git-annex/Command/AddUrl.hs

537 lines
20 KiB
Haskell
Raw Normal View History

2011-07-01 21:15:46 +00:00
{- git-annex command
-
- Copyright 2011-2020 Joey Hess <id@joeyh.name>
2011-07-01 21:15:46 +00:00
-
- Licensed under the GNU AGPL version 3 or higher.
2011-07-01 21:15:46 +00:00
-}
module Command.AddUrl where
import Command
import Backend
import qualified Annex
import qualified Annex.Url as Url
import qualified Backend.URL
import qualified Remote
import qualified Types.Remote as Remote
2015-12-22 17:23:33 +00:00
import qualified Command.Add
2011-10-04 04:40:47 +00:00
import Annex.Content
2015-12-22 17:23:33 +00:00
import Annex.Ingest
import Annex.CheckIgnore
import Annex.Perms
2014-12-17 17:57:52 +00:00
import Annex.UUID
import Annex.YoutubeDl
import Annex.UntrustedFilePath
import Logs.Web
import Types.KeySource
import Types.UrlContents
import Annex.FileMatcher
import Logs.Location
import Utility.Metered
2017-11-28 21:17:40 +00:00
import Utility.HtmlDetect
import Utility.Path.Max
import qualified Utility.RawFilePath as R
import qualified Annex.Transfer as Transfer
2011-07-01 21:15:46 +00:00
import Network.URI
import qualified System.FilePath.ByteString as P
cmd :: Command
cmd = notBareRepo $ withGlobalOptions [jobsOption, jsonOptions, jsonProgressOption] $
command "addurl" SectionCommon "add urls to annex"
2015-07-13 14:57:49 +00:00
(paramRepeating paramUrl) (seek <$$> optParser)
2015-07-13 14:57:49 +00:00
data AddUrlOptions = AddUrlOptions
{ addUrls :: CmdParams
, pathdepthOption :: Maybe Int
, prefixOption :: Maybe String
, suffixOption :: Maybe String
, downloadOptions :: DownloadOptions
2015-12-21 16:57:13 +00:00
, batchOption :: BatchMode
2015-12-22 16:20:39 +00:00
, batchFilesOption :: Bool
2015-07-13 14:57:49 +00:00
}
2011-07-01 21:15:46 +00:00
data DownloadOptions = DownloadOptions
{ relaxedOption :: Bool
, rawOption :: Bool
, fileOption :: Maybe FilePath
, preserveFilenameOption :: Bool
, checkGitIgnoreOption :: CheckGitIgnore
}
2015-07-13 14:57:49 +00:00
optParser :: CmdParamsDesc -> Parser AddUrlOptions
optParser desc = AddUrlOptions
<$> cmdParams desc
<*> optional (option auto
( long "pathdepth" <> metavar paramNumber
<> help "number of url path components to use in filename"
))
<*> optional (strOption
( long "prefix" <> metavar paramValue
<> help "add a prefix to the filename"
))
<*> optional (strOption
( long "suffix" <> metavar paramValue
<> help "add a suffix to the filename"
2015-07-13 14:57:49 +00:00
))
<*> parseDownloadOptions True
2015-12-21 16:57:13 +00:00
<*> parseBatchOption
2015-12-22 16:20:39 +00:00
<*> switch
( long "with-files"
<> help "parse batch mode lines of the form \"$url $file\""
)
2015-07-13 15:06:41 +00:00
parseDownloadOptions :: Bool -> Parser DownloadOptions
parseDownloadOptions withfileoptions = DownloadOptions
<$> switch
( long "relaxed"
<> help "skip size check"
)
<*> switch
( long "raw"
<> help "disable special handling for torrents, youtube-dl, etc"
)
<*> (if withfileoptions
then optional (strOption
( long "file" <> metavar paramFile
<> help "specify what file the url is added to"
))
else pure Nothing)
<*> (if withfileoptions
then switch
( long "preserve-filename"
<> help "use filename provided by server as-is"
)
else pure False)
<*> Command.Add.checkGitIgnoreSwitch
2012-02-16 16:25:19 +00:00
2015-07-13 14:57:49 +00:00
seek :: AddUrlOptions -> CommandSeek
seek o = startConcurrency commandStages $ do
addunlockedmatcher <- addUnlockedMatcher
let go (si, (o', u)) = do
r <- Remote.claimingUrl u
if Remote.uuid r == webUUID || rawOption (downloadOptions o')
then void $ commandAction $
startWeb addunlockedmatcher o' si u
else checkUrl addunlockedmatcher r o' si u
forM_ (addUrls o) (\u -> go (SeekInput [u], (o, u)))
2015-12-21 16:57:13 +00:00
case batchOption o of
Batch fmt -> batchInput fmt (pure . parseBatchInput o) go
2015-12-21 16:57:13 +00:00
NoBatch -> noop
2015-12-22 16:20:39 +00:00
parseBatchInput :: AddUrlOptions -> String -> Either String (AddUrlOptions, URLString)
parseBatchInput o s
| batchFilesOption o =
let (u, f) = separate (== ' ') s
in if null u || null f
then Left ("parsed empty url or filename in input: " ++ s)
else Right (o { downloadOptions = (downloadOptions o) { fileOption = Just f } }, u)
2015-12-22 16:20:39 +00:00
| otherwise = Right (o, s)
checkUrl :: AddUnlockedMatcher -> Remote -> AddUrlOptions -> SeekInput -> URLString -> Annex ()
checkUrl addunlockedmatcher r o si u = do
pathmax <- liftIO $ fileNameLengthLimit "."
