2011-03-27 19:56:43 +00:00
{- Standard git remotes.
-
2018-01-10 18:21:18 +00:00
- Copyright 2011 - 2018 Joey Hess < id @ joeyh . name >
2011-03-27 19:56:43 +00:00
-
2018-03-09 05:03:11 +00:00
- Licensed under the GNU AGPL version 3 or higher .
2011-03-27 19:56:43 +00:00
- }
2013-05-11 20:03:00 +00:00
{- # LANGUAGE CPP # -}
2012-07-22 17:48:50 +00:00
module Remote.Git (
remote ,
configRead ,
repoAvail ,
2017-02-17 19:21:39 +00:00
onLocal ,
2012-07-22 17:48:50 +00:00
) where
2011-03-27 19:56:43 +00:00
2016-01-20 20:36:33 +00:00
import Annex.Common
2013-05-14 17:53:29 +00:00
import Annex.Ssh
2011-06-02 01:56:04 +00:00
import Types.Remote
2013-01-01 17:52:47 +00:00
import Types.GitConfig
2011-06-30 17:16:57 +00:00
import qualified Git
2011-12-13 19:05:07 +00:00
import qualified Git.Config
import qualified Git.Construct
2012-10-12 17:45:14 +00:00
import qualified Git.Command
2013-09-05 20:02:39 +00:00
import qualified Git.GCrypt
2014-05-22 17:42:17 +00:00
import qualified Git.Types as Git
2011-03-27 19:56:43 +00:00
import qualified Annex
2012-01-28 19:54:42 +00:00
import Logs.Presence
2014-03-22 14:42:38 +00:00
import Annex.Transfer
2011-10-15 21:47:03 +00:00
import Annex.UUID
2011-10-04 04:40:47 +00:00
import qualified Annex.Content
2011-12-12 21:38:46 +00:00
import qualified Annex.BranchState
2012-02-25 22:02:49 +00:00
import qualified Annex.Branch
2013-09-28 18:35:21 +00:00
import qualified Annex.Url as Url
2013-05-12 23:19:28 +00:00
import Utility.Tmp
2011-03-28 01:43:25 +00:00
import Config
2013-03-13 20:16:01 +00:00
import Config.Cost
2017-08-17 16:26:14 +00:00
import Config.DynamicConfig
2014-01-26 20:36:31 +00:00
import Annex.Init
deal with unlocked files when calling rsyncParamsRemote
In copyFromRemote, it used to check isDirect, but that was not needed;
the remote is sending the file, so it doesn't matter if the local,
receiving repository is in direct mode or not. And, since the content is not
present, yet, it's certianly not unlocked. Note that, the remote may indeed
be sending an unlocked file, but sendkey uses sendAnnex, which will detect
if the file is modified before or during transfer, and will exit nonzero,
aborting the upload. So, the receiver doesn't need any checks.
In copyToRemote, it forces recvkey to verify content whenever it's being
sent from a v6 repository. recvkey is almost always going to verify content
anyway, unless annex.verify is not set. So, this doesn't make it any more
expensive, except for in that unusual configuration. The alternative would
be to change the recvkey interface, so that the sender checks afterwards if
what it was sending changed, and the receiver then throws out the bad
transfer. That would be less expensive for the reciever, as it would not
need to do a checksum verification. But, it would mean another network
round trip, and since rsync closes the connection, it would need to open
another ssh connection to do this. Even with connction caching, that would
add latency to uploads. It would also complicate the interface, especially
because an older git-annex-shell would not have the new interface
available. For these reasons, I prefer punting on that at this time, and
instead someone might set annex.verify=false and be unhappy that it still
verifies..
(One other gotcha not dealt with is that a v5 repo could be upgraded to v6
while an upload is in progress, and a file unlocked and modified.)
(Also, I double-checked Remote.GCrypt's calls to rsyncParamsRemote, and
they're fine. When a file is being uploaded to gcrypt, or any other special
repository, it is mediated by sendAnnex, so changes will be detected at
that level and the special remote implementation doesn't need to worry
about them.)
2015-12-26 18:05:07 +00:00
import Annex.Version
2014-03-13 23:06:26 +00:00
import Types.CleanupActions
2014-01-26 20:32:55 +00:00
import qualified CmdLine.GitAnnexShell.Fields as Fields
2012-12-12 23:20:38 +00:00
import Logs.Location
2013-03-28 21:03:04 +00:00
import Utility.Metered
2013-08-04 17:12:18 +00:00
# ifndef mingw32_HOST_OS
import Utility.CopyFile
# endif
2013-10-14 19:05:10 +00:00
import Utility.Env
2013-10-11 20:03:18 +00:00
import Utility.Batch
2016-12-09 17:34:00 +00:00
import Utility.SimpleProtocol
2013-09-07 22:38:00 +00:00
import Remote.Helper.Git
2013-09-24 17:37:41 +00:00
import Remote.Helper.Messages
2017-09-01 17:02:07 +00:00
import Remote.Helper.Export
2013-09-24 17:37:41 +00:00
import qualified Remote.Helper.Ssh as Ssh
2013-09-07 22:38:00 +00:00
import qualified Remote.GCrypt
2016-12-06 16:19:47 +00:00
import qualified Remote.P2P
2018-03-08 20:21:16 +00:00
import qualified Remote.Helper.P2P as P2PHelper
import P2P.Address
2015-02-28 21:23:13 +00:00
import Annex.Path
2014-05-22 17:42:17 +00:00
import Creds
2015-10-09 17:07:03 +00:00
import Types.NumCopies
2017-09-30 02:36:08 +00:00
import Annex.Action
2011-03-27 19:56:43 +00:00
2012-09-20 17:35:53 +00:00
import Control.Concurrent
2012-10-05 21:03:58 +00:00
import Control.Concurrent.MSampleVar
2013-09-07 22:38:00 +00:00
import qualified Data.Map as M
2014-08-10 18:52:58 +00:00
import Network.URI
2012-09-20 17:35:53 +00:00
2011-12-31 08:11:39 +00:00
remote :: RemoteType
2017-09-07 17:45:31 +00:00
remote = RemoteType
{ typename = " git "
, enumerate = list
, generate = gen
, setup = gitSetup
, exportSupported = exportUnsupported
}
2011-03-29 03:51:07 +00:00
2015-08-05 17:49:54 +00:00
list :: Bool -> Annex [ Git . Repo ]
list autoinit = do
2011-12-14 19:30:14 +00:00
c <- fromRepo Git . config
2018-01-09 19:36:56 +00:00
rs <- mapM ( tweakurl c ) =<< Annex . getGitRemotes
2015-08-05 17:49:54 +00:00
mapM ( configRead autoinit ) rs
2012-10-29 01:27:15 +00:00
where
annexurl n = " remote. " ++ n ++ " .annexurl "
tweakurl c r = do
let n = fromJust $ Git . remoteName r
case M . lookup ( annexurl n ) c of
Nothing -> return r
Just url -> inRepo $ \ g ->
Git . Construct . remoteNamed n $
Git . Construct . fromRemoteLocation url g
2012-07-22 17:48:50 +00:00
2014-05-22 17:42:17 +00:00
{- Git remotes are normally set up using standard git command, not
- git - annex initremote and enableremote .
-
- For initremote , the git remote must already be set up , and have a uuid .
- Initremote simply remembers its location .
-
- enableremote simply sets up a git remote using the stored location .
- No attempt is made to make the remote be accessible via ssh key setup ,
- etc .
