27 20:53:54 +00:00
-
- Licensed under the GNU AGPL version 3 or higher.
2010-10-27 20:53:54 +00:00
-}
2010-10-10 19:04:07 +00:00
{-# LANGUAGE GeneralizedNewtypeDeriving, BangPatterns, PackageImports #-}
2010-10-11 21:52:46 +00:00
module Annex (
Annex,
AnnexState(..),
AnnexRead(..),
2010-10-14 07:18:11 +00:00
new,
run,
2010-11-01 03:24:16 +00:00
eval,
makeRunner,
getRead,
getState,
changeState,
withState,
setField,
2012-04-30 17:59:05 +00:00
setOutput,
getField,
addCleanupAction,
gitRepo,
inRepo,
fromRepo,
calcRepo,
calcRepo',
getGitConfig,
overrideGitConfig,
changeGitRepo,
adjustGitRepo,
addGitConfigOverride,
getGitConfigOverrides,
2014-05-16 20:08:20 +00:00
getRemoteGitConfig,
withCurrentState,
changeDirectory,
getGitRemotes,
2015-04-30 19:04:01 +00:00
incError,
2010-10-11 21:52:46 +00:00
) where
2010-10-10 19:04:07 +00:00
import Common
import qualified Git
import qualified Git.Config
import qualified Git.Construct
import Annex.Fixup
import Git.HashObject
import Git.CheckAttr
gitignore support for the assistant and watcher Requires git 1.8.4 or newer. When it's installed, a background git check-ignore process is run, and used to efficiently check ignores whenever a new file is added. Thanks to Adam Spiers, for getting the necessary support into git for this. A complication is what to do about files that are gitignored but have been checked into git anyway. git commands assume the ignore has been overridden in this case, and not need any more overriding to commit a changed version. However, for the assistant to do the same, it would have to run git ls-files to check if the ignored file is in git. This is somewhat expensive. Or it could use the running git-cat-file process to query the file that way, but that requires transferring the whole file content over a pipe, so it can be quite expensive too, for files that are not git-annex symlinks. Now imagine if the user knows that a file or directory tree will be getting frequent changes, and doesn't want the assistant to sync it, so gitignores it. The assistant could overload the system with repeated ls-files checks! So, I've decided that the assistant will not automatically commit changes to files that are gitignored. This is a tradeoff. Hopefully it won't be a problem to adjust .gitignore settings to not ignore files you want the assistant to autocommit, or to manually git annex add files that are listed in .gitignore. (This could be revisited if git-annex gets access to an interface to check the content of the index w/o forking a git command. This could be libgit2, or perhaps a separate git cat-file --batch-check process, so it wouldn't need to ship over the whole file content.) This commit was sponsored by Francois Marier. Thanks!
2013-08-02 23:31:55 +00:00
import Git.CheckIgnore
import qualified Git.Hook
2011-12-20 18:37:53 +00:00
import qualified Git.Queue
import Types.Key
import Types.Backend
import Types.GitConfig
import qualified Types.Remote
import Types.Crypto
import Types.BranchState
import Types.TrustLevel
import Types.Group
import Types.Messages
import Types.Concurrency
import Types.UUID
2013-05-25 03:07:26 +00:00
import Types.FileMatcher
import Types.NumCopies
import Types.LockCache
import Types.DesktopNotify
2014-03-13 23:06:26 +00:00
import Types.CleanupActions
import Types.AdjustedBranch
import Types.WorkerPool
import Types.IndexFiles
import Types.CatFileHandles
import Types.RemoteConfig
import Types.TransferrerPool
import Types.VectorClock
2024-06-14 15:16:01 +00:00
import Types.Cluster
import Types.RepoSize
import Annex.VectorClock.Utility
import Annex.Debug.Utility
import qualified Database.Keys.Handle as Keys
import Database.RepoSize.Handle
import Utility.InodeCache
import Utility.Url
import Utility.ResourcePool
import Utility.HumanTime
import Git.Credential (CredentialCache(..))
import "mtl" Control.Monad.Reader
import Control.Concurrent
2017-05-11 22:29:51 +00:00
import Control.Concurrent.STM
import qualified Control.Monad.Fail as Fail
import qualified Data.Map.Strict as M
import qualified Data.Set as S
import Data.Time.Clock.POSIX
{- git-annex's monad is a ReaderT around an AnnexState stored in a MVar,
- and an AnnexRead. The MVar is not exposed outside this module.
-
- Note that when an Annex action fails and the exception is caught,
- any changes the action has made to the AnnexState are retained,
- due to the use of the MVar to store the state.