let deffile = fromMaybe (urlString2file u (pathdepthOption o) pathmax) (fileOption (downloadOptions o))
go deffile =<< maybe
(error $ "unable to checkUrl of " ++ Remote.name r)
(tryNonAsync . flip id u)
(Remote.checkUrl r)
where
go _ (Left e) = void $ commandAction $ startingAddUrl si u o $ do
warning (show e)
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
next $ return False
go deffile (Right (UrlContents sz mf)) = do
f <- maybe (pure deffile) (sanitizeOrPreserveFilePath o) mf
let f' = adjustFile o (fromMaybe f (fileOption (downloadOptions o)))
void $ commandAction $ startRemote addunlockedmatcher r o si f' u sz
go deffile (Right (UrlMulti l)) = case fileOption (downloadOptions o) of
Nothing ->
forM_ l $ \(u', sz, f) -> do
f' <- sanitizeOrPreserveFilePath o f
let f'' = adjustFile o (deffile </> f')
void $ commandAction $ startRemote addunlockedmatcher r o si f'' u' sz
Just f -> case l of
[] -> noop
((u',sz,_):[]) -> do
let f' = adjustFile o f
void $ commandAction $ startRemote addunlockedmatcher r o si f' u' sz
_ -> giveup $ unwords
[ "That url contains multiple files according to the"
, Remote.name r
, " remote; cannot add it to a single file."
]
startRemote :: AddUnlockedMatcher -> Remote -> AddUrlOptions -> SeekInput -> FilePath -> URLString -> Maybe Integer -> CommandStart
startRemote addunlockedmatcher r o si file uri sz = do
2014-12-12 00:10:45 +00:00
pathmax <- liftIO $ fileNameLengthLimit "."
let file' = joinPath $ map (truncateFilePath pathmax) $
splitDirectories file
startingAddUrl si uri o $ do
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
showNote $ "from " ++ Remote.name r
showDestinationFile file'
performRemote addunlockedmatcher r o uri (toRawFilePath file') sz
performRemote :: AddUnlockedMatcher -> Remote -> AddUrlOptions -> URLString -> RawFilePath -> Maybe Integer -> CommandPerform
performRemote addunlockedmatcher r o uri file sz = ifAnnexed file adduri geturi
where
loguri = setDownloader uri OtherDownloader
adduri = addUrlChecked o loguri file (Remote.uuid r) checkexistssize
checkexistssize key = return $ case sz of
Nothing -> (True, True, loguri)
Just n -> (True, n == fromMaybe n (fromKey keySize key), loguri)
geturi = next $ isJust <$> downloadRemoteFile addunlockedmatcher r (downloadOptions o) uri file sz
downloadRemoteFile :: AddUnlockedMatcher -> Remote -> DownloadOptions -> URLString -> RawFilePath -> Maybe Integer -> Annex (Maybe Key)
downloadRemoteFile addunlockedmatcher r o uri file sz = checkCanAdd o file $ \canadd -> do
let urlkey = Backend.URL.fromUrl uri sz
createWorkTreeDirectory (parentDir file)
ifM (Annex.getState Annex.fast <||> pure (relaxedOption o))
( do
addWorkTree canadd addunlockedmatcher (Remote.uuid r) loguri file urlkey Nothing
return (Just urlkey)
, do
-- Set temporary url for the urlkey
-- so that the remote knows what url it
-- should use to download it.
