2012-03-09 23:08:10 +00:00
|
|
|
{- GHC File system encoding handling.
|
2012-03-09 22:52:03 +00:00
|
|
|
-
|
2021-08-11 00:45:02 +00:00
|
|
|
- Copyright 2012-2021 Joey Hess <id@joeyh.name>
|
2012-03-09 22:52:03 +00:00
|
|
|
-
|
2014-05-10 14:01:27 +00:00
|
|
|
- License: BSD-2-clause
|
2012-03-09 22:52:03 +00:00
|
|
|
-}
|
|
|
|
|
2014-03-19 18:49:01 +00:00
|
|
|
{-# LANGUAGE CPP #-}
|
2015-05-10 20:31:50 +00:00
|
|
|
{-# OPTIONS_GHC -fno-warn-tabs #-}
|
2014-03-19 18:49:01 +00:00
|
|
|
|
2012-09-13 23:14:00 +00:00
|
|
|
module Utility.FileSystemEncoding (
|
2016-12-24 18:46:31 +00:00
|
|
|
useFileSystemEncoding,
|
2016-12-30 22:14:19 +00:00
|
|
|
fileEncoding,
|
2018-05-08 18:26:57 +00:00
|
|
|
RawFilePath,
|
|
|
|
fromRawFilePath,
|
|
|
|
toRawFilePath,
|
2019-01-01 18:54:06 +00:00
|
|
|
decodeBL,
|
|
|
|
encodeBL,
|
2019-01-02 02:44:04 +00:00
|
|
|
decodeBS,
|
Fix setting/setting/viewing metadata that contains unicode or other special characters, when in a non-unicode locale.
Oh boy, not again. So, another place that the filesystem encoding needs to
be applied. Yay.
In passing, I changed decodeBS so if a NUL is embedded in the input, the
resulting FilePath doesn't get truncated at that NUL. This was needed to
make prop_b64_roundtrips pass, and on reviewing the callers of decodeBS, I
didn't see any where this wouldn't make sense. When a FilePath is used to
operate on the filesystem, it'll get truncated at a NUL anyway, whereas if
a String is being used for something else, it might conceivably have a NUL
in it, and we wouldn't want it to get truncated when going through
decodeBS.
(NB: There may be a speed impact from this change.)
2015-08-11 22:40:59 +00:00
|
|
|
encodeBS,
|
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
|
|
|
truncateFilePath,
|
2012-09-13 23:14:00 +00:00
|
|
|
) where
|
2012-03-09 22:52:03 +00:00
|
|
|
|
fix key directory hash calculation code
Fix Key directory hash calculation code to behave as it did before version
3.20120227 when a key contains non-ascii.
The hash directories for a given Key are based on its md5sum.
Prior to ghc 7.4, Keys contained raw, undecoded bytes, so the md5sum was
taken of each byte in turn. With the ghc 7.4 filename encoding change,
keys contains decoded unicode characters (possibly with surrigates for
undecodable bytes). This changes the result of the md5sum, since the md5sum
used is pure haskell and supports unicode. And that won't do, as git-annex
will start looking in a different hash directory for the content of a key.
The surrigates are particularly bad, since that's essentially a ghc
implementation detail, so could change again at any time. Also, changing
the locale changes how the bytes are decoded, which can also change
the md5sum.
Symptoms would include things like:
* git annex fsck would complain that no copies existed of a file,
despite its symlink pointing to the content that was locally present
* git annex fix would change the symlink to use the wrong hash
directory.
Only WORM backend is likely to have been affected, since only it tends
to include much filename data (SHA1E could in theory also be affected).
I have not tried to support the hash directories used by git-annex versions
3.20120227 to 3.20120308, so things added with those versions with WORM
will require manual fixups. Sorry for the inconvenience!
