git-annex/Command.hs

159 lines
5.9 KiB
Haskell
Raw Normal View History

2011-09-15 20:57:02 +00:00
{- git-annex command infrastructure
-
- Copyright 2010-2019 Joey Hess <id@joeyh.name>
-
- Licensed under the GNU AGPL version 3 or higher.
-}
module Command (
2016-01-21 17:14:38 +00:00
module Command,
2011-11-11 05:52:58 +00:00
module ReExported
) where
import Annex.Common as ReExported
2011-11-11 05:52:58 +00:00
import Types.Command as ReExported
2015-07-09 20:20:30 +00:00
import Types.DeferredParse as ReExported
import CmdLine.Seek as ReExported
2014-01-26 20:25:55 +00:00
import CmdLine.Usage as ReExported
import CmdLine.Action as ReExported
2014-01-26 20:25:55 +00:00
import CmdLine.Option as ReExported
import CmdLine.AnnexSetter as ReExported
2014-01-26 20:25:55 +00:00
import CmdLine.GitAnnex.Options as ReExported
import CmdLine.Batch as ReExported
import Options.Applicative as ReExported hiding (command)
import Annex.RepoSize.LiveUpdate as ReExported
2016-01-21 17:14:38 +00:00
import qualified Git
import Annex.Init
remove dead nodes when loading the cluster log This is to avoid inserting a cluster uuid into the location log when only dead nodes in the cluster contain the content of a key. One reason why this is necessary is Remote.keyLocations, which excludes dead repositories from the list. But there are probably many more. Implementing this was challenging, because Logs.Location importing Logs.Cluster which imports Logs.Trust which imports Remote.List resulted in an import cycle through several other modules. Resorted to making Logs.Location not import Logs.Cluster, and instead it assumes that Annex.clusters gets populated when necessary before it's called. That's done in Annex.Startup, which is run by the git-annex command (but not other commands) at early startup in initialized repos. Or, is run after initialization. Note that is Remote.Git, it is unable to import Annex.Startup, because Remote.Git importing Logs.Cluster leads the the same import cycle. So ensureInitialized is not passed annexStartup in there. Other commands, like git-annex-shell currently don't run annexStartup either. So there are cases where Logs.Location will not see clusters. So it won't add any cluster UUIDs when loading the log. That's ok, the only reason to do that is to make display of where objects are located include clusters, and to make commands like git-annex get --from treat keys as being located in a cluster. git-annex-shell certainly does not do anything like that, and I'm pretty sure Remote.Git (and callers to Remote.Git.onLocalRepo) don't either.
2024-06-16 18:35:07 +00:00
import Annex.Startup
2016-01-21 17:14:38 +00:00
import Utility.Daemon
import Types.Transfer
import Types.ActionItem as ReExported
import Types.WorkerPool as ReExported
import Remote.List
{- Generates a normal Command -}
command :: String -> CommandSection -> String -> CmdParamsDesc -> (CmdParamsDesc -> CommandParser) -> Command
command name section desc paramdesc mkparser =
Command commonChecks False False name paramdesc
section desc (mkparser paramdesc) mempty [] Nothing
{- Simple option parser that takes all non-option params as-is. -}
2016-01-21 17:14:38 +00:00
withParams :: (CmdParams -> v) -> CmdParamsDesc -> Parser v
withParams mkseek paramdesc = mkseek <$> cmdParams paramdesc
withParams' :: (CmdParams -> v) -> Mod ArgumentFields String -> String -> Parser v
withParams' mkseek completers paramdesc = mkseek
<$> cmdParamsWithCompleter paramdesc completers
2015-07-09 20:05:45 +00:00
{- Uses the supplied option parser, which yields a deferred parse,
- and calls finishParse on the result before passing it to the
- CommandSeek constructor. -}
(<--<) :: DeferredParseClass a
=> (a -> CommandSeek)
-> (CmdParamsDesc -> Parser a)
-> CmdParamsDesc
-> Parser CommandSeek
(<--<) mkseek optparser paramsdesc =
(mkseek <=< finishParse) <$> optparser paramsdesc
2012-09-16 00:46:38 +00:00
{- Indicates that a command doesn't need to commit any changes to
- the git-annex branch. -}
noCommit :: Command -> Command
noCommit c = c { cmdnocommit = True }
{- Indicates that a command should not output the usual messages when
- starting or stopping processing a file or other item. Unless --json mode
- is enabled, this also enables quiet output mode, so only things
- explicitly output by the command are shown and not progress messages
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
- etc.