- }
2017-02-07 18:35:58 +00:00
gitSetup :: SetupStage -> Maybe UUID -> Maybe CredPair -> RemoteConfig -> RemoteGitConfig -> Annex ( RemoteConfig , UUID )
gitSetup Init mu _ c _ = do
2016-11-16 01:29:54 +00:00
let location = fromMaybe ( giveup " Specify location=url " ) $
2014-05-22 17:42:17 +00:00
Url . parseURIRelaxed =<< M . lookup " location " c
2018-01-09 19:36:56 +00:00
rs <- Annex . getGitRemotes
u <- case filter ( \ r -> Git . location r == Git . Url location ) rs of
2014-05-22 17:42:17 +00:00
[ r ] -> getRepoUUID r
2016-11-16 01:29:54 +00:00
[] -> giveup " could not find existing git remote with specified location "
_ -> giveup " found multiple git remotes with specified location "
2017-02-07 18:35:58 +00:00
if isNothing mu || mu == Just u
then return ( c , u )
else error " git remote did not have specified uuid "
2017-09-04 16:40:33 +00:00
gitSetup ( Enable _ ) ( Just u ) _ c _ = do
2014-05-22 17:42:17 +00:00
inRepo $ Git . Command . run
[ Param " remote "
, Param " add "
2016-11-16 01:29:54 +00:00
, Param $ fromMaybe ( giveup " no name " ) ( M . lookup " name " c )
, Param $ fromMaybe ( giveup " no location " ) ( M . lookup " location " c )
2014-05-22 17:42:17 +00:00
]
return ( c , u )
2017-09-04 16:40:33 +00:00
gitSetup ( Enable _ ) Nothing _ _ _ = error " unable to enable git remote with no specified uuid "
2014-05-22 17:42:17 +00:00
2012-07-22 17:48:50 +00:00
{- It's assumed to be cheap to read the config of non - URL remotes, so this is
2018-01-10 18:21:18 +00:00
- done each time git - annex is run in a way that uses remotes , unless
- annex - checkuuid is false .
2012-07-22 17:48:50 +00:00
-
- Conversely , the config of an URL remote is only read when there is no
- cached UUID value . - }
2015-08-05 17:49:54 +00:00
configRead :: Bool -> Git . Repo -> Annex Git . Repo
configRead autoinit r = do
2014-05-16 20:08:20 +00:00
gc <- Annex . getRemoteGitConfig r
2012-07-22 17:48:50 +00:00
u <- getRepoUUID r
2017-08-17 16:26:14 +00:00
annexignore <- liftIO $ getDynamicConfig ( remoteAnnexIgnore gc )
case ( repoCheap r , annexignore , u ) of
2013-01-01 17:52:47 +00:00
( _ , True , _ ) -> return r
2018-01-10 18:21:18 +00:00
( True , _ , _ )
| remoteAnnexCheckUUID gc -> tryGitConfigRead autoinit r
| otherwise -> return r
2015-08-05 17:49:54 +00:00
( False , _ , NoUUID ) -> tryGitConfigRead autoinit r
2012-07-22 17:48:50 +00:00
_ -> return r
2011-03-28 01:43:25 +00:00
2013-09-12 19:54:35 +00:00
gen :: Git . Repo -> UUID -> RemoteConfig -> RemoteGitConfig -> Annex ( Maybe Remote )
2013-09-07 22:38:00 +00:00
gen r u c gc
2014-08-04 12:42:04 +00:00
| Git . GCrypt . isEncrypted r = Remote . GCrypt . chainGen r u c gc
2016-12-06 16:19:47 +00:00
| otherwise = case repoP2PAddress r of
2018-01-10 18:21:18 +00:00
Nothing -> do
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
st <- mkState r u gc
go st <$> remoteCost gc defcst
2016-12-06 16:19:47 +00:00
Just addr -> Remote . P2P . chainGen addr r u c gc
2012-10-29 01:27:15 +00:00
where
defcst = if repoCheap r then cheapRemoteCost else expensiveRemoteCost
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
go st cst = Just new
2013-01-01 17:52:47 +00:00
where
new = Remote
{ uuid = u
, cost = cst
, name = Git . repoDescribe r
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
, storeKey = copyToRemote new st
, retrieveKeyFile = copyFromRemote new st
, retrieveKeyFileCheap = copyFromRemoteCheap new st
, removeKey = dropKey new st
, lockContent = Just ( lockKey new st )
, checkPresent = inAnnex new st
2014-08-06 17:45:19 +00:00
, checkPresentCheap = repoCheap r
2017-09-01 17:02:07 +00:00
, exportActions = exportUnsupported
2013-01-01 17:52:47 +00:00
, whereisKey = Nothing
2013-10-11 20:03:18 +00:00
, remoteFsck = if Git . repoIsUrl r
then Nothing
else Just $ fsckOnRemote r
2013-10-27 19:38:59 +00:00
, repairRepo = if Git . repoIsUrl r
then Nothing
else Just $ repairRemote r
2013-11-02 20:37:28 +00:00
, config = c
2013-09-07 22:38:00 +00:00
, localpath = localpathCalc r
2013-01-01 17:52:47 +00:00
, repo = r
2017-05-11 20:03:50 +00:00
, gitconfig = gc { remoteGitConfig = extractGitConfig r }
2013-01-01 17:52:47 +00:00
, readonly = Git . repoIsHttp r
2014-01-13 18:41:10 +00:00
, availability = availabilityCalc r
2013-01-01 17:52:47 +00:00
, remotetype = remote
2014-08-10 18:52:58 +00:00
, mkUnavailable = unavailable r u c gc
2015-01-13 22:11:03 +00:00
, getInfo = gitRepoInfo new
2014-12-08 17:40:15 +00:00
, claimUrl = Nothing
2014-12-11 19:32:42 +00:00
, checkUrl = Nothing
2013-01-01 17:52:47 +00:00
}
2011-03-27 19:56:43 +00:00
2014-08-10 18:52:58 +00:00
unavailable :: Git . Repo -> UUID -> RemoteConfig -> RemoteGitConfig -> Annex ( Maybe Remote )
unavailable r u c gc = gen r' u c gc
where
r' = case Git . location r of
Git . Local { Git . gitdir = d } ->
r { Git . location = Git . LocalUnknown d }
Git . Url url -> case uriAuthority url of
Just auth ->
let auth' = auth { uriRegName = " !dne! " }
in r { Git . location = Git . Url ( url { uriAuthority = Just auth' } ) }
Nothing -> r { Git . location = Git . Unknown }
_ -> r -- already unavailable
2012-06-27 02:27:30 +00:00
{- Checks relatively inexpensively if a repository is available for use. -}
repoAvail :: Git . Repo -> Annex Bool
repoAvail r
| Git . repoIsHttp r = return True
2013-09-09 13:58:17 +00:00
| Git . GCrypt . isEncrypted r = do
g <- gitRepo
liftIO $ do
2013-09-19 16:53:24 +00:00
er <- Git . GCrypt . encryptedRemote g r
2013-09-09 13:58:17 +00:00
if Git . repoIsLocal er || Git . repoIsLocalUnknown er
then catchBoolIO $
void ( Git . Config . read er ) >> return True
else return True
2012-06-27 02:27:30 +00:00
| Git . repoIsUrl r = return True
| Git . repoIsLocalUnknown r = return False
2014-03-06 21:12:50 +00:00
| otherwise = liftIO $ isJust <$> catchMaybeIO ( Git . Config . read r )
2012-06-27 02:27:30 +00:00
