git-annex/Command/Sync.hs

843 lines
28 KiB
Haskell
Raw Normal View History

{- git-annex command
-
2011-12-30 20:24:30 +00:00
- Copyright 2011 Joachim Breitner <mail@joachim-breitner.de>
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
- Copyright 2011-2020 Joey Hess <id@joeyh.name>
-
- Licensed under the GNU AGPL version 3 or higher.
-}
{-# LANGUAGE FlexibleContexts #-}
{-# LANGUAGE OverloadedStrings #-}
module Command.Sync (
cmd,
CurrBranch,
2016-04-22 18:35:48 +00:00
mergeConfig,
merge,
prepMerge,
mergeLocal,
mergeRemote,
commitStaged,
commitMsg,
pushBranch,
updateBranch,
syncBranch,
updateBranches,
seekExportContent,
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
SyncOptions(..),
) where
import Command
import qualified Annex
import qualified Annex.Branch
import qualified Remote
import qualified Types.Remote as Remote
import Annex.Hook
2011-12-14 19:56:11 +00:00
import qualified Git.Command
import qualified Git.LsFiles as LsFiles
import qualified Git.Branch
import qualified Git.Merge
import qualified Git.Types as Git
import qualified Git.Ref
import qualified Git
import Git.FilePath
import qualified Remote.Git
import Config
import Config.GitConfig
import Annex.SpecialRemote.Config
import Config.DynamicConfig
import Config.Files
import Annex.Wanted
import Annex.Content
import Command.Get (getKey')
import qualified Command.Move
import qualified Command.Export
import qualified Command.Import
import Annex.Drop
import Annex.UUID
import Logs.UUID
import Logs.Export
import Annex.AutoMerge
import Annex.AdjustedBranch
import Annex.Ssh
import Annex.BloomFilter
import Annex.UpdateInstead
import Annex.Export
import Annex.TaggedPush
import Annex.CurrentBranch
import qualified Database.Export as Export
import Utility.Bloom
import Utility.OptParse
import Utility.Process.Transcript
import Control.Concurrent.MVar
import qualified Data.Map as M
cmd :: Command
cmd = withGlobalOptions [jobsOption] $
command "sync" SectionCommon
"synchronize local repository with remotes"
(paramRepeating paramRemote) (seek <--< optParser)
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
data SyncOptions = SyncOptions
{ syncWith :: CmdParams
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
, onlyAnnexOption :: Bool
, notOnlyAnnexOption :: Bool
, commitOption :: Bool
, noCommitOption :: Bool
, messageOption :: Maybe String
, pullOption :: Bool
, pushOption :: Bool
, contentOption :: Bool
, noContentOption :: Bool
, contentOfOption :: [FilePath]
, cleanupOption :: Bool
, keyOptions :: Maybe KeyOptions
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
, resolveMergeOverride :: Bool
}
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
instance Default SyncOptions where
def = SyncOptions
{ syncWith = []
, onlyAnnexOption = False
, notOnlyAnnexOption = False
, commitOption = False
, noCommitOption = False
, messageOption = Nothing
, pullOption = False
, pushOption = False
, contentOption = False
, noContentOption = False
, contentOfOption = []
, cleanupOption = False
, keyOptions = Nothing
, resolveMergeOverride = False
}
optParser :: CmdParamsDesc -> Parser SyncOptions
optParser desc = SyncOptions
<$> (many $ argument str
( metavar desc
<> completeRemotes
))
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
<*> switch
( long "only-annex"
<> short 'a'
<> help "only sync git-annex branch and annexed file contents"
)
<*> switch
( long "not-only-annex"
<> help "sync git branches as well as annex"
)
<*> switch
( long "commit"
<> help "commit changes to git"
)
<*> switch
( long "no-commit"
<> help "avoid git commit"
)
<*> optional (strOption
( long "message" <> short 'm' <> metavar "MSG"
<> help "commit message"
))
<*> invertableSwitch "pull" True
( help "avoid git pulls from remotes"
)
<*> invertableSwitch "push" True
( help "avoid git pushes to remotes"
)
<*> switch
( long "content"
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
<> help "transfer annexed file contents"
)
<*> switch
( long "no-content"
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
<> help "do not transfer annexed file contents"
)
<*> many (strOption
( long "content-of"
<> short 'C'
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
<> help "transfer contents of annexed files in a given location"
<> metavar paramPath
))
<*> switch
( long "cleanup"
<> help "remove synced/ branches from previous sync"
)
<*> optional parseAllOption
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
<*> invertableSwitch "resolvemerge" True
( help "do not automatically resolve merge conflicts"
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
)
-- Since prepMerge changes the working directory, FilePath options
-- have to be adjusted.