2014-12-17 18:34:42 +00:00
setTempUrl urlkey loguri
let downloader = \dest p ->
2020-05-14 18:19:28 +00:00
fst <$> Remote.verifiedAction (Remote.retrieveKeyFile r urlkey af dest p)
ret <- downloadWith canadd addunlockedmatcher downloader urlkey (Remote.uuid r) loguri file
removeTempUrl urlkey
return ret
)
where
loguri = setDownloader uri OtherDownloader
af = AssociatedFile (Just file)
startWeb :: AddUnlockedMatcher -> AddUrlOptions -> SeekInput -> URLString -> CommandStart
startWeb addunlockedmatcher o si urlstring = go $ fromMaybe bad $ parseURI urlstring
2012-11-12 05:05:04 +00:00
where
bad = fromMaybe (giveup $ "bad url " ++ urlstring) $
Url.parseURIRelaxed $ urlstring
go url = startingAddUrl si urlstring o $
if relaxedOption (downloadOptions o)
then go' url Url.assumeUrlExists
else Url.withUrlOptions (Url.getUrlInfo urlstring) >>= \case
Right urlinfo -> go' url urlinfo
Left err -> do
warning err
next $ return False
go' url urlinfo = do
2013-09-09 06:16:22 +00:00
pathmax <- liftIO $ fileNameLengthLimit "."
file <- adjustFile o <$> case fileOption (downloadOptions o) of
Just f -> pure f
Nothing -> case Url.urlSuggestedFile urlinfo of
Just sf -> do
f <- sanitizeOrPreserveFilePath o sf
if preserveFilenameOption (downloadOptions o)
then pure f
else ifM (liftIO $ doesFileExist f <||> doesDirectoryExist f)
( pure $ url2file url (pathdepthOption o) pathmax
, pure f
)
_ -> pure $ url2file url (pathdepthOption o) pathmax
performWeb addunlockedmatcher o urlstring (toRawFilePath file) urlinfo
sanitizeOrPreserveFilePath :: AddUrlOptions -> FilePath -> Annex FilePath
sanitizeOrPreserveFilePath o f
| preserveFilenameOption (downloadOptions o) && not (null f) = do
checkPreserveFileNameSecurity f
return f
| otherwise = do
pathmax <- liftIO $ fileNameLengthLimit "."
return $ truncateFilePath pathmax $ sanitizeFilePath f
-- sanitizeFilePath avoids all these security problems
-- (and probably others, but at least this catches the most egrarious ones).
checkPreserveFileNameSecurity :: FilePath -> Annex ()
checkPreserveFileNameSecurity f = do
checksecurity escapeSequenceInFilePath False "escape sequence"
checksecurity pathTraversalInFilePath True "path traversal"
checksecurity gitDirectoryInFilePath True "contains a .git directory"
where
checksecurity p canshow d = when (p f) $
giveup $ concat
[ "--preserve-filename was used, but the filename "
, if canshow then "(" ++ f ++ ") " else ""
, "has a security problem (" ++ d ++ "), not adding."
]
performWeb :: AddUnlockedMatcher -> AddUrlOptions -> URLString -> RawFilePath -> Url.UrlInfo -> CommandPerform
performWeb addunlockedmatcher o url file urlinfo = ifAnnexed file addurl geturl
2014-12-11 20:11:38 +00:00
where
geturl = next $ isJust <$> addUrlFile addunlockedmatcher (downloadOptions o) url urlinfo file
addurl = addUrlChecked o url file webUUID $ \k ->
ifM (pure (not (rawOption (downloadOptions o))) <&&> youtubeDlSupported url)
( return (True, True, setDownloader url YoutubeDownloader)
, return (Url.urlExists urlinfo, Url.urlSize urlinfo == fromKey keySize k, url)
)
2014-12-11 20:11:38 +00:00
{- Check that the url exists, and has the same size as the key,
- and add it as an url to the key. -}
addUrlChecked :: AddUrlOptions -> URLString -> RawFilePath -> UUID -> (Key -> Annex (Bool, Bool, URLString)) -> Key -> CommandPerform
addUrlChecked o url file u checkexistssize key =
ifM ((elem url <$> getUrls key) <&&> (elem u <$> loggedLocations key))
( do
showDestinationFile (fromRawFilePath file)
next $ return True
, do
(exists, samesize, url') <- checkexistssize key
if exists && (samesize || relaxedOption (downloadOptions o))
then do
setUrlPresent key url'
logChange key u InfoPresent
next $ return True
else do
warning $ "while adding a new url to an already annexed file, " ++ if exists
then "url does not have expected file size (use --relaxed to bypass this check) " ++ url
else "failed to verify url exists: " ++ url
stop
)
{- Downloads an url (except in fast or relaxed mode) and adds it to the
- repository, normally at the specified FilePath.