2011-03-28 01:43:25 +00:00
{- Tries to read the config for a specified remote, updates state, and
- returns the updated repo . - }
2015-08-05 17:49:54 +00:00
tryGitConfigRead :: Bool -> Git . Repo -> Annex Git . Repo
tryGitConfigRead autoinit r
2013-04-24 00:06:02 +00:00
| haveconfig r = return r -- already read
2012-10-12 17:45:14 +00:00
| Git . repoIsSsh r = store $ do
2017-02-15 19:08:46 +00:00
v <- Ssh . onRemote NoConsumeStdin r
( pipedconfig , return ( Left $ giveup " configlist failed " ) )
" configlist " [] configlistfields
2013-04-24 00:06:02 +00:00
case v of
Right r'
| haveconfig r' -> return r'
| otherwise -> configlist_failed
Left _ -> configlist_failed
2014-02-25 01:29:37 +00:00
| Git . repoIsHttp r = store geturlconfig
2013-09-08 19:19:14 +00:00
| Git . GCrypt . isEncrypted r = handlegcrypt =<< getConfigMaybe ( remoteConfig r " uuid " )
2011-03-28 01:43:25 +00:00
| Git . repoIsUrl r = return r
2014-07-15 18:45:27 +00:00
| otherwise = store $ liftIO $
readlocalannexconfig ` catchNonAsync ` ( const $ return r )
2012-10-29 01:27:15 +00:00
where
2013-04-24 00:06:02 +00:00
haveconfig = not . M . null . Git . config
2013-09-24 21:51:12 +00:00
pipedconfig cmd params = do
v <- Git . Config . fromPipe r cmd params
case v of
Right ( r' , val ) -> do
2015-10-15 19:28:29 +00:00
unless ( isUUIDConfigured r' || null val ) $ do
2013-09-24 21:51:12 +00:00
warningIO $ " Failed to get annex.uuid configuration of repository " ++ Git . repoDescribe r
warningIO $ " Instead, got: " ++ show val
warningIO $ " This is unexpected; please check the network transport! "
return $ Right r'
Left l -> return $ Left l
2011-08-17 00:48:11 +00:00
2014-02-25 01:29:37 +00:00
geturlconfig = do
2014-02-25 02:00:25 +00:00
uo <- Url . getUrlOptions
2013-05-25 05:47:19 +00:00
v <- liftIO $ withTmpFile " git-annex.tmp " $ \ tmpfile h -> do
2012-10-29 01:27:15 +00:00
hClose h
2015-04-19 04:38:29 +00:00
let url = Git . repoLocation r ++ " /config "
ifM ( Url . downloadQuiet url tmpfile uo )
2016-05-03 16:53:42 +00:00
( Just <$> pipedconfig " git " [ Param " config " , Param " --null " , Param " --list " , Param " --file " , File tmpfile ]
, return Nothing
2013-05-25 05:47:19 +00:00
)
case v of
2016-05-03 16:53:42 +00:00
Just ( Right r' ) -> do
2014-01-26 17:03:25 +00:00
-- Cache when http remote is not bare for
-- optimisation.
unless ( Git . Config . isBare r' ) $
2016-05-24 19:48:22 +00:00
setremote setRemoteBare False
2014-01-26 17:03:25 +00:00
return r'
2017-03-29 16:43:47 +00:00
_ -> do
set_ignore " not usable by git-annex " False
return r
2011-08-17 00:48:11 +00:00
2012-10-29 01:27:15 +00:00
store = observe $ \ r' -> do
2018-01-09 19:36:56 +00:00
l <- Annex . getGitRemotes
let rs = exchange l r'
Annex . changeState $ \ s -> s { Annex . gitremotes = Just rs }
2011-08-17 00:48:11 +00:00
2012-10-29 01:27:15 +00:00
exchange [] _ = []
exchange ( old : ls ) new
| Git . remoteName old == Git . remoteName new =
new : exchange ls new
| otherwise =
old : exchange ls new
2011-03-27 19:56:43 +00:00
2013-04-24 00:06:02 +00:00
{- Is this remote just not available, or does
- it not have git - annex - shell ?
- Find out by trying to fetch from the remote . - }
configlist_failed = case Git . remoteName r of
Nothing -> return r
Just n -> do
2014-05-16 16:58:50 +00:00
whenM ( inRepo $ Git . Command . runBool [ Param " fetch " , Param " --quiet " , Param n ] ) $ do
set_ignore " does not have git-annex installed " True
2013-04-24 00:06:02 +00:00
return r
2013-05-25 05:47:19 +00:00
2014-05-16 16:58:50 +00:00
set_ignore msg longmessage = do
2014-01-26 17:03:25 +00:00
case Git . remoteName r of
Nothing -> noop
2014-05-16 16:58:50 +00:00
Just n -> do
2016-05-24 19:48:22 +00:00
warning $ " Remote " ++ n ++ " " ++ msg ++ " ; setting annex-ignore "
2014-05-16 16:58:50 +00:00
when longmessage $
2016-05-24 19:48:22 +00:00
warning $ " This could be a problem with the git-annex installation on the remote. Please make sure that git-annex-shell is available in PATH when you ssh into the remote. Once you have fixed the git-annex installation, run: git annex enableremote " ++ n
setremote setRemoteIgnore True
2014-01-26 17:03:25 +00:00
2016-05-24 19:48:22 +00:00
setremote setter v = case Git . remoteName r of
2013-05-25 05:47:19 +00:00
Nothing -> noop
2016-05-27 15:15:52 +00:00
Just _ -> setter r v
2016-05-24 19:48:22 +00:00
2013-09-08 19:19:14 +00:00
handlegcrypt Nothing = return r
handlegcrypt ( Just _cacheduuid ) = do
-- Generate UUID from the gcrypt-id
g <- gitRepo
case Git . GCrypt . remoteRepoId g ( Git . remoteName r ) of
Nothing -> return r
Just v -> store $ liftIO $ setUUID r $
genUUIDInNameSpace gCryptNameSpace v
2013-04-24 00:06:02 +00:00
2014-07-15 18:45:27 +00:00
{- The local repo may not yet be initialized, so try to initialize
- it if allowed . However , if that fails , still return the read
- git config . - }
2014-07-15 18:27:43 +00:00
readlocalannexconfig = do
2017-09-30 02:36:08 +00:00
let check = do
2014-07-15 18:27:43 +00:00
Annex . BranchState . disableUpdate
unify exception handling into Utility.Exception
Removed old extensible-exceptions, only needed for very old ghc.
Made webdav use Utility.Exception, to work after some changes in DAV's
exception handling.
Removed Annex.Exception. Mostly this was trivial, but note that
tryAnnex is replaced with tryNonAsync and catchAnnex replaced with
catchNonAsync. In theory that could be a behavior change, since the former
caught all exceptions, and the latter don't catch async exceptions.
However, in practice, nothing in the Annex monad uses async exceptions.
Grepping for throwTo and killThread only find stuff in the assistant,
which does not seem related.
Command.Add.undo is changed to accept a SomeException, and things
that use it for rollback now catch non-async exceptions, rather than
only IOExceptions.