instance DeferredParseClass SyncOptions where
finishParse v = SyncOptions
<$> pure (syncWith v)
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
<*> pure (onlyAnnexOption v)
<*> pure (notOnlyAnnexOption v)
<*> pure (commitOption v)
<*> pure (noCommitOption v)
<*> pure (messageOption v)
<*> pure (pullOption v)
<*> pure (pushOption v)
<*> pure (contentOption v)
<*> pure (noContentOption v)
<*> liftIO (mapM absPath (contentOfOption v))
<*> pure (cleanupOption v)
<*> pure (keyOptions v)
<*> pure (resolveMergeOverride v)
seek :: SyncOptions -> CommandSeek
seek o = do
prepMerge
startConcurrency transferStages (seek' o)
seek' :: SyncOptions -> CommandSeek
seek' o = do
let withbranch a = a =<< getCurrentBranch
remotes <- syncRemotes (syncWith o)
let gitremotes = filter (Remote.gitSyncableRemoteType . Remote.remotetype) remotes
dataremotes <- filter (\r -> Remote.uuid r /= NoUUID)
<$> filterM (not <$$> liftIO . getDynamicConfig . remoteAnnexIgnore . Remote.gitconfig) remotes
let (exportremotes, keyvalueremotes) = partition (exportTree . Remote.config) dataremotes
let importremotes = filter (importTree . Remote.config) dataremotes
if cleanupOption o
then do
commandAction (withbranch cleanupLocal)
mapM_ (commandAction . withbranch . cleanupRemote) gitremotes
else do
-- Syncing involves many actions, any of which
-- can independently fail, without preventing
-- the others from running.
-- These actions cannot be run concurrently.
mapM_ includeCommandAction $ concat
[ [ commit o ]
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
, [ withbranch (mergeLocal mergeConfig o) ]
, map (withbranch . pullRemote o mergeConfig) gitremotes
, [ mergeAnnex ]
]
2020-02-18 16:29:31 +00:00
whenM (shouldSyncContent o) $ do
mapM_ (withbranch . importRemote o mergeConfig) importremotes
-- Send content to any exports before other
-- repositories, in case that lets content
-- be dropped from other repositories.
exportedcontent <- withbranch $
seekExportContent (Just o) exportremotes
syncedcontent <- withbranch $
seekSyncContent o keyvalueremotes
-- Transferring content can take a while,
-- and other changes can be pushed to the
-- git-annex branch on the remotes in the
-- meantime, so pull and merge again to
-- avoid our push overwriting those changes.
when (syncedcontent || exportedcontent) $ do
mapM_ includeCommandAction $ concat
[ map (withbranch . pullRemote o mergeConfig) gitremotes
, [ commitAnnex, mergeAnnex ]
]
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
void $ includeCommandAction $ withbranch $ pushLocal o
-- Pushes to remotes can run concurrently.
mapM_ (commandAction . withbranch . pushRemote o) gitremotes
{- Merging may delete the current directory, so go to the top
- of the repo. This also means that sync always acts on all files in the
- repository, not just on a subdirectory. -}
prepMerge :: Annex ()
prepMerge = Annex.changeDirectory . fromRawFilePath =<< fromRepo Git.repoPath
2016-04-22 18:35:48 +00:00
mergeConfig :: [Git.Merge.MergeConfig]
mergeConfig =
[ Git.Merge.MergeNonInteractive
-- In several situations, unrelated histories should be merged
-- together. This includes pairing in the assistant, merging
-- from a remote into a newly created direct mode repo,
-- and an initial merge from an import from a special remote.
2016-11-16 20:03:23 +00:00
-- (Once direct mode is removed, this could be changed, so only
-- the assistant and import from special remotes use it.)