Switch to MonadCatchIO-transformers for better handling of state while catching exceptions. As seen in this bug report, the lifted exception handling using the StateT monad throws away state changes when an action throws an exception. http://git-annex.branchable.com/bugs/git_annex_fork_bombs_on_gpg_file/ .. Which can result in cached values being redundantly calculated, or other possibly worse bugs when the annex state gets out of sync with reality. This switches from a StateT AnnexState to a ReaderT (MVar AnnexState). All changes to the state go via the MVar. So when an Annex action is running inside an exception handler, and it makes some changes, they immediately go into affect in the MVar. If it then throws an exception (or even crashes its thread!), the state changes are still in effect. The MonadCatchIO-transformers change is actually only incidental. I could have kept on using lifted-base for the exception handling. However, I'd have needed to write a new instance of MonadBaseControl for the new monad.. and I didn't write the old instance.. I begged Bas and he kindly sent it to me. Happily, MonadCatchIO-transformers is able to derive a MonadCatchIO instance for my monad. This is a deep level change. It passes the test suite! What could it break? Well.. The most likely breakage would be to code that runs an Annex action in an exception handler, and *wants* state changes to be thrown away. Perhaps the state changes leaves the state inconsistent, or wrong. Since there are relatively few places in git-annex that catch exceptions in the Annex monad, and the AnnexState is generally just used to cache calculated data, this is unlikely to be a problem. Oh yeah, this change also makes Assistant.Types.ThreadedMonad a bit redundant. It's now entirely possible to run concurrent Annex actions in different threads, all sharing access to the same state! The ThreadedMonad just adds some extra work on top of that, with its own MVar, and avoids such actions possibly stepping on one-another's toes. I have not gotten rid of it, but might try that later. Being able to run concurrent Annex actions would simplify parts of the Assistant code.
2013-05-19 18:16:36 +00:00
-}
newtype Annex a = Annex { runAnnex :: ReaderT (MVar AnnexState, AnnexRead) IO a }
deriving (
Monad,
MonadIO,
MonadReader (MVar AnnexState, AnnexRead),
MonadCatch,
MonadThrow,
MonadMask,
Fail.MonadFail,
Functor,
Applicative,
Alternative
)
-- Values that can be read, but not modified by an Annex action.
data AnnexRead = AnnexRead
share single BranchState amoung all threads This fixes a problem when git-annex testremote is run against a cluster accessed via the http server. Annex.Cluster uses the location log to find nodes that contain a key when checking if the key is present or getting it. Just after a key was stored to a cluster node, reading the location log was not getting the UUID of that node. Apparently the Annex action that wrote to the location log, and the one that read from it were run with two different Annex states. The http server does use several different Annex threads. BranchState was part of the AnnexState, and so two threads could have different BranchStates. Moved BranchState to the AnnexRead, so all threads will see the common state. This might possibly impact performance. If one thread is writing changes to the branch, and another thread is reading from the branch, the writing thread will now invalidate the BranchState's cache, which will cause the reading thread to need to do extra work. But correctness is surely more important. If did is found to have impacted performance, it could probably be dealt with by doing smarter BranchState cache invalidation. Another way this might impact performance is that the BranchState has a small cache. If several threads were reading from the branch and relying on the value they just read still being in the case, now a cache miss will be more likely. Increasing the BranchState cache to the number of jobs might be a good idea to amelorate that. But the cache is currently an innefficient list, so making it large would need changes to the data types. (Commit 4304f1b6aea19a5c402dc4f9d69aa4ff1c104c9b dealt with a follow-on effect of the bug fixed here.)
2024-07-28 16:17:16 +00:00
{ branchstate :: MVar BranchState
, activekeys :: TVar (M.Map Key ThreadId)
, activeremotes :: MVar (M.Map (Types.Remote.RemoteA Annex) Integer)
, keysdbhandle :: Keys.DbHandle
, sshstalecleaned :: TMVar Bool
, signalactions :: TVar (M.Map SignalAction (Int -> IO ()))
, transferrerpool :: TransferrerPool
, debugenabled :: Bool
, debugselector :: DebugSelector
, explainenabled :: Bool
, ciphers :: TMVar (M.Map StorableCipher Cipher)
, fast :: Bool
, force :: Bool
, forcenumcopies :: Maybe NumCopies
, forcemincopies :: Maybe MinCopies
, forcebackend :: Maybe String
, reposizes :: MVar (Maybe (M.Map UUID (RepoSize, SizeOffset)))
, rebalance :: Bool
, useragent :: Maybe String
, desktopnotify :: DesktopNotify
, gitcredentialcache :: TMVar CredentialCache
}
newAnnexRead :: GitConfig -> IO AnnexRead
newAnnexRead c = do
share single BranchState amoung all threads This fixes a problem when git-annex testremote is run against a cluster accessed via the http server. Annex.Cluster uses the location log to find nodes that contain a key when checking if the key is present or getting it. Just after a key was stored to a cluster node, reading the location log was not getting the UUID of that node. Apparently the Annex action that wrote to the location log, and the one that read from it were run with two different Annex states. The http server does use several different Annex threads. BranchState was part of the AnnexState, and so two threads could have different BranchStates. Moved BranchState to the AnnexRead, so all threads will see the common state. This might possibly impact performance. If one thread is writing changes to the branch, and another thread is reading from the branch, the writing thread will now invalidate the BranchState's cache, which will cause the reading thread to need to do extra work. But correctness is surely more important. If did is found to have impacted performance, it could probably be dealt with by doing smarter BranchState cache invalidation. Another way this might impact performance is that the BranchState has a small cache. If several threads were reading from the branch and relying on the value they just read still being in the case, now a cache miss will be more likely. Increasing the BranchState cache to the number of jobs might be a good idea to amelorate that. But the cache is currently an innefficient list, so making it large would need changes to the data types. (Commit 4304f1b6aea19a5c402dc4f9d69aa4ff1c104c9b dealt with a follow-on effect of the bug fixed here.)