2014-08-08 01:55:44 +00:00
void $ tryNonAsync $ ensureInitialized
2014-07-15 18:27:43 +00:00
Annex . getState Annex . repo
2017-09-30 02:36:08 +00:00
s <- Annex . new r
Annex . eval s $ check ` finally ` stopCoProcesses
2015-08-05 17:49:54 +00:00
configlistfields = if autoinit
then [ ( Fields . autoInit , " 1 " ) ]
else []
2014-07-15 18:27:43 +00:00
2014-08-06 17:45:19 +00:00
{- Checks if a given remote has the content for a key in its annex. -}
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
inAnnex :: Remote -> State -> Key -> Annex Bool
inAnnex rmt ( State connpool duc ) key
2014-02-25 01:29:37 +00:00
| Git . repoIsHttp r = checkhttp
2011-08-17 01:04:23 +00:00
| Git . repoIsUrl r = checkremote
2011-11-09 22:33:15 +00:00
| otherwise = checklocal
2012-10-29 01:27:15 +00:00
where
2014-10-09 18:53:13 +00:00
r = repo rmt
2014-02-25 01:29:37 +00:00
checkhttp = do
2013-09-24 17:37:41 +00:00
showChecking r
2014-02-25 02:00:25 +00:00
ifM ( Url . withUrlOptions $ \ uo -> anyM ( \ u -> Url . checkBoth u ( keySize key ) uo ) ( keyUrls rmt key ) )
2014-08-06 17:45:19 +00:00
( return True
2016-11-16 01:29:54 +00:00
, giveup " not found "
2013-07-18 18:11:56 +00:00
)
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
checkremote =
let fallback = Ssh . inAnnex r key
2018-03-09 17:48:10 +00:00
in P2PHelper . checkpresent ( Ssh . runProto rmt connpool fallback ) key
2018-01-10 18:21:18 +00:00
checklocal = ifM duc
( guardUsable r ( cantCheck r ) $
maybe ( cantCheck r ) return
=<< onLocalFast rmt ( Annex . Content . inAnnexSafe key )
, cantCheck r
)
2011-08-17 01:04:23 +00:00
2014-01-26 17:03:25 +00:00
keyUrls :: Remote -> Key -> [ String ]
keyUrls r key = map tourl locs'
2012-10-29 01:27:15 +00:00
where
2014-01-26 17:03:25 +00:00
tourl l = Git . repoLocation ( repo r ) ++ " / " ++ l
-- If the remote is known to not be bare, try the hash locations
-- used for non-bare repos first, as an optimisation.
locs
2015-01-28 20:51:40 +00:00
| remoteAnnexBare remoteconfig == Just False = reverse ( annexLocations cfg key )
| otherwise = annexLocations cfg key
2013-08-02 16:27:32 +00:00
# ifndef mingw32_HOST_OS
2014-01-26 17:03:25 +00:00
locs' = locs
2013-07-07 17:35:06 +00:00
# else
2014-02-25 17:47:09 +00:00
locs' = map ( replace " \ \ " " / " ) locs
2013-07-07 17:35:06 +00:00
# endif
2015-01-28 20:51:40 +00:00
remoteconfig = gitconfig r
2017-05-11 20:03:50 +00:00
cfg = remoteGitConfig remoteconfig
2011-08-17 01:04:23 +00:00
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
dropKey :: Remote -> State -> Key -> Annex Bool
dropKey r ( State connpool duc ) key
2018-01-10 18:21:18 +00:00
| not $ Git . repoIsUrl ( repo r ) = ifM duc
( guardUsable ( repo r ) ( return False ) $
2017-02-17 19:21:39 +00:00
commitOnCleanup r $ onLocalFast r $ do
2014-08-08 23:18:08 +00:00
ensureInitialized
whenM ( Annex . Content . inAnnex key ) $ do
2016-12-09 16:56:38 +00:00
Annex . Content . lockContentForRemoval key $ \ lock -> do
Annex . Content . removeAnnex lock
2016-12-09 16:54:18 +00:00
logStatus key InfoMissing
2014-08-08 23:18:08 +00:00
Annex . Content . saveState True
return True
2018-01-10 18:21:18 +00:00
, return False
)
2016-11-16 01:29:54 +00:00
| Git . repoIsHttp ( repo r ) = giveup " dropping from http remote not supported "
2018-03-08 20:21:16 +00:00
| otherwise = commitOnCleanup r $ do
let fallback = Ssh . dropKey ( repo r ) key
2018-03-09 17:48:10 +00:00
P2PHelper . remove ( Ssh . runProto r connpool fallback ) key
2011-03-27 19:56:43 +00:00
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
lockKey :: Remote -> State -> Key -> ( VerifiedCopy -> Annex r ) -> Annex r
2018-03-09 17:42:55 +00:00
lockKey r ( State connpool duc ) key callback
2018-01-10 18:21:18 +00:00
| not $ Git . repoIsUrl ( repo r ) = ifM duc
( guardUsable ( repo r ) failedlock $ do
2015-10-09 17:35:28 +00:00
inorigrepo <- Annex . makeRunner
-- Lock content from perspective of remote,
-- and then run the callback in the original
-- annex monad, not the remote's.
2017-02-17 19:21:39 +00:00
onLocalFast r $
make sure that lockContentShared is always paired with an inAnnex check
lockContentShared had a screwy caveat that it didn't verify that the content
was present when locking it, but in the most common case, eg indirect mode,
it failed to lock when the content is not present.
That led to a few callers forgetting to check inAnnex when using it,
but the potential data loss was unlikely to be noticed because it only
affected direct mode I think.
Fix data loss bug when the local repository uses direct mode, and a
locally modified file is dropped from a remote repsitory. The bug
caused the modified file to be counted as a copy of the original file.
(This is not a severe bug because in such a situation, dropping
from the remote and then modifying the file is allowed and has the same
end result.)
And, in content locking over tor, when the remote repository is
in direct mode, it neglected to check that the content was actually
present when locking it. This could cause git annex drop to remove
the only copy of a file when it thought the tor remote had a copy.
So, make lockContentShared do its own inAnnex check. This could perhaps
be optimised for direct mode, to avoid the check then, since locking
the content necessarily verifies it exists there, but I have not bothered
with that.
This commit was sponsored by Jeff Goeke-Smith on Patreon.
2018-03-07 18:13:02 +00:00
Annex . Content . lockContentShared key $
liftIO . inorigrepo . callback
2018-01-10 18:21:18 +00:00
, failedlock
)
2015-10-09 20:55:41 +00:00
| Git . repoIsSsh ( repo r ) = do
2015-10-09 21:21:02 +00:00
showLocking r
2018-03-09 17:48:10 +00:00
let withconn = Ssh . withP2PSshConnection r connpool fallback
P2PHelper . lock withconn Ssh . runProtoConn ( uuid r ) key callback
2018-03-09 17:42:55 +00:00
| otherwise = failedlock
where
fallback = do
2017-02-15 19:08:46 +00:00
Just ( cmd , params ) <- Ssh . git_annex_shell ConsumeStdin
( repo r ) " lockcontent "
2015-10-09 20:55:41 +00:00
[ Param $ key2file key ] []
2015-10-09 21:21:02 +00:00
( Just hin , Just hout , Nothing , p ) <- liftIO $
withFile devNull WriteMode $ \ nullh ->
createProcess $
( proc cmd ( toCommand params ) )
{ std_in = CreatePipe
, std_out = CreatePipe
, std_err = UseHandle nullh
}
2016-12-09 17:34:00 +00:00
v <- liftIO $ tryIO $ getProtocolLine hout
2016-04-18 18:04:20 +00:00
let signaldone = void $ tryNonAsync $ liftIO $ mapM_ tryNonAsync
[ hPutStrLn hout " "
, hFlush hout
, hClose hin
, hClose hout
, void $ waitForProcess p
]
2015-10-09 20:55:41 +00:00
let checkexited = not . isJust <$> getProcessExitCode p
case v of
Left _exited -> do
2015-10-09 21:21:02 +00:00
showNote " lockcontent failed "
2015-10-09 20:55:41 +00:00
liftIO $ do
hClose hin
hClose hout
2016-04-18 18:04:20 +00:00
void $ waitForProcess p
2015-10-09 21:21:02 +00:00
failedlock
2015-10-09 20:55:41 +00:00
Right l
2016-12-09 17:34:00 +00:00
| l == Just Ssh . contentLockedMarker -> bracket_
2015-10-09 20:55:41 +00:00
noop
signaldone
( withVerifiedCopy LockedCopy r checkexited callback )
| otherwise -> do
2015-10-09 21:21:02 +00:00
showNote " lockcontent failed "
2015-10-09 20:55:41 +00:00
signaldone
2015-10-09 21:21:02 +00:00
failedlock
2016-11-16 01:29:54 +00:00
failedlock = giveup " can't lock content "
2015-10-09 17:07:03 +00:00
2011-03-27 19:56:43 +00:00
{- Tries to copy a key's content from a remote's annex to a file. -}
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
copyFromRemote :: Remote -> State -> Key -> AssociatedFile -> FilePath -> MeterUpdate -> Annex ( Bool , Verification )
copyFromRemote = copyFromRemote' False
2015-04-10 19:15:01 +00:00
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
copyFromRemote' :: Bool -> Remote -> State -> Key -> AssociatedFile -> FilePath -> MeterUpdate -> Annex ( Bool , Verification )
copyFromRemote' forcersync r ( State connpool _ ) key file dest meterupdate
| Git . repoIsHttp ( repo r ) = unVerified $
Annex . Content . downloadUrl key meterupdate ( keyUrls r key ) dest
other 80% of avoding verification when hard linking to objects in shared repo
In c6632ee5c8e66c26ef18317f56ae02bae1e7e280, it actually only handled
uploading objects to a shared repository. To avoid verification when
downloading objects from a shared repository, was a lot harder.