, Git.Merge.MergeUnrelatedHistories
]
2016-04-22 18:35:48 +00:00
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
merge :: CurrBranch -> [Git.Merge.MergeConfig] -> SyncOptions -> Git.Branch.CommitMode -> Git.Branch -> Annex Bool
merge currbranch mergeconfig o commitmode tomerge = case currbranch of
(Just b, Just adj) -> mergeToAdjustedBranch tomerge (b, adj) mergeconfig canresolvemerge commitmode
(b, _) -> autoMergeFrom tomerge b mergeconfig canresolvemerge commitmode
where
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
canresolvemerge = if resolveMergeOverride o
then getGitConfigVal annexResolveMerge
else return False
syncBranch :: Git.Branch -> Git.Branch
syncBranch = Git.Ref.underBase "refs/heads/synced" . fromAdjustedBranch
2011-12-31 08:11:39 +00:00
remoteBranch :: Remote -> Git.Ref -> Git.Ref
remoteBranch remote = Git.Ref.underBase $ "refs/remotes/" ++ Remote.name remote
2011-12-31 07:01:18 +00:00
2015-10-15 19:10:14 +00:00
-- Do automatic initialization of remotes when possible when getting remote
-- list.
2011-12-31 08:11:39 +00:00
syncRemotes :: [String] -> Annex [Remote]
syncRemotes ps = do
remotelist <- Remote.remoteList' True
available <- filterM (liftIO . getDynamicConfig . remoteAnnexSync . Remote.gitconfig) remotelist
syncRemotes' ps available
syncRemotes' :: [String] -> [Remote] -> Annex [Remote]
syncRemotes' ps available =
ifM (Annex.getState Annex.fast) ( nub <$> pickfast , wanted )
2012-11-12 05:05:04 +00:00
where
pickfast = (++) <$> listed <*> (filterM good (fastest available))
2012-11-12 05:05:04 +00:00
wanted
| null ps = filterM good (concat $ Remote.byCost available)
2012-11-12 05:05:04 +00:00
| otherwise = listed
listed = concat <$> mapM Remote.byNameOrGroup ps
good r
| Remote.gitSyncableRemoteType (Remote.remotetype r) =
Remote.Git.repoAvail =<< Remote.getRepo r
| otherwise = return True
2012-11-12 05:05:04 +00:00
fastest = fromMaybe [] . headMaybe . Remote.byCost
commit :: SyncOptions -> 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
commit o = stopUnless shouldcommit $ starting "commit" (ActionItemOther Nothing) $ do
commitmessage <- maybe commitMsg return (messageOption o)
Annex.Branch.commit =<< Annex.Branch.commitMessage
next $ do
showOutput
void $ inRepo $ Git.Branch.commitCommand Git.Branch.ManualCommit
[ Param "-a"
, Param "-m"
, Param commitmessage
]
return True
where
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
shouldcommit = notOnlyAnnex o <&&>
( pure (commitOption o)
<||> (pure (not (noCommitOption o)) <&&> getGitConfigVal annexAutoCommit)
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
)
commitMsg :: Annex String
commitMsg = do
u <- getUUID
m <- uuidDescMap
return $ "git-annex in " ++ maybe "unknown" fromUUIDDesc (M.lookup u m)
commitStaged :: Git.Branch.CommitMode -> String -> Annex Bool
commitStaged commitmode commitmessage = do
runAnnexHook preCommitAnnexHook
mb <- inRepo Git.Branch.currentUnsafe
let (getparent, branch) = case mb of
Just b -> (Git.Ref.sha b, b)
Nothing -> (Git.Ref.headSha, Git.Ref.headRef)
parents <- maybeToList <$> inRepo getparent
void $ inRepo $ Git.Branch.commit commitmode False commitmessage branch parents
return True
2011-12-10 16:21:22 +00:00
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
mergeLocal :: [Git.Merge.MergeConfig] -> SyncOptions -> CurrBranch -> CommandStart
mergeLocal mergeconfig o currbranch = stopUnless (notOnlyAnnex o) $
mergeLocal' mergeconfig o currbranch
mergeLocal' :: [Git.Merge.MergeConfig] -> SyncOptions -> CurrBranch -> CommandStart
mergeLocal' mergeconfig o currbranch@(Just _, _) =
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
needMerge currbranch >>= \case
Nothing -> stop
Just syncbranch ->
starting "merge" (ActionItemOther (Just $ Git.Ref.describe syncbranch)) $
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
next $ merge currbranch mergeconfig o Git.Branch.ManualCommit syncbranch
mergeLocal' _ _ (Nothing, madj) = do
b <- inRepo Git.Branch.currentUnsafe
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
needMerge (b, madj) >>= \case
Nothing -> stop
Just syncbranch ->
starting "merge" (ActionItemOther (Just $ Git.Ref.describe syncbranch)) $ do
warning $ "There are no commits yet in the currently checked out branch, so cannot merge any remote changes into it."