2024-07-28 16:17:16 +00:00
bs <- newMVar startBranchState
emptyactivekeys <- newTVarIO M.empty
emptyactiveremotes <- newMVar M.empty
kh <- Keys.newDbHandle
sc <- newTMVarIO False
si <- newTVarIO M.empty
tp <- newTransferrerPool
cm <- newTMVarIO M.empty
cc <- newTMVarIO (CredentialCache M.empty)
rs <- newMVar Nothing
return $ AnnexRead
share single BranchState amoung all threads This fixes a problem when git-annex testremote is run against a cluster accessed via the http server. Annex.Cluster uses the location log to find nodes that contain a key when checking if the key is present or getting it. Just after a key was stored to a cluster node, reading the location log was not getting the UUID of that node. Apparently the Annex action that wrote to the location log, and the one that read from it were run with two different Annex states. The http server does use several different Annex threads. BranchState was part of the AnnexState, and so two threads could have different BranchStates. Moved BranchState to the AnnexRead, so all threads will see the common state. This might possibly impact performance. If one thread is writing changes to the branch, and another thread is reading from the branch, the writing thread will now invalidate the BranchState's cache, which will cause the reading thread to need to do extra work. But correctness is surely more important. If did is found to have impacted performance, it could probably be dealt with by doing smarter BranchState cache invalidation. Another way this might impact performance is that the BranchState has a small cache. If several threads were reading from the branch and relying on the value they just read still being in the case, now a cache miss will be more likely. Increasing the BranchState cache to the number of jobs might be a good idea to amelorate that. But the cache is currently an innefficient list, so making it large would need changes to the data types. (Commit 4304f1b6aea19a5c402dc4f9d69aa4ff1c104c9b dealt with a follow-on effect of the bug fixed here.)
2024-07-28 16:17:16 +00:00
{ branchstate = bs
, activekeys = emptyactivekeys
, activeremotes = emptyactiveremotes
, keysdbhandle = kh
, sshstalecleaned = sc
, signalactions = si
, transferrerpool = tp
, debugenabled = annexDebug c
, debugselector = debugSelectorFromGitConfig c
, explainenabled = False
, ciphers = cm
, fast = False
, force = False
, forcebackend = Nothing
, forcenumcopies = Nothing
, forcemincopies = Nothing
, reposizes = rs
, rebalance = False
, useragent = Nothing
, desktopnotify = mempty
, gitcredentialcache = cc
}
-- Values that can change while running an Annex action.
data AnnexState = AnnexState
{ repo :: Git.Repo
, repoadjustment :: (Git.Repo -> IO Git.Repo)
, gitconfig :: GitConfig
, gitconfigadjustment :: (GitConfig -> GitConfig)
, gitconfigoverride :: [String]
, gitremotes :: Maybe [Git.Repo]
, gitconfiginodecache :: Maybe InodeCache
, backend :: Maybe (BackendA Annex)
2011-12-31 08:11:39 +00:00
, remotes :: [Types.Remote.RemoteA Annex]
, output :: MessageState
, concurrency :: ConcurrencySetting
, daemon :: Bool
, repoqueue :: Maybe (Git.Queue.Queue Annex)
, catfilehandles :: CatFileHandles
, hashobjecthandle :: Maybe (ResourcePool HashObjectHandle)
, checkattrhandle :: Maybe (ResourcePool CheckAttrHandle)
, checkignorehandle :: Maybe (ResourcePool CheckIgnoreHandle)
, globalnumcopies :: Maybe (Maybe NumCopies)
, globalmincopies :: Maybe (Maybe MinCopies)
, limit :: ExpandableMatcher Annex
, timelimit :: Maybe (Duration, POSIXTime)
, sizelimit :: Maybe (TVar Integer)
, uuiddescmap :: Maybe UUIDDescMap
, preferredcontentmap :: Maybe (FileMatcherMap Annex)
, requiredcontentmap :: Maybe (FileMatcherMap Annex)
, remoteconfigmap :: Maybe (M.Map UUID RemoteConfig)
, clusters :: Maybe (Annex Clusters)
, maxsizes :: Maybe (M.Map UUID MaxSize)
, forcetrust :: TrustMap
, trustmap :: Maybe TrustMap
, groupmap :: Maybe GroupMap
, lockcache :: LockCache
, fields :: M.Map String String
, cleanupactions :: M.Map CleanupAction (Annex ())
, sentinalstatus :: Maybe SentinalStatus
, errcounter :: Integer
, reachedlimit :: Bool
, adjustedbranchrefreshcounter :: Integer
, unusedkeys :: Maybe (S.Set Key)
, tempurls :: M.Map Key URLString
, existinghooks :: M.Map Git.Hook.Hook Bool
, workers :: Maybe (TMVar (WorkerPool (AnnexState, AnnexRead)))
, cachedcurrentbranch :: (Maybe (Maybe Git.Branch, Maybe Adjustment))
, cachedgitenv :: Maybe (AltIndexFile, FilePath, [(String, String)])
, urloptions :: Maybe UrlOptions
, insmudgecleanfilter :: Bool
deal better with clock skew situations, using vector clocks * Deal with clock skew, both forwards and backwards, when logging information to the git-annex branch. * GIT_ANNEX_VECTOR_CLOCK can now be set to a fixed value (eg 1) rather than needing to be advanced each time a new change is made. * Misuse of GIT_ANNEX_VECTOR_CLOCK will no longer confuse git-annex. When changing a file in the git-annex branch, the vector clock to use is now determined by first looking at the current time (or GIT_ANNEX_VECTOR_CLOCK when set), and comparing it to the newest vector clock already in use in that file. If a newer time stamp was already in use, advance it forward by a second instead. When the clock is set to a time in the past, this avoids logging with an old timestamp, which would risk that log line later being ignored in favor of "newer" line that is really not newer. When a log entry has been made with a clock that was set far ahead in the future, this avoids newer information being logged with an older timestamp and so being ignored in favor of that future-timestamped information. Once all clocks get fixed, this will result