git-annex/Command/Drop.hs

233 lines
7.1 KiB
Haskell
Raw Normal View History

{- git-annex command
-
- Copyright 2010-2020 Joey Hess <id@joeyh.name>
-
- Licensed under the GNU AGPL version 3 or higher.
-}
module Command.Drop where
import Command
import qualified Remote
import qualified Annex
import Annex.UUID
2011-10-15 20:21:08 +00:00
import Logs.Location
import Logs.Trust
import Logs.PreferredContent
2015-04-30 18:02:56 +00:00
import Annex.NumCopies
2011-10-04 04:40:47 +00:00
import Annex.Content
import Annex.Wanted
2014-03-22 19:01:48 +00:00
import Annex.Notification
import System.Log.Logger (debugM)
import qualified Data.Set as S
cmd :: Command
cmd = withGlobalOptions [jobsOption, jsonOptions, annexedMatchingOptions] $
command "drop" SectionCommon
"remove content of files from repository"
paramPaths (seek <$$> optParser)
2015-07-08 21:59:06 +00:00
data DropOptions = DropOptions
{ dropFiles :: CmdParams
, dropFrom :: Maybe (DeferredParse Remote)
2015-07-08 21:59:06 +00:00
, autoMode :: Bool
, keyOptions :: Maybe KeyOptions
2016-07-06 15:54:46 +00:00
, batchOption :: BatchMode
2015-07-08 21:59:06 +00:00
}
optParser :: CmdParamsDesc -> Parser DropOptions
optParser desc = DropOptions
<$> cmdParams desc
<*> optional parseDropFromOption
2015-07-08 21:59:06 +00:00
<*> parseAutoOption
<*> optional parseKeyOptions
2016-07-06 15:54:46 +00:00
<*> parseBatchOption
2015-07-08 21:59:06 +00:00
parseDropFromOption :: Parser (DeferredParse Remote)
parseDropFromOption = parseRemoteOption <$> strOption
2015-07-09 14:41:17 +00:00
( long "from" <> short 'f' <> metavar paramRemote
<> help "drop content from a remote"
<> completeRemotes
2015-07-08 21:59:06 +00:00
)
seek :: DropOptions -> CommandSeek
seek o = startConcurrency commandStages $ do
from <- case dropFrom o of
Nothing -> pure Nothing
Just f -> getParsed f >>= \remote -> do
u <- getUUID
if Remote.uuid remote == u
then pure Nothing
else pure (Just remote)
let seeker = AnnexedFileSeeker
{ startAction = start o from
, checkContentPresent = case from of
Nothing -> Just True
Just _ -> Nothing
, usesLocationLog = True
}
2016-07-06 15:54:46 +00:00
case batchOption o of
Batch fmt -> batchAnnexedFilesMatching fmt seeker
NoBatch -> withKeyOptions (keyOptions o) (autoMode o) seeker
(commandAction . startKeys o from)
(withFilesInGitAnnex ww seeker)
=<< workTreeItems ww (dropFiles o)
2016-07-06 15:54:46 +00:00
where
ww = WarnUnmatchLsFiles
2015-07-08 21:59:06 +00:00
start :: DropOptions -> Maybe Remote -> RawFilePath -> Key -> CommandStart
start o from file key = start' o from key afile ai
where
afile = AssociatedFile (Just file)
2019-06-06 16:53:24 +00:00
ai = mkActionItem (key, afile)
2015-07-08 21:59:06 +00:00
start' :: DropOptions -> Maybe Remote -> Key -> AssociatedFile -> ActionItem -> CommandStart
start' o from key afile ai =
2015-07-08 21:59:06 +00:00
checkDropAuto (autoMode o) from afile key $ \numcopies ->
stopUnless want $
2015-07-08 21:59:06 +00:00
case from of
Nothing -> startLocal afile ai numcopies key []
Just remote -> startRemote afile ai numcopies key remote
where
want
| autoMode o = wantDrop False (Remote.uuid <$> from) (Just key) afile
| otherwise = return True
2015-07-08 21:59:06 +00:00
startKeys :: DropOptions -> Maybe Remote -> (Key, ActionItem) -> CommandStart
startKeys o from (key, ai) = start' o from key (AssociatedFile Nothing) ai
startLocal :: AssociatedFile -> ActionItem -> NumCopies -> Key -> [VerifiedCopy] -> CommandStart
make CommandStart return a StartMessage The goal is to be able to run CommandStart in the main thread when -J is used, rather than unncessarily passing it off to a worker thread, which incurs overhead that is signficant when the CommandStart is going to quickly decide to stop. To do that, the message it displays needs to be displayed in the worker thread, after the CommandStart has run. Also, the change will mean that CommandStart will no longer necessarily run with the same Annex state as CommandPerform. While its docs already said it should avoid modifying Annex state, I audited all the CommandStart code as part of the conversion. (Note that CommandSeek already sometimes runs with a different Annex state, and that has not been a source of any problems, so I am not too worried that this change will lead to breakage going forward.) The only modification of Annex state I found was it calling allowMessages in some Commands that default to noMessages. Dealt with that by adding a startCustomOutput and a startingUsualMessages. This lets a command start with noMessages and then select the output it wants for each CommandStart. One bit of breakage: onlyActionOn has been removed from commands that used it. The plan is that, since a StartMessage contains an ActionItem, when a Key can be extracted from that, the parallel job runner can run onlyActionOn' automatically. Then commands won't need to worry about this detail. Future work. Otherwise, this was a fairly straightforward process of making each CommandStart compile again. Hopefully other behavior changes were mostly avoided. In a few cases, a command had a CommandStart that called a CommandPerform that then called showStart multiple times. I have collapsed those down to a single start action. The main command to perhaps suffer from it is Command.Direct, which used to show a start for each file, and no longer does. Another minor behavior change is that some commands used showStart before, but had an associated file and a Key available, so were changed to ShowStart with an ActionItemAssociatedFile. That will not change the normal output or behavior, but --json output will now include the key. This should not break it for anyone using a real json parser.