next $ return False
-- Returns the branch that should be merged, if any.
needMerge :: CurrBranch -> Annex (Maybe Git.Branch)
needMerge (Nothing, _) = return Nothing
needMerge (Just branch, madj) = ifM (allM id checks)
( return (Just syncbranch)
, return Nothing
)
where
checks =
[ not <$> isBareRepo
, inRepo (Git.Ref.exists syncbranch)
, inRepo (Git.Branch.changed branch' syncbranch)
]
syncbranch = syncBranch branch
branch' = maybe branch (adjBranch . originalToAdjusted branch) madj
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
pushLocal :: SyncOptions -> CurrBranch -> CommandStart
pushLocal o b = stopUnless (notOnlyAnnex o) $ do
updateBranches b
stop
updateBranches :: CurrBranch -> Annex ()
updateBranches (Nothing, _) = noop
updateBranches (Just branch, madj) = do
-- When in an adjusted branch, propigate any changes made to it
-- back to the original branch. The adjusted branch may also need
-- to be updated to hide/expose files.
case madj of
Nothing -> noop
Just adj -> do
let origbranch = branch
propigateAdjustedCommits origbranch adj
when (adjustmentHidesFiles adj) $ do
showSideAction "updating adjusted branch"
let adjbranch = originalToAdjusted origbranch adj
unlessM (updateAdjustedBranch adj adjbranch origbranch) $
warning $ unwords [ "Updating adjusted branch failed." ]
-- Update the sync branch to match the new state of the branch
inRepo $ updateBranch (syncBranch branch) branch
updateBranch :: Git.Branch -> Git.Branch -> Git.Repo -> IO ()
updateBranch syncbranch updateto g =
unlessM go $ giveup $ "failed to update " ++ Git.fromRef syncbranch
2012-11-12 05:05:04 +00:00
where
go = Git.Command.runBool
[ Param "branch"
, Param "-f"
, Param $ Git.fromRef $ Git.Ref.base syncbranch
, Param $ Git.fromRef $ updateto
2012-11-12 05:05:04 +00:00
] g
2011-12-30 21:38:38 +00:00
pullRemote :: SyncOptions -> [Git.Merge.MergeConfig] -> Remote -> CurrBranch -> 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
pullRemote o mergeconfig remote branch = stopUnless (pure $ pullOption o && wantpull) $
starting "pull" (ActionItemOther (Just (Remote.name remote))) $ do
2011-12-30 21:38:38 +00:00
showOutput
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
ifM (onlyAnnex o)
( do
void $ fetch $ map Git.fromRef
[ Annex.Branch.name
, syncBranch $ Annex.Branch.name
]
next $ return True
, ifM (fetch [])
( next $ mergeRemote remote branch mergeconfig o
, next $ return True
)
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
)
2012-11-12 05:05:04 +00:00
where
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
fetch bs = do
repo <- Remote.getRepo remote
inRepoWithSshOptionsTo repo (Remote.gitconfig remote) $
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
Git.Command.runBool $
[Param "fetch", Param $ Remote.name remote]
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
++ map Param bs
wantpull = remoteAnnexPull (Remote.gitconfig remote)
importRemote :: SyncOptions -> [Git.Merge.MergeConfig] -> Remote -> CurrBranch -> CommandSeek
importRemote o mergeconfig remote currbranch
| not (pullOption o) || not wantpull = noop
| otherwise = case remoteAnnexTrackingBranch (Remote.gitconfig remote) of
Nothing -> noop
Just tb -> do
let (b, s) = separate (== ':') (Git.fromRef tb)
let branch = Git.Ref b
let subdir = if null s
then Nothing
else Just (asTopFilePath (toRawFilePath s))
Command.Import.seekRemote remote branch subdir
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
void $ mergeRemote remote currbranch mergeconfig o
where
wantpull = remoteAnnexPull (Remote.gitconfig remote)
{- The remote probably has both a master and a synced/master branch.