2012-03-09 23:26:02 +00:00
|
|
|
import qualified GHC.IO.Encoding as Encoding
|
|
|
|
import System.IO
|
2020-11-24 16:35:09 +00:00
|
|
|
import System.FilePath.ByteString (RawFilePath, encodeFilePath, decodeFilePath)
|
2018-05-08 18:26:57 +00:00
|
|
|
import qualified Data.ByteString as S
|
2014-03-19 18:49:01 +00:00
|
|
|
import qualified Data.ByteString.Lazy as L
|
|
|
|
#ifdef mingw32_HOST_OS
|
2019-01-01 18:54:06 +00:00
|
|
|
import qualified Data.ByteString.UTF8 as S8
|
2014-03-19 18:49:01 +00:00
|
|
|
import qualified Data.ByteString.Lazy.UTF8 as L8
|
2024-03-26 17:06:52 +00:00
|
|
|
#else
|
|
|
|
import qualified GHC.Foreign as GHC
|
|
|
|
import System.IO.Unsafe
|
|
|
|
import Data.ByteString.Unsafe (unsafePackMallocCStringLen)
|
2014-03-19 18:49:01 +00:00
|
|
|
#endif
|
2012-03-09 23:08:10 +00:00
|
|
|
|
2016-12-24 18:46:31 +00:00
|
|
|
{- Makes all subsequent Handles that are opened, as well as stdio Handles,
|
|
|
|
- use the filesystem encoding, instead of the encoding of the current
|
|
|
|
- locale.
|
|
|
|
-
|
|
|
|
- The filesystem encoding allows "arbitrary undecodable bytes to be
|
|
|
|
- round-tripped through it". This avoids encoded failures when data is not
|
|
|
|
- encoded matching the current locale.
|
|
|
|
-
|
|
|
|
- Note that code can still use hSetEncoding to change the encoding of a
|
|
|
|
- Handle. This only affects the default encoding.
|
2014-03-19 18:49:01 +00:00
|
|
|
-}
|
2016-12-24 18:46:31 +00:00
|
|
|
useFileSystemEncoding :: IO ()
|
|
|
|
useFileSystemEncoding = do
|
2014-03-19 18:49:01 +00:00
|
|
|
#ifndef mingw32_HOST_OS
|
2016-12-24 18:46:31 +00:00
|
|
|
e <- Encoding.getFileSystemEncoding
|
2014-03-19 18:49:01 +00:00
|
|
|
#else
|
2016-12-24 18:46:31 +00:00
|
|
|
{- The file system encoding does not work well on Windows,
|
|
|
|
- and Windows only has utf FilePaths anyway. -}
|
|
|
|
let e = Encoding.utf8
|
2014-03-19 18:49:01 +00:00
|
|
|
#endif
|
2016-12-24 18:46:31 +00:00
|
|
|
hSetEncoding stdin e
|
|
|
|
hSetEncoding stdout e
|
|
|
|
hSetEncoding stderr e
|
|
|
|
Encoding.setLocaleEncoding e
|
2012-03-09 22:52:03 +00:00
|
|
|
|
2016-12-30 22:14:19 +00:00
|
|
|
fileEncoding :: Handle -> IO ()
|
|
|
|
#ifndef mingw32_HOST_OS
|
|
|
|
fileEncoding h = hSetEncoding h =<< Encoding.getFileSystemEncoding
|
|
|
|
#else
|
|
|
|
fileEncoding h = hSetEncoding h Encoding.utf8
|
|
|
|
#endif
|
|
|
|
|
2014-03-19 18:49:01 +00:00
|
|
|
{- Decodes a ByteString into a FilePath, applying the filesystem encoding. -}
|
2019-01-01 18:54:06 +00:00
|
|
|
decodeBL :: L.ByteString -> FilePath
|
2014-03-19 18:49:01 +00:00
|
|
|
#ifndef mingw32_HOST_OS
|
2021-08-11 00:45:02 +00:00
|
|
|
decodeBL = decodeBS . L.toStrict
|
2014-03-19 18:49:01 +00:00
|
|
|
#else
|
|
|
|
{- On Windows, we assume that the ByteString is utf-8, since Windows
|
|
|
|
- only uses unicode for filenames. -}
|
2019-01-01 18:54:06 +00:00
|
|
|
decodeBL = L8.toString
|
|
|
|
#endif
|
|
|
|
|
Fix setting/setting/viewing metadata that contains unicode or other special characters, when in a non-unicode locale.