2019-06-06 19:42:30 +00:00
startLocal afile ai numcopies key preverified =
starting "drop" (OnlyActionOn key ai) $
performLocal key afile numcopies preverified
startRemote :: AssociatedFile -> ActionItem -> NumCopies -> Key -> Remote -> CommandStart
make CommandStart return a StartMessage The goal is to be able to run CommandStart in the main thread when -J is used, rather than unncessarily passing it off to a worker thread, which incurs overhead that is signficant when the CommandStart is going to quickly decide to stop. To do that, the message it displays needs to be displayed in the worker thread, after the CommandStart has run. Also, the change will mean that CommandStart will no longer necessarily run with the same Annex state as CommandPerform. While its docs already said it should avoid modifying Annex state, I audited all the CommandStart code as part of the conversion. (Note that CommandSeek already sometimes runs with a different Annex state, and that has not been a source of any problems, so I am not too worried that this change will lead to breakage going forward.) The only modification of Annex state I found was it calling allowMessages in some Commands that default to noMessages. Dealt with that by adding a startCustomOutput and a startingUsualMessages. This lets a command start with noMessages and then select the output it wants for each CommandStart. One bit of breakage: onlyActionOn has been removed from commands that used it. The plan is that, since a StartMessage contains an ActionItem, when a Key can be extracted from that, the parallel job runner can run onlyActionOn' automatically. Then commands won't need to worry about this detail. Future work. Otherwise, this was a fairly straightforward process of making each CommandStart compile again. Hopefully other behavior changes were mostly avoided. In a few cases, a command had a CommandStart that called a CommandPerform that then called showStart multiple times. I have collapsed those down to a single start action. The main command to perhaps suffer from it is Command.Direct, which used to show a start for each file, and no longer does. Another minor behavior change is that some commands used showStart before, but had an associated file and a Key available, so were changed to ShowStart with an ActionItemAssociatedFile. That will not change the normal output or behavior, but --json output will now include the key. This should not break it for anyone using a real json parser.
2019-06-06 19:42:30 +00:00
startRemote afile ai numcopies key remote =
starting ("drop " ++ Remote.name remote) (OnlyActionOn key ai) $
make CommandStart return a StartMessage The goal is to be able to run CommandStart in the main thread when -J is used, rather than unncessarily passing it off to a worker thread, which incurs overhead that is signficant when the CommandStart is going to quickly decide to stop. To do that, the message it displays needs to be displayed in the worker thread, after the CommandStart has run. Also, the change will mean that CommandStart will no longer necessarily run with the same Annex state as CommandPerform. While its docs already said it should avoid modifying Annex state, I audited all the CommandStart code as part of the conversion. (Note that CommandSeek already sometimes runs with a different Annex state, and that has not been a source of any problems, so I am not too worried that this change will lead to breakage going forward.) The only modification of Annex state I found was it calling allowMessages in some Commands that default to noMessages. Dealt with that by adding a startCustomOutput and a startingUsualMessages. This lets a command start with noMessages and then select the output it wants for each CommandStart. One bit of breakage: onlyActionOn has been removed from commands that used it. The plan is that, since a StartMessage contains an ActionItem, when a Key can be extracted from that, the parallel job runner can run onlyActionOn' automatically. Then commands won't need to worry about this detail. Future work. Otherwise, this was a fairly straightforward process of making each CommandStart compile again. Hopefully other behavior changes were mostly avoided. In a few cases, a command had a CommandStart that called a CommandPerform that then called showStart multiple times. I have collapsed those down to a single start action. The main command to perhaps suffer from it is Command.Direct, which used to show a start for each file, and no longer does. Another minor behavior change is that some commands used showStart before, but had an associated file and a Key available, so were changed to ShowStart with an ActionItemAssociatedFile. That will not change the normal output or behavior, but --json output will now include the key. This should not break it for anyone using a real json parser.