- Which to merge from? Well, the master has whatever latest changes
- were committed (or pushed changes, if this is a bare remote),
- while the synced/master may have changes that some
- other remote synced to this remote. So, merge them both. -}
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
mergeRemote :: Remote -> CurrBranch -> [Git.Merge.MergeConfig] -> SyncOptions -> CommandCleanup
mergeRemote remote currbranch mergeconfig o = ifM isBareRepo
( return True
, case currbranch of
(Nothing, _) -> do
branch <- inRepo Git.Branch.currentUnsafe
mergelisted (pure (branchlist branch))
(Just branch, _) -> do
inRepo $ updateBranch (syncBranch branch) branch
mergelisted (tomerge (branchlist (Just branch)))
)
2012-11-12 05:05:04 +00:00
where
mergelisted getlist = and <$>
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
(mapM (merge currbranch mergeconfig o Git.Branch.ManualCommit . remoteBranch remote) =<< getlist)
tomerge = filterM (changed remote)
2012-11-12 05:05:04 +00:00
branchlist Nothing = []
branchlist (Just branch) = [fromAdjustedBranch branch, syncBranch branch]
pushRemote :: SyncOptions -> Remote -> CurrBranch -> CommandStart
pushRemote _o _remote (Nothing, _) = stop
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
pushRemote o remote (Just branch, _) = do
onlyannex <- onlyAnnex o
let mainbranch = if onlyannex then Nothing else Just branch
stopUnless (pure (pushOption o) <&&> needpush mainbranch) $
starting "push" (ActionItemOther (Just (Remote.name remote))) $ next $ do
repo <- Remote.getRepo remote
showOutput
ok <- inRepoWithSshOptionsTo repo gc $
pushBranch remote mainbranch
if ok
then postpushupdate repo
else do
warning $ unwords [ "Pushing to " ++ Remote.name remote ++ " failed." ]
return ok
2012-11-12 05:05:04 +00:00
where
gc = Remote.gitconfig remote
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
needpush mainbranch
| remoteAnnexReadOnly gc = return False
| not (remoteAnnexPush gc) = return False
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
| otherwise = anyM (newer remote) $ catMaybes
[ syncBranch <$> mainbranch
, Just (Annex.Branch.name)
]
-- Older remotes on crippled filesystems may not have a
-- post-receive hook set up, so when updateInstead emulation
-- is needed, run post-receive manually.
postpushupdate repo = case Git.repoWorkTree repo of
Nothing -> return True
Just wt -> ifM needemulation
( liftIO $ do
p <- readProgramFile
boolSystem' p [Param "post-receive"]
(\cp -> cp { cwd = Just (fromRawFilePath wt) })
, return True
)
where
needemulation = Remote.Git.onLocal repo remote $
(annexCrippledFileSystem <$> Annex.getGitConfig)
<&&>
needUpdateInsteadEmulation
{- Pushes a regular branch like master to a remote. Also pushes the git-annex
- branch.
-
- If the remote is a bare git repository, it's best to push the regular
- branch directly to it, so that cloning/pulling will get it.
- On the other hand, if it's not bare, pushing to the checked out branch
- will generally fail (except with receive.denyCurrentBranch=updateInstead),
- and this is why we push to its syncBranch.
-
- Git offers no way to tell if a remote is bare or not, so both methods
- are tried.
-
- The direct push is likely to spew an ugly error message, so its stderr is
- often elided. Since git progress display goes to stderr too, the
- sync push is done first, and actually sends the data. Then the
- direct push is tried, with stderr discarded, to update the branch ref
- on the remote.
-
- The sync push forces the update of the remote synced/git-annex branch.
- This is necessary if a transition has rewritten the git-annex branch.
- Normally any changes to the git-annex branch get pulled and merged before
- this push, so this forcing is unlikely to overwrite new data pushed
- in from another repository that is also syncing.
-
- But overwriting of data on synced/git-annex can happen, in a race.
- The only difference caused by using a forced push in that case is that
- the last repository to push wins the race, rather than the first to push.
-}
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
pushBranch :: Remote -> Maybe Git.Branch -> Git.Repo -> IO Bool
pushBranch remote mbranch g = directpush `after` annexpush `after` syncpush
where
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
syncpush = flip Git.Command.runBool g $ pushparams $ catMaybes
[ Just $ Git.Branch.forcePush $ refspec Annex.Branch.name
, (refspec . fromAdjustedBranch) <$> mbranch
]
annexpush = void $ tryIO $ flip Git.Command.runQuiet g $ pushparams
[ Git.fromRef $ Git.Ref.base $ Annex.Branch.name ]
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
directpush = case mbranch of
Nothing -> noop
-- Git prints out an error message when this fails.