in the vector clocks being incremented, until finally enough time has passed that time gets back ahead of the vector clock value, and then it will return to usual operation. (This latter situation is not ideal, but it seems the best that can be done. The issue with it is, since all writers will be incrementing the last vector clock they saw, there's no way to tell when one writer made a write significantly later in time than another, so the earlier write might arbitrarily be picked when merging. This problem is why git-annex uses timestamps in the first place, rather than pure vector clocks.) Advancing forward by 1 second is somewhat arbitrary. setDead advances a timestamp by just 1 picosecond, and the vector clock could too. But then it would interfere with setDead, which wants to be overrulled by any change. So it could use 2 picoseconds or something, but that seems weird. It could just as well advance it forward by a minute or whatever, but then it would be harder for real time to catch up with the vector clock when forward clock slew had happened. A complication is that many log files contain several different peices of information, and it may be best to only use vector clocks for the same peice of information. For example, a key's location log file contains InfoPresent/InfoMissing for each UUID, and it only looks at the vector clocks for the UUID that is being changed, and not other UUIDs. Although exactly where the dividing line is can be hard to determine. Consider metadata logs, where a field "tag" can have multiple values set at different times. Should it advance forward past the last tag? Probably. What about when a different field is set, should it look at the clocks of other fields? Perhaps not, but currently it does, and this does not seems like it will cause any problems. Another one I'm not entirely sure about is the export log, which is keyed by (fromuuid, touuid). So if multiple repos are exporting to the same remote, different vector clocks can be used for that remote. It looks like that's probably ok, because it does not try to determine what order things occurred when there was an export conflict. Sponsored-by: Jochen Bartl on Patreon
2021-08-03 20:45:20 +00:00
, getvectorclock :: IO CandidateVectorClock
, proxyremote :: Maybe (Either ClusterUUID (Types.Remote.RemoteA Annex))
, reposizehandle :: Maybe RepoSizeHandle
2011-04-16 20:41:46 +00:00
}
newAnnexState :: GitConfig -> Git.Repo -> IO AnnexState
newAnnexState c r = do
o <- newMessageState
vc <- startVectorClock
return $ AnnexState
{ repo = r
, repoadjustment = return
, gitconfig = c
, gitconfigadjustment = id
, gitconfigoverride = []
, gitremotes = Nothing
, gitconfiginodecache = Nothing
, backend = Nothing
, remotes = []
, output = o
, concurrency = ConcurrencyCmdLine NonConcurrent
, daemon = False
, repoqueue = Nothing
, catfilehandles = catFileHandlesNonConcurrent
, hashobjecthandle = Nothing
, checkattrhandle = Nothing
, checkignorehandle = Nothing
, globalnumcopies = Nothing
, globalmincopies = Nothing
, limit = BuildingMatcher []
, timelimit = Nothing
, sizelimit = Nothing
, uuiddescmap = Nothing
, preferredcontentmap = Nothing
, requiredcontentmap = Nothing
, remoteconfigmap = Nothing
2024-06-14 15:16:01 +00:00
, clusters = Nothing
, maxsizes = Nothing
, forcetrust = M.empty
, trustmap = Nothing
, groupmap = Nothing
, lockcache = M.empty
, fields = M.empty
, cleanupactions = M.empty
, sentinalstatus = Nothing
, errcounter = 0
, reachedlimit = False
, adjustedbranchrefreshcounter = 0
, unusedkeys = Nothing
, tempurls = M.empty
, existinghooks = M.empty
, workers = Nothing
, cachedcurrentbranch = Nothing
, cachedgitenv = Nothing
, urloptions = Nothing
, insmudgecleanfilter = False
, getvectorclock = vc
git-annex-shell: accept uuid of remote that proxying is enabled for For NotifyChanges and also for the fallthrough case where git-annex-shell passes a command off to git-shell, proxying is currently ignored. So every remote that is accessed via a proxy will be treated as the same git repository. Every other command listed in cmdsMap will need to check if Annex.proxyremote is set, and if so handle the proxying appropriately. Probably only P2PStdio will need to support proxying. For now, everything else refuses to work when proxying. The part of that I don't like is that there's the possibility a command later gets added to the list that doesn't check proxying. When proxying is not enabled, it's important that git-annex-shell not leak information that it would not have exposed before. Such as the names or uuids of remotes. I decided that, in the case where a repository used to have proxying enabled, but no longer supports any proxies, it's ok to give the user a clear error message indicating that proxying is not configured, rather than a confusing uuid mismatch message. Similarly, if a repository has proxying enabled, but not for the requested repository, give a clear error message. A tricky thing here is how to handle the case where there is more than one remote, with proxying enabled, with the specified uuid. One way to handle that would be to plumb the proxyRemoteName all the way through from the remote git-annex to git-annex-shell, eg as a field, and use only a remote with the same name. That would be very intrusive though. Instead, I decided to let the proxy pick which remote it uses to access a given Remote. And so it picks the least expensive one. The client after all doesn't necessarily know any details about the proxy's configuration. This does mean though, that if the least expensive remote is not accessible, but another remote would have worked, an access via the proxy will fail.