On the plus side, if the process of downloading a file from a remote
is able to verify its content on the side, the remote can indicate this
now, and avoid the extra post-download verification.
As of yet, I don't have any remotes (except Git) using this ability.
Some more work would be needed to support it in special remotes.
It would make sense for tahoe to implicitly verify things downloaded from it;
as long as you trust your tahoe server (which typically runs locally),
there's cryptographic integrity. OTOH, despite bup being based on shas,
a bup repo under an attacker's control could have the git ref used for an
object changed, and so a bup repo shouldn't implicitly verify. Indeed,
tahoe seems unique in being trustworthy enough to implicitly verify.
2015-10-02 17:56:42 +00:00
| not $ Git . repoIsUrl ( repo r ) = guardUsable ( repo r ) ( unVerified ( return False ) ) $ do
2014-04-17 18:31:42 +00:00
params <- Ssh . rsyncParams r Download
2012-07-01 20:59:54 +00:00
u <- getUUID
2015-09-14 16:13:38 +00:00
hardlink <- wantHardLink
2012-07-01 20:59:54 +00:00
-- run copy from perspective of remote
2017-02-17 19:21:39 +00:00
onLocalFast r $ do
2012-07-01 20:59:54 +00:00
ensureInitialized
2013-01-10 15:52:11 +00:00
v <- Annex . Content . prepSendAnnex key
case v of
other 80% of avoding verification when hard linking to objects in shared repo
In c6632ee5c8e66c26ef18317f56ae02bae1e7e280, it actually only handled
uploading objects to a shared repository. To avoid verification when
downloading objects from a shared repository, was a lot harder.
On the plus side, if the process of downloading a file from a remote
is able to verify its content on the side, the remote can indicate this
now, and avoid the extra post-download verification.
As of yet, I don't have any remotes (except Git) using this ability.
Some more work would be needed to support it in special remotes.
It would make sense for tahoe to implicitly verify things downloaded from it;
as long as you trust your tahoe server (which typically runs locally),
there's cryptographic integrity. OTOH, despite bup being based on shas,
a bup repo under an attacker's control could have the git ref used for an
object changed, and so a bup repo shouldn't implicitly verify. Indeed,
tahoe seems unique in being trustworthy enough to implicitly verify.
2015-10-02 17:56:42 +00:00
Nothing -> return ( False , UnVerified )
2014-09-05 17:44:09 +00:00
Just ( object , checksuccess ) -> do
2015-10-02 16:36:03 +00:00
copier <- mkCopier hardlink params
2014-09-05 17:44:09 +00:00
runTransfer ( Transfer Download u key )
2016-10-26 19:38:22 +00:00
file forwardRetry
2015-11-16 23:32:30 +00:00
( \ p -> copier object dest ( combineMeterUpdate p meterupdate ) checksuccess )
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
| Git . repoIsSsh ( repo r ) = if forcersync
then unVerified fallback
else P2PHelper . retrieve
2018-03-09 17:48:10 +00:00
( Ssh . runProto r connpool fallback )
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
key file dest meterupdate
2016-11-16 01:29:54 +00:00
| otherwise = giveup " copying from non-ssh, non-http remote not supported "
2012-10-29 01:27:15 +00:00
where
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
fallback = feedprogressback $ \ p -> do
Ssh . rsyncHelper ( Just ( combineMeterUpdate meterupdate p ) )
=<< Ssh . rsyncParamsRemote False r Download key dest file
2012-10-29 01:27:15 +00:00
{- Feed local rsync's progress info back to the remote,
- by forking a feeder thread that runs
- git - annex - shell transferinfo at the same time
- git - annex - shell sendkey is running .
-
2013-05-14 17:51:14 +00:00
- To avoid extra password prompts , this is only done when ssh
- connection caching is supported .
2012-10-29 01:27:15 +00:00
- Note that it actually waits for rsync to indicate
- progress before starting transferinfo , in order
- to ensure ssh connection caching works and reuses
- the connection set up for the sendkey .
-
- Also note that older git - annex - shell does not support
- transferinfo , so stderr is dropped and failure ignored .
- }
2013-05-14 17:51:14 +00:00
feedprogressback a = ifM ( isJust <$> sshCacheDir )
( feedprogressback' a
2013-05-19 22:15:29 +00:00
, a $ const noop
2013-05-14 17:52:30 +00:00
)
2013-05-14 17:51:14 +00:00
feedprogressback' a = do
2012-10-29 01:27:15 +00:00
u <- getUUID
2017-03-10 17:12:24 +00:00
let AssociatedFile afile = file
2012-10-29 01:27:15 +00:00
let fields = ( Fields . remoteUUID , fromUUID u )
2017-03-10 17:12:24 +00:00
: maybe [] ( \ f -> [ ( Fields . associatedFile , f ) ] ) afile
2017-02-15 19:08:46 +00:00
Just ( cmd , params ) <- Ssh . git_annex_shell ConsumeStdin
( repo r ) " transferinfo "
2012-10-29 01:27:15 +00:00
[ Param $ key2file key ] fields
2013-09-26 03:19:01 +00:00
v <- liftIO ( newEmptySV :: IO ( MSampleVar Integer ) )
2014-08-04 00:14:20 +00:00
pidv <- liftIO $ newEmptyMVar
2012-10-29 01:27:15 +00:00
tid <- liftIO $ forkIO $ void $ tryIO $ do
bytes <- readSV v
p <- createProcess $
( proc cmd ( toCommand params ) )
{ std_in = CreatePipe
, std_err = CreatePipe
}
2014-08-04 00:14:20 +00:00
putMVar pidv ( processHandle p )
2012-10-29 01:27:15 +00:00
hClose $ stderrHandle p
let h = stdinHandle p
let send b = do
2013-09-26 03:19:01 +00:00
hPrint h b
2012-10-29 01:27:15 +00:00
hFlush h
send bytes
forever $
send =<< readSV v
2015-04-10 19:15:01 +00:00
let feeder = \ n -> do
meterupdate n
writeSV v ( fromBytesProcessed n )
2015-08-13 18:20:28 +00:00
-- It can easily take 0.3 seconds to clean up after
-- the transferinfo, and all that's involved is shutting
-- down the process and associated thread cleanly. So,
-- do it in the background.