-- In the default configuration of receive.denyCurrentBranch,
-- the error message mentions that config setting
-- (and should even if it is localized), and is quite long,
-- and the user was not intending to update the checked out
-- branch, so in that case, avoid displaying the error
-- message. Do display other error messages though,
-- including the error displayed when
-- receive.denyCurrentBranch=updateInstead -- the user
-- will want to see that one.
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
Just branch -> do
let p = flip Git.Command.gitCreateProcess g $ pushparams
[ Git.fromRef $ Git.Ref.base $ fromAdjustedBranch branch ]
(transcript, ok) <- processTranscript' p Nothing
when (not ok && not ("denyCurrentBranch" `isInfixOf` transcript)) $
hPutStr stderr transcript
pushparams branches =
[ Param "push"
, Param $ Remote.name remote
] ++ map Param branches
2012-11-12 05:05:04 +00:00
refspec b = concat
[ Git.fromRef $ Git.Ref.base b
2012-11-12 05:05:04 +00:00
, ":"
, Git.fromRef $ Git.Ref.base $ syncBranch b
2012-11-12 05:05:04 +00:00
]
commitAnnex :: CommandStart
commitAnnex = do
Annex.Branch.commit =<< Annex.Branch.commitMessage
stop
mergeAnnex :: CommandStart
2011-12-30 21:38:38 +00:00
mergeAnnex = do
2013-09-25 07:09:06 +00:00
void Annex.Branch.forceUpdate
2011-12-30 21:38:38 +00:00
stop
2011-12-31 06:18:16 +00:00
2011-12-31 08:11:39 +00:00
changed :: Remote -> Git.Ref -> Annex Bool
2011-12-31 07:01:18 +00:00
changed remote b = do
let r = remoteBranch remote b
ifM (inRepo $ Git.Ref.exists r)
( inRepo $ Git.Branch.changed b r
, return False
)
2011-12-31 07:01:18 +00:00
2011-12-31 08:11:39 +00:00
newer :: Remote -> Git.Ref -> Annex Bool
2011-12-31 07:01:18 +00:00
newer remote b = do
let r = remoteBranch remote b
ifM (inRepo $ Git.Ref.exists r)
( inRepo $ Git.Branch.changed r b
, return True
)
{- Without --all, only looks at files in the work tree.
- (Or, when in an ajusted branch where some files are hidden, at files in
- the original branch.)
-
- With --all, makes a second pass over all keys.
- This ensures that preferred content expressions that match on
- filenames work, even when in --all mode.
-
- Returns true if any file transfers were made.
-
- When concurrency is enabled, files are processed concurrently.
-}
seekSyncContent :: SyncOptions -> [Remote] -> CurrBranch -> Annex Bool
seekSyncContent _ [] _ = return False
seekSyncContent o rs currbranch = do
mvar <- liftIO newEmptyMVar
bloom <- case keyOptions o of
Just WantAllKeys -> Just <$> genBloomFilter (seekworktree mvar [])
_ -> case currbranch of
(Just origbranch, Just adj) | adjustmentHidesFiles adj -> do
l <- workTreeItems' (AllowHidden True) (contentOfOption o)
seekincludinghidden origbranch mvar l (const noop)
pure Nothing
_ -> do
l <- workTreeItems (contentOfOption o)
seekworktree mvar l (const noop)
pure Nothing
withKeyOptions' (keyOptions o) False
(return (gokey mvar bloom))
(const noop)
[]
finishCommandActions
liftIO $ not <$> isEmptyMVar mvar
where
seekworktree mvar l bloomfeeder =
seekHelper LsFiles.inRepo l
>>= gofiles bloomfeeder mvar
seekincludinghidden origbranch mvar l bloomfeeder =
seekHelper (LsFiles.inRepoOrBranch origbranch) l
>>= gofiles bloomfeeder mvar
gofiles bloomfeeder mvar = mapM_ $ \f ->
ifAnnexed f
(go (Right bloomfeeder) mvar (AssociatedFile (Just f)))
noop
gokey mvar bloom (k, _) = go (Left bloom) mvar (AssociatedFile Nothing) k
go ebloom mvar af k = do
-- Run syncFile as a command action so file transfers run
-- concurrently.
let ai = OnlyActionOn k (ActionItemKey k)
commandAction $ startingNoMessage ai $ do
whenM (syncFile ebloom rs af k) $
void $ liftIO $ tryPutMVar mvar ()
next $ return True
2015-06-16 23:07:14 +00:00
{- If it's preferred content, and we don't have it, get it from one of the
- listed remotes (preferring the cheaper earlier ones).