2024-06-10 16:05:03 +00:00
, proxyremote = Nothing
, reposizehandle = Nothing
}
2010-10-14 07:18:11 +00:00
{- Makes an Annex state object for the specified git repo.
- Ensures the config is read, if it was not already, and performs
- any necessary git repo fixups. -}
new :: Git.Repo -> IO (AnnexState, AnnexRead)
new r = do
r' <- Git.Config.read r
let c = extractGitConfig FromGitConfig r'
st <- newAnnexState c =<< fixupRepo r' c
rd <- newAnnexRead c
return (st, rd)
2010-10-14 07:18:11 +00:00
Switch to MonadCatchIO-transformers for better handling of state while catching exceptions. As seen in this bug report, the lifted exception handling using the StateT monad throws away state changes when an action throws an exception. http://git-annex.branchable.com/bugs/git_annex_fork_bombs_on_gpg_file/ .. Which can result in cached values being redundantly calculated, or other possibly worse bugs when the annex state gets out of sync with reality. This switches from a StateT AnnexState to a ReaderT (MVar AnnexState). All changes to the state go via the MVar. So when an Annex action is running inside an exception handler, and it makes some changes, they immediately go into affect in the MVar. If it then throws an exception (or even crashes its thread!), the state changes are still in effect. The MonadCatchIO-transformers change is actually only incidental. I could have kept on using lifted-base for the exception handling. However, I'd have needed to write a new instance of MonadBaseControl for the new monad.. and I didn't write the old instance.. I begged Bas and he kindly sent it to me. Happily, MonadCatchIO-transformers is able to derive a MonadCatchIO instance for my monad. This is a deep level change. It passes the test suite! What could it break? Well.. The most likely breakage would be to code that runs an Annex action in an exception handler, and *wants* state changes to be thrown away. Perhaps the state changes leaves the state inconsistent, or wrong. Since there are relatively few places in git-annex that catch exceptions in the Annex monad, and the AnnexState is generally just used to cache calculated data, this is unlikely to be a problem. Oh yeah, this change also makes Assistant.Types.ThreadedMonad a bit redundant. It's now entirely possible to run concurrent Annex actions in different threads, all sharing access to the same state! The ThreadedMonad just adds some extra work on top of that, with its own MVar, and avoids such actions possibly stepping on one-another's toes. I have not gotten rid of it, but might try that later. Being able to run concurrent Annex actions would simplify parts of the Assistant code.
2013-05-19 18:16:36 +00:00
{- Performs an action in the Annex monad from a starting state,
- returning a new state. -}
run :: (AnnexState, AnnexRead) -> Annex a -> IO (a, (AnnexState, AnnexRead))
run (st, rd) a = do
mv <- newMVar st
run' mv rd a
run' :: MVar AnnexState -> AnnexRead -> Annex a -> IO (a, (AnnexState, AnnexRead))
run' mvar rd a = do
r <- runReaderT (runAnnex a) (mvar, rd)
st <- takeMVar mvar
return (r, (st, rd))
Switch to MonadCatchIO-transformers for better handling of state while catching exceptions. As seen in this bug report, the lifted exception handling using the StateT monad throws away state changes when an action throws an exception. http://git-annex.branchable.com/bugs/git_annex_fork_bombs_on_gpg_file/ .. Which can result in cached values being redundantly calculated, or other possibly worse bugs when the annex state gets out of sync with reality. This switches from a StateT AnnexState to a ReaderT (MVar AnnexState). All changes to the state go via the MVar. So when an Annex action is running inside an exception handler, and it makes some changes, they immediately go into affect in the MVar. If it then throws an exception (or even crashes its thread!), the state changes are still in effect. The MonadCatchIO-transformers change is actually only incidental. I could have kept on using lifted-base for the exception handling. However, I'd have needed to write a new instance of MonadBaseControl for the new monad.. and I didn't write the old instance.. I begged Bas and he kindly sent it to me. Happily, MonadCatchIO-transformers is able to derive a MonadCatchIO instance for my monad. This is a deep level change. It passes the test suite! What could it break? Well.. The most likely breakage would be to code that runs an Annex action in an exception handler, and *wants* state changes to be thrown away. Perhaps the state changes leaves the state inconsistent, or wrong. Since there are relatively few places in git-annex that catch exceptions in the Annex monad, and the AnnexState is generally just used to cache calculated data, this is unlikely to be a problem. Oh yeah, this change also makes Assistant.Types.ThreadedMonad a bit redundant. It's now entirely possible to run concurrent Annex actions in different threads, all sharing access to the same state! The ThreadedMonad just adds some extra work on top of that, with its own MVar, and avoids such actions possibly stepping on one-another's toes. I have not gotten rid of it, but might try that later. Being able to run concurrent Annex actions would simplify parts of the Assistant code.
