2011-08-22 20:14:12 +00:00
|
|
|
{- path manipulation
|
|
|
|
-
|
2013-05-12 18:58:46 +00:00
|
|
|
- Copyright 2010-2013 Joey Hess <joey@kitenet.net>
|
2011-08-22 20:14:12 +00:00
|
|
|
-
|
|
|
|
- Licensed under the GNU GPL version 3 or higher.
|
|
|
|
-}
|
2013-04-14 16:51:05 +00:00
|
|
|
|
2013-05-12 21:36:44 +00:00
|
|
|
{-# LANGUAGE PackageImports, CPP #-}
|
2011-08-22 20:14:12 +00:00
|
|
|
|
|
|
|
module Utility.Path where
|
|
|
|
|
|
|
|
import Data.String.Utils
|
|
|
|
import System.FilePath
|
|
|
|
import System.Directory
|
|
|
|
import Data.List
|
|
|
|
import Data.Maybe
|
2013-08-22 22:25:21 +00:00
|
|
|
import Data.Char
|
2011-08-25 04:28:55 +00:00
|
|
|
import Control.Applicative
|
2011-10-16 04:31:25 +00:00
|
|
|
|
2013-08-02 16:27:32 +00:00
|
|
|
#ifdef mingw32_HOST_OS
|
2013-05-14 17:24:15 +00:00
|
|
|
import Data.Char
|
2013-05-14 19:21:35 +00:00
|
|
|
import qualified System.FilePath.Posix as Posix
|
|
|
|
#else
|
|
|
|
import qualified "MissingH" System.Path as MissingH
|
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
|
|
|
import System.Posix.Files
|
2013-05-14 17:24:15 +00:00
|
|
|
#endif
|
|
|
|
|
2011-10-16 04:31:25 +00:00
|
|
|
import Utility.Monad
|
2012-10-25 22:17:32 +00:00
|
|
|
import Utility.UserInfo
|
2011-08-22 20:14:12 +00:00
|
|
|
|
2013-05-12 21:36:44 +00:00
|
|
|
{- Makes a path absolute if it's not already.
|
|
|
|
- The first parameter is a base directory (ie, the cwd) to use if the path
|
|
|
|
- is not already absolute.
|
|
|
|
-
|
|
|
|
- On Unix, collapses and normalizes ".." etc in the path. May return Nothing
|
|
|
|
- if the path cannot be normalized.
|
|
|
|
-
|
|
|
|
- MissingH's absNormPath does not work on Windows, so on Windows
|
|
|
|
- no normalization is done.
|
|
|
|
-}
|
|
|
|
absNormPath :: FilePath -> FilePath -> Maybe FilePath
|
2013-08-02 16:27:32 +00:00
|
|
|
#ifndef mingw32_HOST_OS
|
2013-05-12 21:36:44 +00:00
|
|
|
absNormPath dir path = MissingH.absNormPath dir path
|
|
|
|
#else
|
|
|
|
absNormPath dir path = Just $ combine dir path
|
|
|
|
#endif
|
|
|
|
|
2013-05-12 18:58:46 +00:00
|
|
|
{- Returns the parent directory of a path.
|
|
|
|
-
|
|
|
|
- To allow this to be easily used in loops, which terminate upon reaching the
|
|
|
|
- top, the parent of / is "" -}
|
2011-08-22 20:14:12 +00:00
|
|
|
parentDir :: FilePath -> FilePath
|
2011-10-11 18:43:45 +00:00
|
|
|
parentDir dir
|
2013-05-12 18:58:46 +00:00
|
|
|
| null dirs = ""
|
|
|
|
| otherwise = joinDrive drive (join s $ init dirs)
|
2012-12-13 04:24:19 +00:00
|
|
|
where
|
2013-05-12 18:58:46 +00:00
|
|
|
-- on Unix, the drive will be "/" when the dir is absolute, otherwise ""
|
|
|
|
(drive, path) = splitDrive dir
|
|
|
|
dirs = filter (not . null) $ split s path
|
2012-12-13 04:24:19 +00:00
|
|
|
s = [pathSeparator]
|
2011-08-22 20:14:12 +00:00
|
|
|
|
|
|
|
prop_parentDir_basics :: FilePath -> Bool
|
|
|
|
prop_parentDir_basics dir
|
|
|
|
| null dir = True
|
|
|
|
| dir == "/" = parentDir dir == ""
|
|
|
|
| otherwise = p /= dir
|
2012-12-13 04:24:19 +00:00
|
|
|
where
|
|
|
|
p = parentDir dir
|
2011-08-22 20:14:12 +00:00
|
|
|
|
|
|
|
{- Checks if the first FilePath is, or could be said to contain the second.