-
- Send it to each remote that doesn't have it, and for which it's
- preferred content.
-
- Drop it locally if it's not preferred content (honoring numcopies).
-
- Drop it from each remote that has it, where it's not preferred content
- (honoring numcopies).
-
- Returns True if any file transfers were made.
2015-06-16 23:07:14 +00:00
-}
syncFile :: Either (Maybe (Bloom Key)) (Key -> Annex ()) -> [Remote] -> AssociatedFile -> Key -> Annex Bool
syncFile ebloom rs af k = do
inhere <- inAnnex k
locs <- map Remote.uuid <$> Remote.keyPossibilities k
let (have, lack) = partition (\r -> Remote.uuid r `elem` locs) rs
got <- anyM id =<< handleget have inhere
putrs <- handleput lack
u <- getUUID
let locs' = concat [if inhere || got then [u] else [], putrs, locs]
-- A bloom filter is populated with all the keys in the first pass.
-- On the second pass, avoid dropping keys that were seen in the
-- first pass, which would happen otherwise when preferred content
-- matches on the filename, which is not available in the second
-- pass.
--
-- When there's a false positive in the bloom filter, the result
-- is keeping a key that preferred content doesn't really want.
seenbloom <- case ebloom of
Left Nothing -> pure False
Left (Just bloom) -> pure (elemB k bloom)
Right bloomfeeder -> bloomfeeder k >> return False
unless seenbloom $
-- Using callCommandAction rather than
-- includeCommandAction for drops,
-- because a failure to drop does not mean
-- the sync failed.
handleDropsFrom locs' rs "unwanted" True k af []
callCommandAction
return (got || not (null putrs))
where
wantget have inhere = allM id
[ pure (not $ null have)
, pure (not inhere)
2015-06-16 19:07:03 +00:00
, wantGet True (Just k) af
]
handleget have inhere = ifM (wantget have inhere)
( return [ get have ]
, return []
)
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
get have = includeCommandAction $ starting "get" ai $
stopUnless (getKey' k af have) $
next $ return True
wantput r
| Remote.readonly r || remoteAnnexReadOnly (Remote.gitconfig r) = return False
2015-06-16 19:07:03 +00:00
| otherwise = wantSend True (Just k) af (Remote.uuid r)
handleput lack = catMaybes <$> ifM (inAnnex k)
( forM lack $ \r ->
ifM (wantput r <&&> put r)
( return (Just (Remote.uuid r))
, return Nothing
)
, return []
)
2016-03-09 19:04:00 +00:00
put dest = includeCommandAction $
2019-06-06 16:53:24 +00:00
Command.Move.toStart' dest Command.Move.RemoveNever af k ai
ai = mkActionItem (k, af)
{- When a remote has an annex-tracking-branch configuration, change the export
- to contain the current content of the branch. Otherwise, transfer any files
- that were part of an export but are not in the remote yet.
-
- Returns True if any file transfers were made.
-}
seekExportContent :: Maybe SyncOptions -> [Remote] -> CurrBranch -> Annex Bool
seekExportContent o rs (currbranch, _) = or <$> forM rs go
where
go r
| not (maybe True pullOption o) = return False
| not (remoteAnnexPush (Remote.gitconfig r)) = return False
| otherwise = bracket
(Export.openDb (Remote.uuid r))
Export.closeDb
(\db -> Export.writeLockDbWhile db (go' r db))
go' r db = case remoteAnnexTrackingBranch (Remote.gitconfig r) of
Nothing -> nontracking r db
Just b -> do
mtree <- inRepo $ Git.Ref.tree b
mtbcommitsha <- Command.Export.getExportCommit r b
case (mtree, mtbcommitsha) of
(Just tree, Just _) -> do
filteredtree <- Command.Export.filterPreferredContent r tree
Command.Export.changeExport r db filteredtree
Command.Export.fillExport r db filteredtree mtbcommitsha
_ -> nontracking r db
nontracking r db = do
exported <- getExport (Remote.uuid r)
maybe noop (warnnontracking r exported) currbranch
fillexport r db (exportedTreeishes exported) Nothing
warnnontracking r exported currb = inRepo (Git.Ref.tree currb) >>= \case
Just currt | not (any (== currt) (exportedTreeishes exported)) ->
showLongNote $ unwords
[ "Not updating export to " ++ Remote.name r
, "to reflect changes to the tree, because export"
, "tracking is not enabled. "
, "(Set " ++ gitconfig ++ " to enable it.)"