2013-05-19 18:16:36 +00:00
{- Performs an action in the Annex monad from a starting state,
- and throws away the changed state. -}
eval :: (AnnexState, AnnexRead) -> Annex a -> IO a
eval v a = fst <$> run v a
Switch to MonadCatchIO-transformers for better handling of state while catching exceptions. As seen in this bug report, the lifted exception handling using the StateT monad throws away state changes when an action throws an exception. http://git-annex.branchable.com/bugs/git_annex_fork_bombs_on_gpg_file/ .. Which can result in cached values being redundantly calculated, or other possibly worse bugs when the annex state gets out of sync with reality. This switches from a StateT AnnexState to a ReaderT (MVar AnnexState). All changes to the state go via the MVar. So when an Annex action is running inside an exception handler, and it makes some changes, they immediately go into affect in the MVar. If it then throws an exception (or even crashes its thread!), the state changes are still in effect. The MonadCatchIO-transformers change is actually only incidental. I could have kept on using lifted-base for the exception handling. However, I'd have needed to write a new instance of MonadBaseControl for the new monad.. and I didn't write the old instance.. I begged Bas and he kindly sent it to me. Happily, MonadCatchIO-transformers is able to derive a MonadCatchIO instance for my monad. This is a deep level change. It passes the test suite! What could it break? Well.. The most likely breakage would be to code that runs an Annex action in an exception handler, and *wants* state changes to be thrown away. Perhaps the state changes leaves the state inconsistent, or wrong. Since there are relatively few places in git-annex that catch exceptions in the Annex monad, and the AnnexState is generally just used to cache calculated data, this is unlikely to be a problem. Oh yeah, this change also makes Assistant.Types.ThreadedMonad a bit redundant. It's now entirely possible to run concurrent Annex actions in different threads, all sharing access to the same state! The ThreadedMonad just adds some extra work on top of that, with its own MVar, and avoids such actions possibly stepping on one-another's toes. I have not gotten rid of it, but might try that later. Being able to run concurrent Annex actions would simplify parts of the Assistant code.
2013-05-19 18:16:36 +00:00
{- Makes a runner action, that allows diving into IO and from inside
- the IO action, running an Annex action. -}
makeRunner :: Annex (Annex a -> IO a)
makeRunner = do
(mvar, rd) <- ask
return $ \a -> do
(r, (s, _rd)) <- run' mvar rd a
putMVar mvar s
return r
getRead :: (AnnexRead -> v) -> Annex v
getRead selector = selector . snd <$> ask
Switch to MonadCatchIO-transformers for better handling of state while catching exceptions. As seen in this bug report, the lifted exception handling using the StateT monad throws away state changes when an action throws an exception. http://git-annex.branchable.com/bugs/git_annex_fork_bombs_on_gpg_file/ .. Which can result in cached values being redundantly calculated, or other possibly worse bugs when the annex state gets out of sync with reality. This switches from a StateT AnnexState to a ReaderT (MVar AnnexState). All changes to the state go via the MVar. So when an Annex action is running inside an exception handler, and it makes some changes, they immediately go into affect in the MVar. If it then throws an exception (or even crashes its thread!), the state changes are still in effect. The MonadCatchIO-transformers change is actually only incidental. I could have kept on using lifted-base for the exception handling. However, I'd have needed to write a new instance of MonadBaseControl for the new monad.. and I didn't write the old instance.. I begged Bas and he kindly sent it to me. Happily, MonadCatchIO-transformers is able to derive a MonadCatchIO instance for my monad. This is a deep level change. It passes the test suite! What could it break? Well.. The most likely breakage would be to code that runs an Annex action in an exception handler, and *wants* state changes to be thrown away. Perhaps the state changes leaves the state inconsistent, or wrong. Since there are relatively few places in git-annex that catch exceptions in the Annex monad, and the AnnexState is generally just used to cache calculated data, this is unlikely to be a problem. Oh yeah, this change also makes Assistant.Types.ThreadedMonad a bit redundant. It's now entirely possible to run concurrent Annex actions in different threads, all sharing access to the same state! The ThreadedMonad just adds some extra work on top of that, with its own MVar, and avoids such actions possibly stepping on one-another's toes. I have not gotten rid of it, but might try that later. Being able to run concurrent Annex actions would simplify parts of the Assistant code.
2013-05-19 18:16:36 +00:00
getState :: (AnnexState -> v) -> Annex v
getState selector = do
mvar <- fst <$> ask
st <- liftIO $ readMVar mvar
return $ selector st
Switch to MonadCatchIO-transformers for better handling of state while catching exceptions. As seen in this bug report, the lifted exception handling using the StateT monad throws away state changes when an action throws an exception. http://git-annex.branchable.com/bugs/git_annex_fork_bombs_on_gpg_file/ .. Which can result in cached values being redundantly calculated, or other possibly worse bugs when the annex state gets out of sync with reality. This switches from a StateT AnnexState to a ReaderT (MVar AnnexState). All changes to the state go via the MVar. So when an Annex action is running inside an exception handler, and it makes some changes, they immediately go into affect in the MVar. If it then throws an exception (or even crashes its thread!), the state changes are still in effect. The MonadCatchIO-transformers change is actually only incidental. I could have kept on using lifted-base for the exception handling. However, I'd have needed to write a new instance of MonadBaseControl for the new monad.. and I didn't write the old instance.. I begged Bas and he kindly sent it to me. Happily, MonadCatchIO-transformers is able to derive a MonadCatchIO instance for my monad. This is a deep level change. It passes the test suite! What could it break? Well.. The most likely breakage would be to code that runs an Annex action in an exception handler, and *wants* state changes to be thrown away. Perhaps the state changes leaves the state inconsistent, or wrong. Since there are relatively few places in git-annex that catch exceptions in the Annex monad, and the AnnexState is generally just used to cache calculated data, this is unlikely to be a problem. Oh yeah, this change also makes Assistant.Types.ThreadedMonad a bit redundant. It's now entirely possible to run concurrent Annex actions in different threads, all sharing access to the same state! The ThreadedMonad just adds some extra work on top of that, with its own MVar, and avoids such actions possibly stepping on one-another's toes. I have not gotten rid of it, but might try that later. Being able to run concurrent Annex actions would simplify parts of the Assistant code.