- But, if youtube-dl supports the url, it will be written to a
- different file, based on the title of the media. Unless the user
- specified fileOption, which then forces using the FilePath.
2017-11-28 21:17:40 +00:00
-}
addUrlFile :: AddUnlockedMatcher -> DownloadOptions -> URLString -> Url.UrlInfo -> RawFilePath -> Annex (Maybe Key)
addUrlFile addunlockedmatcher o url urlinfo file =
ifM (Annex.getState Annex.fast <||> pure (relaxedOption o))
( nodownloadWeb addunlockedmatcher o url urlinfo file
, downloadWeb addunlockedmatcher o url urlinfo file
)
downloadWeb :: AddUnlockedMatcher -> DownloadOptions -> URLString -> Url.UrlInfo -> RawFilePath -> Annex (Maybe Key)
downloadWeb addunlockedmatcher o url urlinfo file =
go =<< downloadWith' downloader urlkey webUUID url (AssociatedFile (Just file))
2017-11-28 21:17:40 +00:00
where
urlkey = addSizeUrlKey urlinfo $ Backend.URL.fromUrl url Nothing
downloader f p = Url.withUrlOptions $ downloadUrl urlkey p [url] f
2017-11-28 21:17:40 +00:00
go Nothing = return Nothing
go (Just tmp) = ifM (pure (not (rawOption o)) <&&> liftIO (isHtml <$> readFile (fromRawFilePath tmp)))
( tryyoutubedl tmp
2017-11-28 21:17:40 +00:00
, normalfinish tmp
)
normalfinish tmp = checkCanAdd o file $ \canadd -> do
showDestinationFile (fromRawFilePath file)
createWorkTreeDirectory (parentDir file)
Just <$> finishDownloadWith canadd addunlockedmatcher tmp webUUID url file
-- Ask youtube-dl what filename it will download first,
-- so it's only used when the file contains embedded media.
tryyoutubedl tmp = youtubeDlFileNameHtmlOnly url >>= \case
Right mediafile ->
let f = youtubeDlDestFile o file (toRawFilePath mediafile)
in ifAnnexed f
(alreadyannexed (fromRawFilePath f))
(dl f)
Left _ -> normalfinish tmp
where
dl dest = withTmpWorkDir mediakey $ \workdir -> do
let cleanuptmp = pruneTmpWorkDirBefore tmp (liftIO . removeWhenExistsWith R.removeLink)
showNote "using youtube-dl"
Transfer.notifyTransfer Transfer.Download url $
Transfer.download webUUID mediakey (AssociatedFile Nothing) Transfer.noRetry $ \p ->
youtubeDl url (fromRawFilePath workdir) p >>= \case
Right (Just mediafile) -> do
cleanuptmp
checkCanAdd o dest $ \canadd -> do
showDestinationFile (fromRawFilePath dest)
addWorkTree canadd addunlockedmatcher webUUID mediaurl dest mediakey (Just (toRawFilePath mediafile))
return $ Just mediakey
Right Nothing -> normalfinish tmp
Left msg -> do
cleanuptmp
warning msg
return Nothing
mediaurl = setDownloader url YoutubeDownloader
mediakey = Backend.URL.fromUrl mediaurl Nothing
-- Does the already annexed file have the mediaurl
-- as an url? If so nothing to do.
alreadyannexed dest k = do
us <- getUrls k
if mediaurl `elem` us
then return (Just k)
else do
warning $ dest ++ " already exists; not overwriting"
return Nothing
2017-11-28 21:17:40 +00:00
{- The destination file is not known at start time unless the user provided
- a filename. It's not displayed then for output consistency,
- but is added to the json when available. -}
startingAddUrl :: SeekInput -> URLString -> AddUrlOptions -> CommandPerform -> CommandStart
startingAddUrl si url o p = starting "addurl" (ActionItemOther (Just url)) si $ do
case fileOption (downloadOptions o) of
Nothing -> noop
Just file -> maybeShowJSON $ JSONChunk [("file", file)]
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
p
2017-11-28 21:17:40 +00:00
showDestinationFile :: FilePath -> Annex ()
showDestinationFile file = do
showNote ("to " ++ file)
maybeShowJSON $ JSONChunk [("file", file)]
{- The Key should be a dummy key, based on the URL, which is used
- for this download, before we can examine the file and find its real key.