]
_ -> noop
where
gitconfig = show (remoteAnnexConfig r "tracking-branch")
fillexport _ _ [] _ = return False
fillexport r db (tree:[]) mtbcommitsha = do
let filteredtree = Command.Export.PreferredFiltered tree
Command.Export.fillExport r db filteredtree mtbcommitsha
fillexport r _ _ _ = do
warnExportImportConflict r
return False
cleanupLocal :: CurrBranch -> CommandStart
cleanupLocal (Nothing, _) = stop
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
cleanupLocal (Just currb, _) =
starting "cleanup" (ActionItemOther (Just "local")) $
next $ do
delbranch $ syncBranch currb
delbranch $ syncBranch $ Git.Ref.base $ Annex.Branch.name
mapM_ (\(s,r) -> inRepo $ Git.Ref.delete s r)
=<< listTaggedBranches
return True
where
delbranch b = whenM (inRepo $ Git.Ref.exists $ Git.Ref.branchRef b) $
inRepo $ Git.Branch.delete b
cleanupRemote :: Remote -> CurrBranch -> CommandStart
cleanupRemote _ (Nothing, _) = stop
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
cleanupRemote remote (Just b, _) =
starting "cleanup" (ActionItemOther (Just (Remote.name remote))) $
next $ inRepo $ Git.Command.runBool
[ Param "push"
, Param "--quiet"
, Param "--delete"
, Param $ Remote.name remote
, Param $ Git.fromRef $ syncBranch b
, Param $ Git.fromRef $ syncBranch $
Git.Ref.base $ Annex.Branch.name
]
2020-02-18 16:29:31 +00:00
shouldSyncContent :: SyncOptions -> Annex Bool
shouldSyncContent o
| noContentOption o = pure False
| contentOption o || not (null (contentOfOption o)) = pure True
| otherwise = getGitConfigVal annexSyncContent <||> onlyAnnex o
sync --only-annex and annex.synconlyannex * Added sync --only-annex, which syncs the git-annex branch and annexed content but leaves managing the other git branches up to you. * Added annex.synconlyannex git config setting, which can also be set with git-annex config to configure sync in all clones of the repo. Use case is then the user has their own git workflow, and wants to use git-annex without disrupting that, so they sync --only-annex to get the git-annex stuff in sync in addition to their usual git workflow. When annex.synconlyannex is set, --not-only-annex can be used to override it. It's not entirely clear what --only-annex --commit or --only-annex --push should do, and I left that combination not documented because I don't know if I might want to change the current behavior, which is that such options do not override the --only-annex. My gut feeling is that there is no good reasons to use such combinations; if you want to use your own git workflow, you'll be doing your own committing and pulling and pushing. A subtle question is, how should import/export special remotes be handled? Importing updates their remote tracking branch and merges it into master. If --only-annex prevented that git branch stuff, then it would prevent exporting to the special remote, in the case where it has changes that were not imported yet, because there would be a unresolved conflict. I decided that it's best to treat the fact that there's a remote tracking branch for import/export as an implementation detail in this case. The more important thing is that an import/export special remote is entirely annexed content, and so it makes a lot of sense that --only-annex will still sync with it.
2020-02-17 19:19:58 +00:00
notOnlyAnnex :: SyncOptions -> Annex Bool
notOnlyAnnex o = not <$> onlyAnnex o
onlyAnnex :: SyncOptions -> Annex Bool
2020-02-18 16:29:31 +00:00
onlyAnnex o
| notOnlyAnnexOption o = pure False
| onlyAnnexOption o = pure True
| otherwise = getGitConfigVal annexSyncOnlyAnnex