let cleanup = forkIO $ do
2014-08-04 00:14:20 +00:00
void $ tryIO $ killThread tid
2015-08-13 18:20:28 +00:00
void $ tryNonAsync $
2014-08-04 00:14:20 +00:00
maybe noop ( void . waitForProcess )
=<< tryTakeMVar pidv
bracketIO noop ( const cleanup ) ( const $ a feeder )
2011-08-17 01:04:23 +00:00
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
copyFromRemoteCheap :: Remote -> State -> Key -> AssociatedFile -> FilePath -> Annex Bool
2013-08-02 16:27:32 +00:00
# ifndef mingw32_HOST_OS
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
copyFromRemoteCheap r st key af file
2015-04-18 17:36:12 +00:00
| not $ Git . repoIsUrl ( repo r ) = guardUsable ( repo r ) ( return False ) $ liftIO $ do
loc <- gitAnnexLocation key ( repo r ) $
2017-05-11 20:03:50 +00:00
remoteGitConfig $ gitconfig r
2015-04-18 17:36:12 +00:00
ifM ( doesFileExist loc )
( do
absloc <- absPath loc
catchBoolIO $ do
createSymbolicLink absloc file
return True
, return False
)
2013-01-01 17:52:47 +00:00
| Git . repoIsSsh ( repo r ) =
2012-03-16 00:39:25 +00:00
ifM ( Annex . Content . preseedTmp key file )
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
( fst <$> copyFromRemote' True r st key af file nullMeterUpdate
2012-03-16 00:39:25 +00:00
, return False
)
2012-01-20 17:23:11 +00:00
| otherwise = return False
2013-08-04 17:12:18 +00:00
# else
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
copyFromRemoteCheap _ _ _ _ _ = return False
2013-08-04 17:12:18 +00:00
# endif
2012-01-20 17:23:11 +00:00
2011-03-27 19:56:43 +00:00
{- Tries to copy a key's content to a remote's annex. -}
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
copyToRemote :: Remote -> State -> Key -> AssociatedFile -> MeterUpdate -> Annex Bool
copyToRemote r ( State connpool duc ) key file meterupdate
2018-01-10 18:21:18 +00:00
| not $ Git . repoIsUrl ( repo r ) = ifM duc
( guardUsable ( repo r ) ( return False ) $ commitOnCleanup r $
2013-01-10 15:45:44 +00:00
copylocal =<< Annex . Content . prepSendAnnex key
2018-01-10 18:21:18 +00:00
, return False
)
2013-01-01 17:52:47 +00:00
| Git . repoIsSsh ( repo r ) = commitOnCleanup r $
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
P2PHelper . store
2018-03-09 17:48:10 +00:00
( Ssh . runProto r connpool copyremotefallback )
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
key file meterupdate
2016-11-16 01:29:54 +00:00
| otherwise = giveup " copying to non-ssh repo not supported "
2012-12-08 21:03:39 +00:00
where
2013-01-10 15:45:44 +00:00
copylocal Nothing = return False
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
copylocal ( Just ( object , checksuccess ) ) = do
2013-03-12 20:41:54 +00:00
-- The checksuccess action is going to be run in
deal with unlocked files when calling rsyncParamsRemote
In copyFromRemote, it used to check isDirect, but that was not needed;
the remote is sending the file, so it doesn't matter if the local,
receiving repository is in direct mode or not. And, since the content is not
present, yet, it's certianly not unlocked. Note that, the remote may indeed
be sending an unlocked file, but sendkey uses sendAnnex, which will detect
if the file is modified before or during transfer, and will exit nonzero,
aborting the upload. So, the receiver doesn't need any checks.
In copyToRemote, it forces recvkey to verify content whenever it's being
sent from a v6 repository. recvkey is almost always going to verify content
anyway, unless annex.verify is not set. So, this doesn't make it any more
expensive, except for in that unusual configuration. The alternative would
be to change the recvkey interface, so that the sender checks afterwards if
what it was sending changed, and the receiver then throws out the bad
transfer. That would be less expensive for the reciever, as it would not
need to do a checksum verification. But, it would mean another network
round trip, and since rsync closes the connection, it would need to open
another ssh connection to do this. Even with connction caching, that would
add latency to uploads. It would also complicate the interface, especially
because an older git-annex-shell would not have the new interface
available. For these reasons, I prefer punting on that at this time, and
instead someone might set annex.verify=false and be unhappy that it still
verifies..
(One other gotcha not dealt with is that a v5 repo could be upgraded to v6
while an upload is in progress, and a file unlocked and modified.)
(Also, I double-checked Remote.GCrypt's calls to rsyncParamsRemote, and
they're fine. When a file is being uploaded to gcrypt, or any other special
repository, it is mediated by sendAnnex, so changes will be detected at
that level and the special remote implementation doesn't need to worry
about them.)
2015-12-26 18:05:07 +00:00
-- the remote's Annex, but it needs access to the local
2013-03-12 20:41:54 +00:00
-- Annex monad's state.
checksuccessio <- Annex . withCurrentState checksuccess
2014-04-17 18:31:42 +00:00
params <- Ssh . rsyncParams r Upload
2012-07-01 20:59:54 +00:00
u <- getUUID
2015-09-14 16:13:38 +00:00
hardlink <- wantHardLink
2011-03-27 19:56:43 +00:00
-- run copy from perspective of remote
2017-02-17 19:21:39 +00:00
onLocalFast r $ ifM ( Annex . Content . inAnnex key )
2013-03-10 21:54:27 +00:00
( return True
2012-09-18 17:59:03 +00:00
, do
ensureInitialized
2015-10-02 16:36:03 +00:00
copier <- mkCopier hardlink params
other 80% of avoding verification when hard linking to objects in shared repo
In c6632ee5c8e66c26ef18317f56ae02bae1e7e280, it actually only handled
uploading objects to a shared repository. To avoid verification when
downloading objects from a shared repository, was a lot harder.
On the plus side, if the process of downloading a file from a remote
is able to verify its content on the side, the remote can indicate this
now, and avoid the extra post-download verification.
As of yet, I don't have any remotes (except Git) using this ability.
Some more work would be needed to support it in special remotes.
It would make sense for tahoe to implicitly verify things downloaded from it;
as long as you trust your tahoe server (which typically runs locally),
there's cryptographic integrity. OTOH, despite bup being based on shas,
a bup repo under an attacker's control could have the git ref used for an
object changed, and so a bup repo shouldn't implicitly verify. Indeed,
tahoe seems unique in being trustworthy enough to implicitly verify.
2015-10-02 17:56:42 +00:00
let verify = Annex . Content . RemoteVerify r
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
runTransfer ( Transfer Download u key ) file forwardRetry $ \ p ->
let p' = combineMeterUpdate meterupdate p
2015-11-16 23:32:30 +00:00
in Annex . Content . saveState True ` after `
2015-10-02 16:36:03 +00:00
Annex . Content . getViaTmp verify key
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
( \ dest -> copier object dest p' ( liftIO checksuccessio ) )
2012-09-18 17:59:03 +00:00
)
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
copyremotefallback = Annex . Content . sendAnnex key noop $ \ object -> do
-- This is too broad really, but recvkey normally
-- verifies content anyway, so avoid complicating
-- it with a local sendAnnex check and rollback.