- For resuming downloads to work, the dummy key for a given url should be
2017-11-28 21:17:40 +00:00
- stable. For disk space checking to work, the dummy key should have
- the size of the url already set.
-
- Downloads the url, sets up the worktree file, and returns the
- real key.
-}
downloadWith :: CanAddFile -> AddUnlockedMatcher -> (FilePath -> MeterUpdate -> Annex Bool) -> Key -> UUID -> URLString -> RawFilePath -> Annex (Maybe Key)
downloadWith canadd addunlockedmatcher downloader dummykey u url file =
2017-11-28 21:17:40 +00:00
go =<< downloadWith' downloader dummykey u url afile
where
afile = AssociatedFile (Just file)
2017-11-28 21:17:40 +00:00
go Nothing = return Nothing
go (Just tmp) = Just <$> finishDownloadWith canadd addunlockedmatcher tmp u url file
2017-11-28 21:17:40 +00:00
{- Like downloadWith, but leaves the dummy key content in
- the returned location. -}
downloadWith' :: (FilePath -> MeterUpdate -> Annex Bool) -> Key -> UUID -> URLString -> AssociatedFile -> Annex (Maybe RawFilePath)
2017-11-28 21:17:40 +00:00
downloadWith' downloader dummykey u url afile =
checkDiskSpaceToGet dummykey Nothing $ do
tmp <- fromRepo $ gitAnnexTmpObjectLocation dummykey
ok <- Transfer.notifyTransfer Transfer.Download url $
Transfer.download u dummykey afile Transfer.stdRetry $ \p -> do
createAnnexDirectory (parentDir tmp)
downloader (fromRawFilePath tmp) p
2017-11-28 21:17:40 +00:00
if ok
then return (Just tmp)
else return Nothing
finishDownloadWith :: CanAddFile -> AddUnlockedMatcher -> RawFilePath -> UUID -> URLString -> RawFilePath -> Annex Key
finishDownloadWith canadd addunlockedmatcher tmp u url file = do
2017-11-28 21:17:40 +00:00
backend <- chooseBackend file
let source = KeySource
{ keyFilename = file
, contentLocation = tmp
2017-11-28 21:17:40 +00:00
, inodeCache = Nothing
}
key <- fst <$> genKey source nullMeterUpdate backend
addWorkTree canadd addunlockedmatcher u url file key (Just tmp)
return key
{- Adds the url size to the Key. -}
addSizeUrlKey :: Url.UrlInfo -> Key -> Key
addSizeUrlKey urlinfo key = alterKey key $ \d -> d
{ keySize = Url.urlSize urlinfo
}
{- Adds worktree file to the repository. -}
addWorkTree :: CanAddFile -> AddUnlockedMatcher -> UUID -> URLString -> RawFilePath -> Key -> Maybe RawFilePath -> Annex ()
addWorkTree _ addunlockedmatcher u url file key mtmp = case mtmp of
Nothing -> go
Just tmp -> do
-- Move to final location for large file check.
pruneTmpWorkDirBefore tmp $ \_ -> do
createWorkTreeDirectory (P.takeDirectory file)
liftIO $ renameFile
(fromRawFilePath tmp)
(fromRawFilePath file)
largematcher <- largeFilesMatcher
large <- checkFileMatcher largematcher file
if large
then do
-- Move back to tmp because addAnnexedFile
-- needs the file in a different location
-- than the work tree file.