Oh boy, not again. So, another place that the filesystem encoding needs to
be applied. Yay.
In passing, I changed decodeBS so if a NUL is embedded in the input, the
resulting FilePath doesn't get truncated at that NUL. This was needed to
make prop_b64_roundtrips pass, and on reviewing the callers of decodeBS, I
didn't see any where this wouldn't make sense. When a FilePath is used to
operate on the filesystem, it'll get truncated at a NUL anyway, whereas if
a String is being used for something else, it might conceivably have a NUL
in it, and we wouldn't want it to get truncated when going through
decodeBS.
(NB: There may be a speed impact from this change.)
2015-08-11 22:40:59 +00:00
|
|
|
{- Encodes a FilePath into a ByteString, applying the filesystem encoding. -}
|
2019-01-01 18:54:06 +00:00
|
|
|
encodeBL :: FilePath -> L.ByteString
|
|
|
|
#ifndef mingw32_HOST_OS
|
2021-08-11 00:45:02 +00:00
|
|
|
encodeBL = L.fromStrict . encodeBS
|
2019-01-01 18:54:06 +00:00
|
|
|
#else
|
|
|
|
encodeBL = L8.fromString
|
|
|
|
#endif
|
|
|
|
|
2019-01-02 02:44:04 +00:00
|
|
|
decodeBS :: S.ByteString -> FilePath
|
|
|
|
#ifndef mingw32_HOST_OS
|
2021-12-08 22:59:22 +00:00
|
|
|
-- This does the same thing as System.FilePath.ByteString.decodeFilePath,
|
|
|
|
-- with an identical implementation. However, older versions of that library
|
|
|
|
-- truncated at NUL, which this must not do, because it may end up used on
|
|
|
|
-- something other than a unix filepath.
|
2021-08-11 00:45:02 +00:00
|
|
|
{-# NOINLINE decodeBS #-}
|
|
|
|
decodeBS b = unsafePerformIO $ do
|
|
|
|
enc <- Encoding.getFileSystemEncoding
|
|
|
|
S.useAsCStringLen b (GHC.peekCStringLen enc)
|
2019-01-02 02:44:04 +00:00
|
|
|
#else
|
|
|
|
decodeBS = S8.toString
|
|
|
|
#endif
|
|
|
|
|
2019-01-01 18:54:06 +00:00
|
|
|
encodeBS :: FilePath -> S.ByteString
|
Fix setting/setting/viewing metadata that contains unicode or other special characters, when in a non-unicode locale.
Oh boy, not again. So, another place that the filesystem encoding needs to
be applied. Yay.
In passing, I changed decodeBS so if a NUL is embedded in the input, the
resulting FilePath doesn't get truncated at that NUL. This was needed to
make prop_b64_roundtrips pass, and on reviewing the callers of decodeBS, I
didn't see any where this wouldn't make sense. When a FilePath is used to
operate on the filesystem, it'll get truncated at a NUL anyway, whereas if
a String is being used for something else, it might conceivably have a NUL
in it, and we wouldn't want it to get truncated when going through
decodeBS.
(NB: There may be a speed impact from this change.)
2015-08-11 22:40:59 +00:00
|
|
|
#ifndef mingw32_HOST_OS
|
2021-12-08 22:59:22 +00:00
|
|
|
-- This does the same thing as System.FilePath.ByteString.encodeFilePath,
|
|
|
|
-- with an identical implementation. However, older versions of that library
|
|
|
|
-- truncated at NUL, which this must not do, because it may end up used on
|
|
|
|
-- something other than a unix filepath.
|
2021-08-11 00:45:02 +00:00
|
|
|
{-# NOINLINE encodeBS #-}
|
|
|
|
encodeBS f = unsafePerformIO $ do
|
|
|
|
enc <- Encoding.getFileSystemEncoding
|
|
|
|
GHC.newCStringLen enc f >>= unsafePackMallocCStringLen
|
Fix setting/setting/viewing metadata that contains unicode or other special characters, when in a non-unicode locale.