unlocked <- isDirect <||> versionSupportsUnlockedPointers
Ssh . rsyncHelper ( Just meterupdate )
=<< Ssh . rsyncParamsRemote unlocked r Upload key object file
2011-03-27 19:56:43 +00:00
2013-10-11 20:03:18 +00:00
fsckOnRemote :: Git . Repo -> [ CommandParam ] -> Annex ( IO Bool )
fsckOnRemote r params
2013-10-14 16:23:38 +00:00
| Git . repoIsUrl r = do
2017-02-15 19:08:46 +00:00
s <- Ssh . git_annex_shell NoConsumeStdin r " fsck " params []
2013-10-11 20:03:18 +00:00
return $ case s of
Nothing -> return False
Just ( c , ps ) -> batchCommand c ps
2013-10-14 16:23:38 +00:00
| otherwise = return $ do
2015-02-28 21:23:13 +00:00
program <- programPath
2013-10-14 19:05:10 +00:00
r' <- Git . Config . read r
2014-06-10 23:20:14 +00:00
environ <- getEnvironment
let environ' = addEntries
2013-10-14 19:05:10 +00:00
[ ( " GIT_WORK_TREE " , Git . repoPath r' )
, ( " GIT_DIR " , Git . localGitDir r' )
2014-06-10 23:20:14 +00:00
] environ
2015-02-09 18:16:42 +00:00
batchCommandEnv program ( Param " fsck " : params ) ( Just environ' )
2013-10-11 20:03:18 +00:00
2013-10-27 19:38:59 +00:00
{- The passed repair action is run in the Annex monad of the remote. -}
repairRemote :: Git . Repo -> Annex Bool -> Annex ( IO Bool )
2014-03-06 21:12:50 +00:00
repairRemote r a = return $ do
2013-09-07 22:38:00 +00:00
s <- Annex . new r
Annex . eval s $ do
Annex . BranchState . disableUpdate
2014-03-06 21:12:50 +00:00
ensureInitialized
2017-09-30 02:36:08 +00:00
a ` finally ` stopCoProcesses
2011-03-27 19:56:43 +00:00
2014-03-06 21:12:50 +00:00
{- Runs an action from the perspective of a local remote.
-
- The AnnexState is cached for speed and to avoid resource leaks .
2017-02-17 19:21:39 +00:00
- However , coprocesses are stopped after each call to avoid git
- processes hanging around on removable media .
2014-03-06 21:12:50 +00:00
- }
onLocal :: Remote -> Annex a -> Annex a
onLocal r a = do
m <- Annex . getState Annex . remoteannexstate
2017-02-17 19:21:39 +00:00
go =<< maybe
( liftIO $ Annex . new $ repo r )
return
( M . lookup ( uuid r ) m )
2014-03-06 21:12:50 +00:00
where
cache st = Annex . changeState $ \ s -> s
{ Annex . remoteannexstate = M . insert ( uuid r ) st ( Annex . remoteannexstate s ) }
2017-02-17 19:21:39 +00:00
go st = do
2015-04-04 00:08:38 +00:00
curro <- Annex . getState Annex . output
( ret , st' ) <- liftIO $ Annex . run ( st { Annex . output = curro } ) $
2017-09-30 02:36:08 +00:00
a ` finally ` stopCoProcesses
2014-03-06 21:12:50 +00:00
cache st'
return ret
2017-02-17 19:21:39 +00:00
{- Faster variant of onLocal.
-
- The repository's git - annex branch is not updated , as an optimisation .
- No caller of onLocalFast can query data from the branch and be ensured
- it gets the most current value . Caller of onLocalFast can make changes
- to the branch , however .
- }
onLocalFast :: Remote -> Annex a -> Annex a
onLocalFast r a = onLocal r $ Annex . BranchState . disableUpdate >> a
2011-06-14 00:23:47 +00:00
{- Copys a file with rsync unless both locations are on the same
- filesystem . Then cp could be faster . - }
2012-09-21 18:50:14 +00:00
rsyncOrCopyFile :: [ CommandParam ] -> FilePath -> FilePath -> MeterUpdate -> Annex Bool
2012-09-19 20:08:37 +00:00
rsyncOrCopyFile rsyncparams src dest p =
2013-08-02 16:27:32 +00:00
# ifdef mingw32_HOST_OS
2013-05-11 20:03:00 +00:00
dorsync
where
# else
2012-09-22 00:24:08 +00:00
ifM ( sameDeviceIds src dest ) ( docopy , dorsync )
2012-10-29 01:27:15 +00:00
where
2013-09-26 03:19:01 +00:00
sameDeviceIds a b = ( == ) <$> getDeviceId a <*> getDeviceId b
2015-01-09 17:11:56 +00:00
getDeviceId f = deviceID <$> liftIO ( getFileStatus $ parentDir f )
2015-11-17 00:27:01 +00:00
docopy = liftIO $ watchFileSize dest p $
copyFileExternal CopyTimeStamps src dest
2013-05-11 20:03:00 +00:00
# endif
2013-09-24 17:37:41 +00:00
dorsync = Ssh . rsyncHelper ( Just p ) $
2013-05-14 17:24:15 +00:00
rsyncparams ++ [ File src , File dest ]
2011-06-14 00:23:47 +00:00
2013-01-01 17:52:47 +00:00
commitOnCleanup :: Remote -> Annex a -> Annex a
2012-02-25 22:02:49 +00:00
commitOnCleanup r a = go ` after ` a
2012-10-29 01:27:15 +00:00
where
2014-03-13 23:06:26 +00:00
go = Annex . addCleanup ( RemoteCleanup $ uuid r ) cleanup
2012-10-29 01:27:15 +00:00
cleanup
2017-02-17 19:21:39 +00:00
| not $ Git . repoIsUrl ( repo r ) = onLocalFast r $
2012-10-29 01:27:15 +00:00
doQuietSideAction $
Annex . Branch . commit " update "
| otherwise = void $ do
Just ( shellcmd , shellparams ) <-
2017-02-15 19:08:46 +00:00
Ssh . git_annex_shell NoConsumeStdin
( repo r ) " commit " [] []
2012-10-29 01:27:15 +00:00
-- Throw away stderr, since the remote may not
-- have a new enough git-annex shell to
-- support committing.
2013-09-26 03:19:01 +00:00
liftIO $ catchMaybeIO $
2012-10-29 01:27:15 +00:00
withQuietOutput createProcessSuccess $
proc shellcmd $
toCommand shellparams
2015-09-14 16:13:38 +00:00
wantHardLink :: Annex Bool
2016-01-13 18:19:31 +00:00
wantHardLink = ( annexHardLink <$> Annex . getGitConfig )
-- Not direct mode files because they can be modified at any time.
<&&> ( not <$> isDirect )
-- Not unlocked files that are hard linked in the work tree,
-- because they can be modified at any time.
<&&> ( not <$> annexThin <$> Annex . getGitConfig )
2015-09-14 16:13:38 +00:00
other 80% of avoding verification when hard linking to objects in shared repo
In c6632ee5c8e66c26ef18317f56ae02bae1e7e280, it actually only handled
uploading objects to a shared repository. To avoid verification when
downloading objects from a shared repository, was a lot harder.
On the plus side, if the process of downloading a file from a remote
is able to verify its content on the side, the remote can indicate this
now, and avoid the extra post-download verification.
As of yet, I don't have any remotes (except Git) using this ability.
Some more work would be needed to support it in special remotes.
It would make sense for tahoe to implicitly verify things downloaded from it;
as long as you trust your tahoe server (which typically runs locally),
there's cryptographic integrity. OTOH, despite bup being based on shas,
a bup repo under an attacker's control could have the git ref used for an
object changed, and so a bup repo shouldn't implicitly verify. Indeed,
tahoe seems unique in being trustworthy enough to implicitly verify.
2015-10-02 17:56:42 +00:00
-- Copies from src to dest, updating a meter. If the copy finishes
2015-12-10 18:29:34 +00:00
-- successfully, calls a final check action, which must also succeed, or
other 80% of avoding verification when hard linking to objects in shared repo
In c6632ee5c8e66c26ef18317f56ae02bae1e7e280, it actually only handled
uploading objects to a shared repository. To avoid verification when
downloading objects from a shared repository, was a lot harder.
On the plus side, if the process of downloading a file from a remote
is able to verify its content on the side, the remote can indicate this
now, and avoid the extra post-download verification.
As of yet, I don't have any remotes (except Git) using this ability.
Some more work would be needed to support it in special remotes.