2013-05-19 18:16:36 +00:00
changeState :: (AnnexState -> AnnexState) -> Annex ()
changeState modifier = do
mvar <- fst <$> ask
Switch to MonadCatchIO-transformers for better handling of state while catching exceptions. As seen in this bug report, the lifted exception handling using the StateT monad throws away state changes when an action throws an exception. http://git-annex.branchable.com/bugs/git_annex_fork_bombs_on_gpg_file/ .. Which can result in cached values being redundantly calculated, or other possibly worse bugs when the annex state gets out of sync with reality. This switches from a StateT AnnexState to a ReaderT (MVar AnnexState). All changes to the state go via the MVar. So when an Annex action is running inside an exception handler, and it makes some changes, they immediately go into affect in the MVar. If it then throws an exception (or even crashes its thread!), the state changes are still in effect. The MonadCatchIO-transformers change is actually only incidental. I could have kept on using lifted-base for the exception handling. However, I'd have needed to write a new instance of MonadBaseControl for the new monad.. and I didn't write the old instance.. I begged Bas and he kindly sent it to me. Happily, MonadCatchIO-transformers is able to derive a MonadCatchIO instance for my monad. This is a deep level change. It passes the test suite! What could it break? Well.. The most likely breakage would be to code that runs an Annex action in an exception handler, and *wants* state changes to be thrown away. Perhaps the state changes leaves the state inconsistent, or wrong. Since there are relatively few places in git-annex that catch exceptions in the Annex monad, and the AnnexState is generally just used to cache calculated data, this is unlikely to be a problem. Oh yeah, this change also makes Assistant.Types.ThreadedMonad a bit redundant. It's now entirely possible to run concurrent Annex actions in different threads, all sharing access to the same state! The ThreadedMonad just adds some extra work on top of that, with its own MVar, and avoids such actions possibly stepping on one-another's toes. I have not gotten rid of it, but might try that later. Being able to run concurrent Annex actions would simplify parts of the Assistant code.
2013-05-19 18:16:36 +00:00
liftIO $ modifyMVar_ mvar $ return . modifier
2010-10-14 07:18:11 +00:00
withState :: (AnnexState -> IO (AnnexState, b)) -> Annex b
withState modifier = do
mvar <- fst <$> ask
liftIO $ modifyMVar mvar modifier
{- Sets a field to a value -}
setField :: String -> String -> Annex ()
setField field value = changeState $ \st ->
st { fields = M.insert field value $ fields st }
{- Adds a cleanup action to perform. -}
addCleanupAction :: CleanupAction -> Annex () -> Annex ()
addCleanupAction k a = changeState $ \st ->
st { cleanupactions = M.insert k a $ cleanupactions st }
2012-04-30 17:59:05 +00:00
{- Sets the type of output to emit. -}
setOutput :: OutputType -> Annex ()
setOutput o = changeState $ \st ->
let m = output st
in st { output = m { outputType = adjustOutputType (outputType m) o } }
2012-04-30 17:59:05 +00:00
{- Gets the value of a field. -}
getField :: String -> Annex (Maybe String)
getField field = M.lookup field <$> getState fields
{- Returns the annex's git repository. -}
2010-10-14 07:18:11 +00:00
gitRepo :: Annex Git.Repo
gitRepo = getState repo
{- Runs an IO action in the annex's git repository. -}
inRepo :: (Git.Repo -> IO a) -> Annex a
2011-11-12 18:24:07 +00:00
inRepo a = liftIO . a =<< gitRepo
{- Extracts a value from the annex's git repisitory. -}
fromRepo :: (Git.Repo -> a) -> Annex a
fromRepo a = a <$> gitRepo
Switch to MonadCatchIO-transformers for better handling of state while catching exceptions. As seen in this bug report, the lifted exception handling using the StateT monad throws away state changes when an action throws an exception. http://git-annex.branchable.com/bugs/git_annex_fork_bombs_on_gpg_file/ .. Which can result in cached values being redundantly calculated, or other possibly worse bugs when the annex state gets out of sync with reality. This switches from a StateT AnnexState to a ReaderT (MVar AnnexState). All changes to the state go via the MVar. So when an Annex action is running inside an exception handler, and it makes some changes, they immediately go into affect in the MVar. If it then throws an exception (or even crashes its thread!), the state changes are still in effect. The MonadCatchIO-transformers change is actually only incidental. I could have kept on using lifted-base for the exception handling. However, I'd have needed to write a new instance of MonadBaseControl for the new monad.. and I didn't write the old instance.. I begged Bas and he kindly sent it to me. Happily, MonadCatchIO-transformers is able to derive a MonadCatchIO instance for my monad. This is a deep level change. It passes the test suite! What could it break? Well.. The most likely breakage would be to code that runs an Annex action in an exception handler, and *wants* state changes to be thrown away. Perhaps the state changes leaves the state inconsistent, or wrong. Since there are relatively few places in git-annex that catch exceptions in the Annex monad, and the AnnexState is generally just used to cache calculated data, this is unlikely to be a problem. Oh yeah, this change also makes Assistant.Types.ThreadedMonad a bit redundant. It's now entirely possible to run concurrent Annex actions in different threads, all sharing access to the same state! The ThreadedMonad just adds some extra work on top of that, with its own MVar, and avoids such actions possibly stepping on one-another's toes. I have not gotten rid of it, but might try that later. Being able to run concurrent Annex actions would simplify parts of the Assistant code.