2019-06-06 19:42:30 +00:00
performRemote key afile numcopies remote
performLocal :: Key -> AssociatedFile -> NumCopies -> [VerifiedCopy] -> CommandPerform
performLocal key afile numcopies preverified = lockContentForRemoval key fallback $ \contentlock -> do
u <- getUUID
(tocheck, verified) <- verifiableCopies key [u]
doDrop u (Just contentlock) key afile numcopies [] (preverified ++ verified) tocheck
( \proof -> do
liftIO $ debugM "drop" $ unwords
[ "Dropping from here"
, "proof:"
, show proof
]
removeAnnex contentlock
2014-03-22 19:01:48 +00:00
notifyDrop afile True
next $ cleanupLocal key
2015-10-09 15:09:46 +00:00
, do
2014-03-22 19:01:48 +00:00
notifyDrop afile False
stop
)
where
-- This occurs when, for example, two files are being dropped
-- and have the same content. The seek stage checks if the content
-- is present, but due to buffering, may find it present for the
-- second file before the first is dropped. If so, nothing remains
-- to be done except for cleaning up.
fallback = next $ cleanupLocal key
performRemote :: Key -> AssociatedFile -> NumCopies -> Remote -> CommandPerform
performRemote key afile numcopies remote = do
2019-10-21 17:51:38 +00:00
-- Filter the uuid it's being dropped from out of the lists of
-- places assumed to have the key, and places to check.
(tocheck, verified) <- verifiableCopies key [uuid]
doDrop uuid Nothing key afile numcopies [uuid] verified tocheck
( \proof -> do
liftIO $ debugM "drop" $ unwords
[ "Dropping from remote"
, show remote
, "proof:"
, show proof
]
2020-05-14 18:19:28 +00:00
ok <- Remote.action (Remote.removeKey remote key)
2015-10-09 15:09:46 +00:00
next $ cleanupRemote key remote ok
, stop
)
2012-12-13 04:45:27 +00:00
where
2012-11-12 05:05:04 +00:00
uuid = Remote.uuid remote
cleanupLocal :: Key -> CommandCleanup
cleanupLocal key = do
logStatus key InfoMissing
return True
2011-12-31 08:11:39 +00:00
cleanupRemote :: Key -> Remote -> Bool -> CommandCleanup
cleanupRemote key remote ok = do
when ok $
Remote.logStatus remote key InfoMissing
return ok
2015-10-09 15:09:46 +00:00
{- Before running the dropaction, checks specified remotes to
- verify that enough copies of a key exist to allow it to be
- safely removed (with no data loss).
-
- Also checks if it's required content, and refuses to drop if so.
-
- --force overrides and always allows dropping.
-}
doDrop
:: UUID
-> Maybe ContentRemovalLock
-> Key
-> AssociatedFile
-> NumCopies
-> [UUID]
-> [VerifiedCopy]
-> [UnVerifiedCopy]
-> (Maybe SafeDropProof -> CommandPerform, CommandPerform)
-> CommandPerform
doDrop dropfrom contentlock key afile numcopies skip preverified check (dropaction, nodropaction) =
2015-04-30 18:02:56 +00:00
ifM (Annex.getState Annex.force)
( dropaction Nothing
2015-10-09 15:09:46 +00:00
, ifM (checkRequiredContent dropfrom key afile)
( verifyEnoughCopiesToDrop nolocmsg key
contentlock numcopies
skip preverified check
(dropaction . Just)
(forcehint nodropaction)
2015-10-09 15:09:46 +00:00
, stop
)
2015-04-30 18:02:56 +00:00
)
2012-11-12 05:05:04 +00:00
where
2015-04-30 18:02:56 +00:00
nolocmsg = "Rather than dropping this file, try using: git annex move"
2015-10-09 15:09:46 +00:00
forcehint a = do
showLongNote "(Use --force to override this check, or adjust numcopies.)"
a
checkRequiredContent :: UUID -> Key -> AssociatedFile -> Annex Bool
checkRequiredContent u k afile =
ifM (isRequiredContent (Just u) S.empty (Just k) afile False)
( requiredContent
, return True
)
requiredContent :: Annex Bool
requiredContent = do
showLongNote "That file is required content, it cannot be dropped!"
showLongNote "(Use --force to override this check, or adjust required content configuration.)"
return False
{- In auto mode, only runs the action if there are enough
- copies on other semitrusted repositories. -}
checkDropAuto :: Bool -> Maybe Remote -> AssociatedFile -> Key -> (NumCopies -> CommandStart) -> CommandStart
checkDropAuto automode mremote afile key a =
go =<< getAssociatedFileNumCopies afile
where
go numcopies
2015-07-08 21:59:06 +00:00
| automode = do
locs <- Remote.keyLocations key
uuid <- getUUID
let remoteuuid = fromMaybe uuid $ Remote.uuid <$> mremote
locs' <- trustExclude UnTrusted $ filter (/= remoteuuid) locs
if NumCopies (length locs') >= numcopies
then a numcopies
else stop
| otherwise = a numcopies