liftIO $ renameFile
(fromRawFilePath file)
(fromRawFilePath tmp)
go
else void $ Command.Add.addSmall noci file
where
go = do
maybeShowJSON $ JSONChunk [("key", serializeKey key)]
setUrlPresent key url
logChange key u InfoPresent
ifM (addAnnexedFile noci addunlockedmatcher file key mtmp)
annex.securehashesonly Cryptographically secure hashes can be forced to be used in a repository, by setting annex.securehashesonly. This does not prevent the git repository from containing files with insecure hashes, but it does prevent the content of such files from being pulled into .git/annex/objects from another repository. We want to make sure that at no point does git-annex accept content into .git/annex/objects that is hashed with an insecure key. Here's how it was done: * .git/annex/objects/xx/yy/KEY/ is kept frozen, so nothing can be written to it normally * So every place that writes content must call, thawContent or modifyContent. We can audit for these, and be sure we've considered all cases. * The main functions are moveAnnex, and linkToAnnex; these were made to check annex.securehashesonly, and are the main security boundary for annex.securehashesonly. * Most other calls to modifyContent deal with other files in the KEY directory (inode cache etc). The other ones that mess with the content are: - Annex.Direct.toDirectGen, in which content already in the annex directory is moved to the direct mode file, so not relevant. - fix and lock, which don't add new content - Command.ReKey.linkKey, which manually unlocks it to make a copy. * All other calls to thawContent appear safe. Made moveAnnex return a Bool, so checked all callsites and made them deal with a failure in appropriate ways. linkToAnnex simply returns LinkAnnexFailed; all callsites already deal with it failing in appropriate ways. This commit was sponsored by Riku Voipio.
2017-02-27 17:01:32 +00:00
( do
when (isJust mtmp) $
logStatus key InfoPresent
, maybe noop (\tmp -> pruneTmpWorkDirBefore tmp (liftIO . removeWhenExistsWith R.removeLink)) mtmp
annex.securehashesonly Cryptographically secure hashes can be forced to be used in a repository, by setting annex.securehashesonly. This does not prevent the git repository from containing files with insecure hashes, but it does prevent the content of such files from being pulled into .git/annex/objects from another repository. We want to make sure that at no point does git-annex accept content into .git/annex/objects that is hashed with an insecure key. Here's how it was done: * .git/annex/objects/xx/yy/KEY/ is kept frozen, so nothing can be written to it normally * So every place that writes content must call, thawContent or modifyContent. We can audit for these, and be sure we've considered all cases. * The main functions are moveAnnex, and linkToAnnex; these were made to check annex.securehashesonly, and are the main security boundary for annex.securehashesonly. * Most other calls to modifyContent deal with other files in the KEY directory (inode cache etc). The other ones that mess with the content are: - Annex.Direct.toDirectGen, in which content already in the annex directory is moved to the direct mode file, so not relevant. - fix and lock, which don't add new content - Command.ReKey.linkKey, which manually unlocks it to make a copy. * All other calls to thawContent appear safe. Made moveAnnex return a Bool, so checked all callsites and made them deal with a failure in appropriate ways. linkToAnnex simply returns LinkAnnexFailed; all callsites already deal with it failing in appropriate ways. This commit was sponsored by Riku Voipio.
2017-02-27 17:01:32 +00:00
)
-- git does not need to check ignores, because that has already
-- been done, as witnessed by the CannAddFile.