|
|
|
|
- For example, "foo/" contains "foo/bar". Also, "foo", "./foo", "foo/" etc
|
|
|
|
- are all equivilant.
|
|
|
|
-}
|
|
|
|
dirContains :: FilePath -> FilePath -> Bool
|
2013-05-12 18:58:46 +00:00
|
|
|
dirContains a b = a == b || a' == b' || (a'++[pathSeparator]) `isPrefixOf` b'
|
2012-12-13 04:24:19 +00:00
|
|
|
where
|
|
|
|
norm p = fromMaybe "" $ absNormPath p "."
|
|
|
|
a' = norm a
|
|
|
|
b' = norm b
|
2011-08-22 20:14:12 +00:00
|
|
|
|
2012-01-23 20:57:45 +00:00
|
|
|
{- Converts a filename into a normalized, absolute path.
|
|
|
|
-
|
|
|
|
- Unlike Directory.canonicalizePath, this does not require the path
|
|
|
|
- already exists. -}
|
2011-08-22 20:14:12 +00:00
|
|
|
absPath :: FilePath -> IO FilePath
|
|
|
|
absPath file = do
|
|
|
|
cwd <- getCurrentDirectory
|
|
|
|
return $ absPathFrom cwd file
|
|
|
|
|
|
|
|
{- Converts a filename into a normalized, absolute path
|
|
|
|
- from the specified cwd. -}
|
|
|
|
absPathFrom :: FilePath -> FilePath -> FilePath
|
|
|
|
absPathFrom cwd file = fromMaybe bad $ absNormPath cwd file
|
2012-12-13 04:24:19 +00:00
|
|
|
where
|
|
|
|
bad = error $ "unable to normalize " ++ file
|
2011-08-22 20:14:12 +00:00
|
|
|
|
|
|
|
{- Constructs a relative path from the CWD to a file.
|
|
|
|
-
|
|
|
|
- For example, assuming CWD is /tmp/foo/bar:
|
|
|
|
- relPathCwdToFile "/tmp/foo" == ".."
|
|
|
|
- relPathCwdToFile "/tmp/foo/bar" == ""
|
|
|
|
-}
|
|
|
|
relPathCwdToFile :: FilePath -> IO FilePath
|
2011-08-25 04:28:55 +00:00
|
|
|
relPathCwdToFile f = relPathDirToFile <$> getCurrentDirectory <*> absPath f
|
2011-08-22 20:14:12 +00:00
|
|
|
|
|
|
|
{- Constructs a relative path from a directory to a file.
|
|
|
|
-
|
|
|
|
- Both must be absolute, and normalized (eg with absNormpath).
|
|
|
|
-}
|
|
|
|
relPathDirToFile :: FilePath -> FilePath -> FilePath
|
2011-10-11 18:43:45 +00:00
|
|
|
relPathDirToFile from to = join s $ dotdots ++ uncommon
|
2012-12-13 04:24:19 +00:00
|
|
|
where
|
|
|
|
s = [pathSeparator]
|
|
|
|
pfrom = split s from
|
|
|
|
pto = split s to
|
|
|
|
common = map fst $ takeWhile same $ zip pfrom pto
|
|
|
|
same (c,d) = c == d
|
|
|
|
uncommon = drop numcommon pto
|
|
|
|
dotdots = replicate (length pfrom - numcommon) ".."
|
|
|
|
numcommon = length common
|
2011-08-22 20:14:12 +00:00
|
|
|
|
|
|
|
prop_relPathDirToFile_basics :: FilePath -> FilePath -> Bool
|
|
|
|
prop_relPathDirToFile_basics from to
|
|
|
|
| from == to = null r
|
|
|
|
| otherwise = not (null r)
|
2012-12-13 04:24:19 +00:00
|
|
|
where
|
|
|
|
r = relPathDirToFile from to
|
2011-09-19 05:37:04 +00:00
|
|
|
|
2012-03-05 16:42:52 +00:00
|
|
|
prop_relPathDirToFile_regressionTest :: Bool
|
|
|
|
prop_relPathDirToFile_regressionTest = same_dir_shortcurcuits_at_difference
|
2012-12-13 04:24:19 +00:00
|
|
|
where
|
|
|
|
{- Two paths have the same directory component at the same
|
|
|
|
- location, but it's not really the same directory.