-}
noMessages :: Command -> Command
noMessages c = c { cmdnomessages = True }
{- Adds a fallback action to a command, that will be run if it's used
- outside a git repository. -}
noRepo :: (CmdParamsDesc -> Parser (IO ())) -> Command -> Command
noRepo a c = c { cmdnorepo = Just (a (cmdparamdesc c)) }
2010-11-04 17:28:49 +00:00
{- Adds Annex options to a command. -}
withAnnexOptions :: [[AnnexOption]] -> Command -> Command
withAnnexOptions os c = c { cmdannexoptions = cmdannexoptions c ++ concat os }
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
{- For start stage to indicate what will be done. -}
starting:: MkActionItem actionitem => String -> actionitem -> SeekInput -> CommandPerform -> CommandStart
starting msg ai si a = next
(StartMessage msg (mkActionItem ai) si, a)
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
{- Use when noMessages was used but the command is going to output
- usual messages after all. -}
startingUsualMessages :: MkActionItem t => String -> t -> SeekInput -> CommandPerform -> CommandStart
startingUsualMessages msg t si a = next
(StartUsualMessages msg (mkActionItem t) si, a)
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
{- When no message should be displayed at start/end, but messages can still
- be displayed when using eg includeCommandAction. -}
startingNoMessage :: MkActionItem t => t -> CommandPerform -> CommandStart
startingNoMessage t a = next (StartNoMessage (mkActionItem t), a)
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
{- For commands that do not display usual start or end messages,
- but have some other custom output. -}
startingCustomOutput :: MkActionItem t => t -> CommandPerform -> CommandStart
startingCustomOutput t a = next (CustomOutput (mkActionItem t), a)
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
{- For perform stage to indicate what step to run next. -}
2011-05-15 06:02:46 +00:00
next :: a -> Annex (Maybe a)
next a = return $ Just a
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
{- For start and perform stage to indicate nothing needs to be done. -}
2011-05-15 06:02:46 +00:00
stop :: Annex (Maybe a)
stop = return Nothing
{- Stops unless a condition is met. -}
stopUnless :: Annex Bool -> Annex (Maybe a) -> Annex (Maybe a)
stopUnless c a = ifM c ( a , stop )
{- When doing a dry run, avoid actually performing the action, but pretend
- that it succeeded. -}
skipWhenDryRun :: DryRun -> CommandPerform -> CommandPerform
skipWhenDryRun (DryRun False) a = a
skipWhenDryRun (DryRun True) _ = next $ return True
{- When acting on a failed transfer, stops unless it was in the specified
- direction. -}
checkFailedTransferDirection :: ActionItem -> Direction -> Annex (Maybe a) -> Annex (Maybe a)
checkFailedTransferDirection ai d = stopUnless (pure check)
where
check = case actionItemTransferDirection ai of
Nothing -> True
Just d' -> d' == d
2016-01-21 17:14:38 +00:00
commonChecks :: [CommandCheck]
commonChecks = [repoExists]
repoExists :: CommandCheck
remove dead nodes when loading the cluster log This is to avoid inserting a cluster uuid into the location log when only dead nodes in the cluster contain the content of a key. One reason why this is necessary is Remote.keyLocations, which excludes dead repositories from the list. But there are probably many more. Implementing this was challenging, because Logs.Location importing Logs.Cluster which imports Logs.Trust which imports Remote.List resulted in an import cycle through several other modules. Resorted to making Logs.Location not import Logs.Cluster, and instead it assumes that Annex.clusters gets populated when necessary before it's called. That's done in Annex.Startup, which is run by the git-annex command (but not other commands) at early startup in initialized repos. Or, is run after initialization. Note that is Remote.Git, it is unable to import Annex.Startup, because Remote.Git importing Logs.Cluster leads the the same import cycle. So ensureInitialized is not passed annexStartup in there. Other commands, like git-annex-shell currently don't run annexStartup either. So there are cases where Logs.Location will not see clusters. So it won't add any cluster UUIDs when loading the log. That's ok, the only reason to do that is to make display of where objects are located include clusters, and to make commands like git-annex get --from treat keys as being located in a cluster. git-annex-shell certainly does not do anything like that, and I'm pretty sure Remote.Git (and callers to Remote.Git.onLocalRepo) don't either.
2024-06-16 18:35:07 +00:00
repoExists = CommandCheck RepoExists (ensureInitialized startupAnnex remoteList)
2016-01-21 17:14:38 +00:00
notBareRepo :: Command -> Command
notBareRepo = addCheck CheckNotBareRepo checkNotBareRepo
checkNotBareRepo :: Annex ()
checkNotBareRepo = whenM (fromRepo Git.repoIsLocalBare) $
giveup "You cannot run this command in a bare repository."
2016-01-21 17:14:38 +00:00
noDaemonRunning :: Command -> Command
noDaemonRunning = addCheck NoDaemonRunning $ whenM (isJust <$> daemonpid) $
giveup "You cannot run this command while git-annex watch or git-annex assistant is running."
2016-01-21 17:14:38 +00:00
where
2020-10-30 19:55:59 +00:00
daemonpid = liftIO . checkDaemon . fromRawFilePath
=<< fromRepo gitAnnexPidFile
2016-01-21 17:14:38 +00:00
dontCheck :: CommandCheck -> Command -> Command
dontCheck check cmd = mutateCheck cmd $ \c -> filter (/= check) c
addCheck :: CommandCheckId -> Annex () -> Command -> Command
addCheck cid check cmd = mutateCheck cmd $ \c ->
CommandCheck cid check : c
2016-01-21 17:14:38 +00:00
mutateCheck :: Command -> ([CommandCheck] -> [CommandCheck]) -> Command
mutateCheck cmd@(Command { cmdcheck = c }) a = cmd { cmdcheck = a c }