2013-05-19 18:16:36 +00:00
{- Calculates a value from an annex's git repository and its GitConfig. -}
calcRepo :: (Git.Repo -> GitConfig -> IO a) -> Annex a
calcRepo a = do
s <- getState id
liftIO $ a (repo s) (gitconfig s)
calcRepo' :: (Git.Repo -> GitConfig -> a) -> Annex a
calcRepo' f = do
s <- getState id
pure $ f (repo s) (gitconfig s)
{- Gets the GitConfig settings. -}
getGitConfig :: Annex GitConfig
getGitConfig = getState gitconfig
{- Overrides a GitConfig setting. The modification persists across
- reloads of the repo's config. -}
overrideGitConfig :: (GitConfig -> GitConfig) -> Annex ()
overrideGitConfig f = changeState $ \st -> st
{ gitconfigadjustment = gitconfigadjustment st . f
, gitconfig = f (gitconfig st)
}
{- Adds an adjustment to the Repo data. Adjustments persist across reloads
- of the repo's config.
-
- Note that the action may run more than once, and should avoid eg,
- appending the same value to a repo's config when run repeatedly.
-}
adjustGitRepo :: (Git.Repo -> IO Git.Repo) -> Annex ()
adjustGitRepo a = do
changeState $ \st -> st { repoadjustment = \r -> repoadjustment st r >>= a }
changeGitRepo =<< gitRepo
{- Adds git config setting, like "foo=bar". It will be passed with -c
- to git processes. The config setting is also recorded in the Repo,
- and the GitConfig is updated. -}
addGitConfigOverride :: String -> Annex ()
addGitConfigOverride v = do
adjustGitRepo $ \r ->
Git.Config.store (encodeBS v) Git.Config.ConfigList $
r { Git.gitGlobalOpts = go (Git.gitGlobalOpts r) }
changeState $ \st -> st { gitconfigoverride = v : gitconfigoverride st }
where
2023-03-14 02:39:16 +00:00
-- Remove any prior occurrence of the setting to avoid
-- building up many of them when the adjustment is run repeatedly,
-- and add the setting to the end.
go [] = [Param "-c", Param v]
go (Param "-c": Param v':rest) | v' == v = go rest
go (c:rest) = c : go rest
{- Values that were passed to addGitConfigOverride. -}
getGitConfigOverrides :: Annex [String]
getGitConfigOverrides = reverse <$> getState gitconfigoverride
{- Changing the git Repo data also involves re-extracting its GitConfig. -}
changeGitRepo :: Git.Repo -> Annex ()
changeGitRepo r = do
repoadjuster <- getState repoadjustment
gitconfigadjuster <- getState gitconfigadjustment
r' <- liftIO $ repoadjuster r
changeState $ \st -> st
{ repo = r'
, gitconfig = gitconfigadjuster $
extractGitConfig FromGitConfig r'
, gitremotes = Nothing
}
2014-05-16 20:08:20 +00:00
{- Gets the RemoteGitConfig from a remote, given the Git.Repo for that
- remote. -}
getRemoteGitConfig :: Git.Repo -> Annex RemoteGitConfig
getRemoteGitConfig r = do
g <- gitRepo
liftIO $ atomically $ extractRemoteGitConfig g (Git.repoDescribe r)
2014-05-16 20:08:20 +00:00
{- Converts an Annex action into an IO action, that runs with a copy
- of the current Annex state.
-
- Use with caution; the action should not rely on changing the
- state, as it will be thrown away. -}
withCurrentState :: Annex a -> Annex (IO a)
withCurrentState a = do
(mvar, rd) <- ask
st <- liftIO $ readMVar mvar
return $ eval (st, rd) a
{- It's not safe to use setCurrentDirectory in the Annex monad,
- because the git repo paths are stored relative.
- Instead, use this.
-}
changeDirectory :: FilePath -> Annex ()
changeDirectory d = do
r <- liftIO . Git.adjustPath absPath =<< gitRepo
liftIO $ setCurrentDirectory d
r' <- liftIO $ Git.relPath r
changeState $ \st -> st { repo = r' }
2015-04-30 19:04:01 +00:00
incError :: Annex ()
incError = changeState $ \st ->
let !c = errcounter st + 1
!st' = st { errcounter = c }
in st'
getGitRemotes :: Annex [Git.Repo]
getGitRemotes = do
st <- getState id
case gitremotes st of
Just rs -> return rs
Nothing -> do
rs <- liftIO $ Git.Construct.fromRemotes (repo st)
changeState $ \st' -> st' { gitremotes = Just rs }
return rs