|
|
|
|
- Code used to get this wrong. -}
|
|
|
|
same_dir_shortcurcuits_at_difference =
|
2013-06-18 17:08:28 +00:00
|
|
|
relPathDirToFile (joinPath [pathSeparator : "tmp", "r", "lll", "xxx", "yyy", "18"])
|
|
|
|
(joinPath [pathSeparator : "tmp", "r", ".git", "annex", "objects", "18", "gk", "SHA256-foo", "SHA256-foo"])
|
|
|
|
== joinPath ["..", "..", "..", "..", ".git", "annex", "objects", "18", "gk", "SHA256-foo", "SHA256-foo"]
|
2012-03-05 16:42:52 +00:00
|
|
|
|
2012-11-25 21:54:08 +00:00
|
|
|
{- Given an original list of paths, and an expanded list derived from it,
|
|
|
|
- generates a list of lists, where each sublist corresponds to one of the
|
2013-05-12 18:58:46 +00:00
|
|
|
- original paths. When the original path is a directory, any items
|
2012-11-25 21:54:08 +00:00
|
|
|
- in the expanded list that are contained in that directory will appear in
|
|
|
|
- its segment.
|
2011-09-19 05:37:04 +00:00
|
|
|
-}
|
2012-11-25 21:54:08 +00:00
|
|
|
segmentPaths :: [FilePath] -> [FilePath] -> [[FilePath]]
|
|
|
|
segmentPaths [] new = [new]
|
|
|
|
segmentPaths [_] new = [new] -- optimisation
|
|
|
|
segmentPaths (l:ls) new = [found] ++ segmentPaths ls rest
|
2012-12-13 04:24:19 +00:00
|
|
|
where
|
|
|
|
(found, rest)=partition (l `dirContains`) new
|
2011-09-19 05:37:04 +00:00
|
|
|
|
2012-11-25 21:54:08 +00:00
|
|
|
{- This assumes that it's cheaper to call segmentPaths on the result,
|
|
|
|
- than it would be to run the action separately with each path. In
|
|
|
|
- the case of git file list commands, that assumption tends to hold.
|
2011-09-19 05:37:04 +00:00
|
|
|
-}
|
2012-11-25 21:54:08 +00:00
|
|
|
runSegmentPaths :: ([FilePath] -> IO [FilePath]) -> [FilePath] -> IO [[FilePath]]
|
|
|
|
runSegmentPaths a paths = segmentPaths paths <$> a paths
|
2011-10-16 04:31:25 +00:00
|
|
|
|
2012-08-02 11:47:50 +00:00
|
|
|
{- Converts paths in the home directory to use ~/ -}
|
|
|
|
relHome :: FilePath -> IO String
|
|
|
|
relHome path = do
|
|
|
|
home <- myHomeDir
|
|
|
|
return $ if dirContains home path
|
|
|
|
then "~/" ++ relPathDirToFile home path
|
|
|
|
else path
|
|
|
|
|
2012-12-14 19:52:44 +00:00
|
|
|
{- Checks if a command is available in PATH.
|
|
|
|
-
|
|
|
|
- The command may be fully-qualified, in which case, this succeeds as
|
|
|
|
- long as it exists. -}
|
2011-10-16 04:31:25 +00:00
|
|
|
inPath :: String -> IO Bool
|
2012-12-14 19:52:44 +00:00
|
|
|
inPath command = isJust <$> searchPath command
|
|
|
|
|
|
|
|
{- Finds a command in PATH and returns the full path to it.
|
|
|
|
-
|
|
|
|
- The command may be fully qualified already, in which case it will
|
|
|
|
- be returned if it exists.