Oh boy, not again. So, another place that the filesystem encoding needs to
be applied. Yay.
In passing, I changed decodeBS so if a NUL is embedded in the input, the
resulting FilePath doesn't get truncated at that NUL. This was needed to
make prop_b64_roundtrips pass, and on reviewing the callers of decodeBS, I
didn't see any where this wouldn't make sense. When a FilePath is used to
operate on the filesystem, it'll get truncated at a NUL anyway, whereas if
a String is being used for something else, it might conceivably have a NUL
in it, and we wouldn't want it to get truncated when going through
decodeBS.
(NB: There may be a speed impact from this change.)
2015-08-11 22:40:59 +00:00
|
|
|
#else
|
2019-01-01 18:54:06 +00:00
|
|
|
encodeBS = S8.fromString
|
Fix setting/setting/viewing metadata that contains unicode or other special characters, when in a non-unicode locale.
Oh boy, not again. So, another place that the filesystem encoding needs to
be applied. Yay.
In passing, I changed decodeBS so if a NUL is embedded in the input, the
resulting FilePath doesn't get truncated at that NUL. This was needed to
make prop_b64_roundtrips pass, and on reviewing the callers of decodeBS, I
didn't see any where this wouldn't make sense. When a FilePath is used to
operate on the filesystem, it'll get truncated at a NUL anyway, whereas if
a String is being used for something else, it might conceivably have a NUL
in it, and we wouldn't want it to get truncated when going through
decodeBS.
(NB: There may be a speed impact from this change.)
2015-08-11 22:40:59 +00:00
|
|
|
#endif
|
|
|
|
|
2018-05-08 18:26:57 +00:00
|
|
|
fromRawFilePath :: RawFilePath -> FilePath
|
2020-01-05 00:18:40 +00:00
|
|
|
fromRawFilePath = decodeFilePath
|
2018-05-08 18:26:57 +00:00
|
|
|
|
|
|
|
toRawFilePath :: FilePath -> RawFilePath
|
2020-01-05 00:18:40 +00:00
|
|
|
toRawFilePath = encodeFilePath
|
2018-05-08 18:26:57 +00:00
|
|
|
|
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
|
|
|
{- Truncates a FilePath to the given number of bytes (or less),
|
|
|
|
- as represented on disk.
|
|
|
|
-
|
|
|
|
- Avoids returning an invalid part of a unicode byte sequence, at the
|
|
|
|
- cost of efficiency when running on a large FilePath.
|
|
|
|
-}
|
|
|
|
truncateFilePath :: Int -> FilePath -> FilePath
|
2014-03-19 18:49:01 +00:00
|
|
|
#ifndef 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
|
|
|
truncateFilePath n = go . reverse
|
|
|
|
where
|
2014-10-09 18:53:13 +00:00
|
|
|
go f =
|
2021-08-11 00:45:02 +00:00
|
|
|
let b = encodeBS f
|
|
|
|
in if S.length b <= n
|
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
|
|
|
then reverse f
|
|
|
|
else go (drop 1 f)
|
2014-03-19 18:49:01 +00:00
|
|
|
#else
|
|
|
|
{- On Windows, count the number of bytes used by each utf8 character. -}
|
|
|
|
truncateFilePath n = reverse . go [] n . L8.fromString
|
|
|
|
where
|
|
|
|
go coll cnt bs
|
|
|
|
| cnt <= 0 = coll
|
|
|
|
| otherwise = case L8.decode bs of
|
|
|
|
Just (c, x) | c /= L8.replacement_char ->
|
|
|
|
let x' = fromIntegral x
|
|
|
|
in if cnt - x' < 0
|
|
|
|
then coll
|
|
|
|
else go (c:coll) (cnt - x') (L8.drop 1 bs)
|
|
|
|
_ -> coll
|
|
|
|
#endif
|