It would make sense for tahoe to implicitly verify things downloaded from it;
as long as you trust your tahoe server (which typically runs locally),
there's cryptographic integrity. OTOH, despite bup being based on shas,
a bup repo under an attacker's control could have the git ref used for an
object changed, and so a bup repo shouldn't implicitly verify. Indeed,
tahoe seems unique in being trustworthy enough to implicitly verify.
2015-10-02 17:56:42 +00:00
-- returns false.
--
2015-09-14 16:13:38 +00:00
-- If either the remote or local repository wants to use hard links,
other 80% of avoding verification when hard linking to objects in shared repo
In c6632ee5c8e66c26ef18317f56ae02bae1e7e280, it actually only handled
uploading objects to a shared repository. To avoid verification when
downloading objects from a shared repository, was a lot harder.
On the plus side, if the process of downloading a file from a remote
is able to verify its content on the side, the remote can indicate this
now, and avoid the extra post-download verification.
As of yet, I don't have any remotes (except Git) using this ability.
Some more work would be needed to support it in special remotes.
It would make sense for tahoe to implicitly verify things downloaded from it;
as long as you trust your tahoe server (which typically runs locally),
there's cryptographic integrity. OTOH, despite bup being based on shas,
a bup repo under an attacker's control could have the git ref used for an
object changed, and so a bup repo shouldn't implicitly verify. Indeed,
tahoe seems unique in being trustworthy enough to implicitly verify.
2015-10-02 17:56:42 +00:00
-- the copier will do so (falling back to copying if a hard link cannot be
-- made).
--
-- When a hard link is created, returns Verified; the repo being linked
-- from is implicitly trusted, so no expensive verification needs to be
-- done.
type Copier = FilePath -> FilePath -> MeterUpdate -> Annex Bool -> Annex ( Bool , Verification )
2015-10-02 16:36:03 +00:00
mkCopier :: Bool -> [ CommandParam ] -> Annex Copier
mkCopier remotewanthardlink rsyncparams = do
other 80% of avoding verification when hard linking to objects in shared repo
In c6632ee5c8e66c26ef18317f56ae02bae1e7e280, it actually only handled
uploading objects to a shared repository. To avoid verification when
downloading objects from a shared repository, was a lot harder.
On the plus side, if the process of downloading a file from a remote
is able to verify its content on the side, the remote can indicate this
now, and avoid the extra post-download verification.
As of yet, I don't have any remotes (except Git) using this ability.
Some more work would be needed to support it in special remotes.
It would make sense for tahoe to implicitly verify things downloaded from it;
as long as you trust your tahoe server (which typically runs locally),
there's cryptographic integrity. OTOH, despite bup being based on shas,
a bup repo under an attacker's control could have the git ref used for an
object changed, and so a bup repo shouldn't implicitly verify. Indeed,
tahoe seems unique in being trustworthy enough to implicitly verify.
2015-10-02 17:56:42 +00:00
let copier = \ src dest p check -> unVerified $
rsyncOrCopyFile rsyncparams src dest p <&&> check
2015-09-14 16:13:38 +00:00
localwanthardlink <- wantHardLink
other 80% of avoding verification when hard linking to objects in shared repo
In c6632ee5c8e66c26ef18317f56ae02bae1e7e280, it actually only handled
uploading objects to a shared repository. To avoid verification when
downloading objects from a shared repository, was a lot harder.
On the plus side, if the process of downloading a file from a remote
is able to verify its content on the side, the remote can indicate this
now, and avoid the extra post-download verification.
As of yet, I don't have any remotes (except Git) using this ability.
Some more work would be needed to support it in special remotes.
It would make sense for tahoe to implicitly verify things downloaded from it;
as long as you trust your tahoe server (which typically runs locally),
there's cryptographic integrity. OTOH, despite bup being based on shas,
a bup repo under an attacker's control could have the git ref used for an
object changed, and so a bup repo shouldn't implicitly verify. Indeed,
tahoe seems unique in being trustworthy enough to implicitly verify.
2015-10-02 17:56:42 +00:00
let linker = \ src dest -> createLink src dest >> return True
2016-01-13 18:19:31 +00:00
ifM ( pure ( remotewanthardlink || localwanthardlink ) <&&> not <$> isDirect )
other 80% of avoding verification when hard linking to objects in shared repo
In c6632ee5c8e66c26ef18317f56ae02bae1e7e280, it actually only handled
uploading objects to a shared repository. To avoid verification when
downloading objects from a shared repository, was a lot harder.
On the plus side, if the process of downloading a file from a remote
is able to verify its content on the side, the remote can indicate this
now, and avoid the extra post-download verification.
As of yet, I don't have any remotes (except Git) using this ability.
Some more work would be needed to support it in special remotes.
It would make sense for tahoe to implicitly verify things downloaded from it;
as long as you trust your tahoe server (which typically runs locally),
there's cryptographic integrity. OTOH, despite bup being based on shas,
a bup repo under an attacker's control could have the git ref used for an
object changed, and so a bup repo shouldn't implicitly verify. Indeed,
tahoe seems unique in being trustworthy enough to implicitly verify.
2015-10-02 17:56:42 +00:00
( return $ \ src dest p check ->
ifM ( liftIO ( catchBoolIO ( linker src dest ) ) )
( return ( True , Verified )
, copier src dest p check
)
, return copier
2015-09-14 16:13:38 +00:00
)
2018-01-10 18:21:18 +00:00
use P2P protocol for checkpresent, retrieve, and store
Note that, due to not using rsync to transfer files to ssh remotes
any longer, permissions and other file metadata of annexed files
will no longer be preserved when copying them to ssh remotes.
Other remotes never supported preserving that information, so
this is not considered a regression. Added NEWS item about this.
Another significant side effect of this is that, even when rsync is run to
retrieve a file, its progress display will no longer be shown, and
instead the native git-annex progress display will appear. It would be
possible to use the rsync process display when rsync is used (old
git-annex-shell and also retrieval from a local repository), but it
would have complicated the code unncessarily, and been inconsistent
behavior.
(I'd been thinking for a while about eliminating the rsync progress
display, since it's got some annoying verbosities, including display of
the key and the "(xfr#1, to-chk=0/1)" bit and was already somewhat
inconsistent.)
retrieveKeyFileCheap still uses rsync, since that ensures that it gets
the actual file content from the remote. Using the P2P protocol would
use the local content, as long as the local and remote size are the
same.
This commit was sponsored by John Pellman on Patreon.
2018-03-09 16:57:32 +00:00
data State = State Ssh . P2PSshConnectionPool DeferredUUIDCheck
mkState :: Git . Repo -> UUID -> RemoteGitConfig -> Annex State
mkState r u gc = State
<$> Ssh . mkP2PSshConnectionPool
<*> mkDeferredUUIDCheck r u gc
2018-03-08 18:02:18 +00:00
{- Normally the UUID of a local repository is checked at startup,
- but annex - checkuuid config can prevent that . To avoid getting
- confused , a deferred check is done just before the repository
- is used .
- This returns False when the repository UUID is not as expected . - }
2018-01-10 18:21:18 +00:00
type DeferredUUIDCheck = Annex Bool
mkDeferredUUIDCheck :: Git . Repo -> UUID -> RemoteGitConfig -> Annex DeferredUUIDCheck
mkDeferredUUIDCheck r u gc
| remoteAnnexCheckUUID gc = return ( return True )
| otherwise = do
v <- liftIO newEmptyMVar
return $ ifM ( liftIO $ isEmptyMVar v )
( do
r' <- tryGitConfigRead False r
u' <- getRepoUUID r'
let ok = u' == u
void $ liftIO $ tryPutMVar v ok
unless ok $
warning $ Git . repoDescribe r ++ " is not the expected repository. The remote's annex-checkuuid configuration prevented noticing the change until now. "
return ok
, liftIO $ readMVar v
)