|
|
|
|
-}
|
|
|
|
searchPath :: String -> IO (Maybe FilePath)
|
|
|
|
searchPath command
|
|
|
|
| isAbsolute command = check command
|
|
|
|
| otherwise = getSearchPath >>= getM indir
|
2012-12-13 04:24:19 +00:00
|
|
|
where
|
2012-12-14 19:52:44 +00:00
|
|
|
indir d = check $ d </> command
|
2013-07-06 04:48:47 +00:00
|
|
|
check f = firstM doesFileExist
|
2013-08-02 16:27:32 +00:00
|
|
|
#ifdef mingw32_HOST_OS
|
2013-07-06 04:48:47 +00:00
|
|
|
[f, f ++ ".exe"]
|
|
|
|
#else
|
|
|
|
[f]
|
|
|
|
#endif
|
2012-01-03 04:09:09 +00:00
|
|
|
|
|
|
|
{- Checks if a filename is a unix dotfile. All files inside dotdirs
|
|
|
|
- count as dotfiles. -}
|
|
|
|
dotfile :: FilePath -> Bool
|
|
|
|
dotfile file
|
|
|
|
| f == "." = False
|
|
|
|
| f == ".." = False
|
|
|
|
| f == "" = False
|
|
|
|
| otherwise = "." `isPrefixOf` f || dotfile (takeDirectory file)
|
2012-12-13 04:24:19 +00:00
|
|
|
where
|
|
|
|
f = takeFileName file
|
2013-05-14 17:24:15 +00:00
|
|
|
|
|
|
|
{- Converts a DOS style path to a Cygwin style path. Only on Windows.
|
|
|
|
- Any trailing '\' is preserved as a trailing '/' -}
|
|
|
|
toCygPath :: FilePath -> FilePath
|
2013-08-02 16:27:32 +00:00
|
|
|
#ifndef mingw32_HOST_OS
|
2013-05-14 17:24:15 +00:00
|
|
|
toCygPath = id
|
|
|
|
#else
|
|
|
|
toCygPath p
|
|
|
|
| null drive = recombine parts
|
|
|
|
| otherwise = recombine $ "/cygdrive" : driveletter drive : parts
|
|
|
|
where
|
|
|
|
(drive, p') = splitDrive p
|
|
|
|
parts = splitDirectories p'
|
|
|
|
driveletter = map toLower . takeWhile (/= ':')
|
|
|
|
recombine = fixtrailing . Posix.joinPath
|
|
|
|
fixtrailing s
|
|
|
|
| hasTrailingPathSeparator p = Posix.addTrailingPathSeparator s
|
|
|
|
| otherwise = s
|
|
|
|
#endif
|
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
|
|
|
|
|
|
|
{- Maximum size to use for a file in a specified directory.
|
|
|
|
-
|
|
|
|
- Many systems have a 255 byte limit to the name of a file,
|
|
|
|
- so that's taken as the max if the system has a larger limit, or has no
|
|
|
|
- limit.
|
|
|
|
-}
|
|
|
|
fileNameLengthLimit :: FilePath -> IO Int
|
2013-08-02 16:27:32 +00:00
|
|
|
#ifdef mingw32_HOST_OS
|
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
|
|
|
fileNameLengthLimit _ = return 255
|
|
|
|
#else
|
|
|
|
fileNameLengthLimit dir = do
|
|
|
|
l <- fromIntegral <$> getPathVar dir FileNameLimit
|
|
|
|
if l <= 0
|
|
|
|
then return 255
|
|
|
|
else return $ minimum [l, 255]
|
|
|
|
where
|
|
|
|
#endif
|
2013-08-22 22:25:21 +00:00
|
|
|
|
|
|
|
{- Given a string that we'd like to use as the basis for FilePath, but that
|
|
|
|
- was provided by a third party and is not to be trusted, returns the closest
|
|
|
|
- sane FilePath.
|
|
|
|
-
|
|
|
|
- All spaces and punctuation are replaced with '_', except for '.'
|
|
|
|
- "../" will thus turn into ".._", which is safe.
|
|
|
|
-}
|
|
|
|
sanitizeFilePath :: String -> FilePath
|
|
|
|
sanitizeFilePath = map sanitize
|
|
|
|
where
|
|
|
|
sanitize c
|
|
|
|
| c == '.' = c
|
|
|
|
| isSpace c || isPunctuation c || c == '/' = '_'
|
|
|
|
| otherwise = c
|