noci = CheckGitIgnore False
2011-07-01 21:15:46 +00:00
nodownloadWeb :: AddUnlockedMatcher -> DownloadOptions -> URLString -> Url.UrlInfo -> RawFilePath -> Annex (Maybe Key)
nodownloadWeb addunlockedmatcher o url urlinfo file
| Url.urlExists urlinfo = if rawOption o
then nomedia
else either (const nomedia) (usemedia . toRawFilePath)
=<< youtubeDlFileName url
| otherwise = do
warning $ "unable to access url: " ++ url
return Nothing
where
nomedia = do
let key = Backend.URL.fromUrl url (Url.urlSize urlinfo)
nodownloadWeb' o addunlockedmatcher url key file
usemedia mediafile = do
let dest = youtubeDlDestFile o file mediafile
let mediaurl = setDownloader url YoutubeDownloader
let mediakey = Backend.URL.fromUrl mediaurl Nothing
nodownloadWeb' o addunlockedmatcher mediaurl mediakey dest
youtubeDlDestFile :: DownloadOptions -> RawFilePath -> RawFilePath -> RawFilePath
youtubeDlDestFile o destfile mediafile
| isJust (fileOption o) = destfile
| otherwise = P.takeFileName mediafile
nodownloadWeb' :: DownloadOptions -> AddUnlockedMatcher -> URLString -> Key -> RawFilePath -> Annex (Maybe Key)
nodownloadWeb' o addunlockedmatcher url key file = checkCanAdd o file $ \canadd -> do
showDestinationFile (fromRawFilePath file)
createWorkTreeDirectory (parentDir file)
addWorkTree canadd addunlockedmatcher webUUID url file key Nothing
return (Just key)
Fix a few bugs involving filenames that are at or near the filesystem's maximum filename length limit. Started with a problem when running addurl on a really long url, because the whole url is munged into the filename. Ended up doing a fairly extensive review for places where filenames could get too large, although it's hard to say I'm not missed any.. Backend.Url had a 128 character limit, which is fine when the limit is 255, but not if it's a lot shorter on some systems. So check the pathconf() limit. Note that this could result in fromUrl creating different keys for the same url, if run on systems with different limits. I don't see this is likely to cause any problems. That can already happen when using addurl --fast, or if the content of an url changes. Both Command.AddUrl and Backend.Url assumed that urls don't contain a lot of multi-byte unicode, and would fail to truncate an url that did properly. A few places use a filename as the template to make a temp file. While that's nice in that the temp file name can be easily related back to the original filename, it could lead to `git annex add` failing to add a filename that was at or close to the maximum length. Note that in Command.Add.lockdown, the template is still derived from the filename, just with enough space left to turn it into a temp file. This is an important optimisation, because the assistant may lock down a bunch of files all at once, and using the same template for all of them would cause openTempFile to iterate through the same set of names, looking for an unused temp file. I'm not very happy with the relatedTemplate hack, but it avoids that slowdown. Backend.WORM does not limit the filename stored in the key. I have not tried to change that; so git annex add will fail on really long filenames when using the WORM backend. It seems better to preserve the invariant that a WORM key always contains the complete filename, since the filename is the only unique material in the key, other than mtime and size. Since nobody has complained about add failing (I think I saw it once?) on WORM, probably it's ok, or nobody but me uses it. There may be compatability problems if using git annex addurl --fast or the WORM backend on a system with the 255 limit and then trying to use that repo in a system with a smaller limit. I have not tried to deal with those. This commit was sponsored by Alexander Brem. Thanks!
2013-07-30 21:49:11 +00:00
url2file :: URI -> Maybe Int -> Int -> FilePath
url2file url pathdepth pathmax = case pathdepth of
2013-10-05 17:32:42 +00:00
Nothing -> truncateFilePath pathmax $ sanitizeFilePath fullurl
2012-02-16 16:25:19 +00:00
Just depth
| depth >= length urlbits -> frombits id
| depth > 0 -> frombits $ drop depth
2012-02-16 18:28:17 +00:00
| depth < 0 -> frombits $ reverse . take (negate depth) . reverse
| otherwise -> giveup "bad --pathdepth"
2012-11-12 05:05:04 +00:00
where
fullurl = concat
[ maybe "" uriRegName (uriAuthority url)
, uriPath url
, uriQuery url
]
frombits a = intercalate "/" $ a urlbits
urlbits = map (truncateFilePath pathmax . sanitizeFilePath) $
filter (not . null) $ splitc '/' fullurl
urlString2file :: URLString -> Maybe Int -> Int -> FilePath
urlString2file s pathdepth pathmax = case Url.parseURIRelaxed s of
Nothing -> giveup $ "bad uri " ++ s
Just u -> url2file u pathdepth pathmax
adjustFile :: AddUrlOptions -> FilePath -> FilePath
adjustFile o = addprefix . addsuffix
where
addprefix f = maybe f (++ f) (prefixOption o)
addsuffix f = maybe f (f ++) (suffixOption o)
data CanAddFile = CanAddFile
checkCanAdd :: DownloadOptions -> RawFilePath -> (CanAddFile -> Annex (Maybe a)) -> Annex (Maybe a)
checkCanAdd o file a = ifM (isJust <$> (liftIO $ catchMaybeIO $ R.getSymbolicLinkStatus file))
( do
warning $ fromRawFilePath file ++ " already exists; not overwriting"
return Nothing
, ifM (checkIgnored (checkGitIgnoreOption o) file)
( do
warning $ "not adding " ++ fromRawFilePath file ++ " which is .gitignored (use --no-check-gitignore to override)"
return Nothing
, a CanAddFile
)
)