The initial attempt was to discard cache when appropriate, but it appears
to be error prone. We had to carefully inspect all places where audit() is
called e.g. without actually updating filesystem, before removing files and
directories, etc.
So, this patch disables the cache of audited paths by default, and enables
it only for the following cases:
- short-lived auditor objects
- repo.vfs, repo.svfs, and repo.cachevfs, which are managed directories
and considered sort of append-only (a file/directory would never be
replaced with a symlink)
There would be more cacheable vfs objects (e.g. mq.queue.opener), but I
decided not to inspect all of them in this patch. We can make them cached
later.
Benchmark result:
- using old clone of http://selenic.com/repo/linux-2.6/ (38319 files)
- on tmpfs
- run HGRCPATH=/dev/null hg up -q --time tip && hg up -q null
- try 4 times and take the last three results
original:
real 7.480 secs (user 1.140+22.760 sys 0.150+1.690)
real 8.010 secs (user 1.070+22.280 sys 0.170+2.120)
real 7.470 secs (user 1.120+22.390 sys 0.120+1.910)
clearcache (the other series):
real 7.680 secs (user 1.120+23.420 sys 0.140+1.970)
real 7.670 secs (user 1.110+23.620 sys 0.130+1.810)
real 7.740 secs (user 1.090+23.510 sys 0.160+1.940)
enable cache only for vfs and svfs (this series):
real 8.730 secs (user 1.500+25.190 sys 0.260+2.260)
real 8.750 secs (user 1.490+25.170 sys 0.250+2.340)
real 9.010 secs (user 1.680+25.340 sys 0.280+2.540)
remove cache function at all (for reference):
real 9.620 secs (user 1.440+27.120 sys 0.250+2.980)
real 9.420 secs (user 1.400+26.940 sys 0.320+3.130)
real 9.760 secs (user 1.530+27.270 sys 0.250+2.970)
We want to be able to do more precise check when auditing a path depending of
the intend of the file access (eg read versus write). So we now pass the 'mode'
value to 'audit' and update the audit function to accept them.
This will be put to use in the next changeset.
I always read the name "checkcase(path)" as "do we need to check for
case folding at this path", but it's actually (I think) meant to be
read "check if the file system cares about case at this path". I'm
clearly not the only one confused by this as the dirstate has this
property:
def _checkcase(self):
return not util.checkcase(self._join('.hg'))
Maybe we should even inverse the function and call it fscasefolding()
since that's what all callers care about?
Previously, when we verified the parts of a path in the auditor, we would
validate the deepest directory first, then it's parent, and so on up to the
root. If there happened to be a symlink in the chain, that meant our first check
would likely traverse that symlink. In some cases that symlink might point to
a network filesystem that is expensive, and therefore this simple check could be
very slow.
The fix is to check the path parts starting at the root and working our way
down.
This has a minor performance difference in that we used to be able to short
circuit from the audit if we reached a directory that had already been checked.
Now we can't, but the cost is N dictionary look ups, where N is the number of
parts in the path, which should be fairly minor.
The home of 'Abort' is 'error' not 'util' however, a lot of code seems to be
confused about that and gives all the credit to 'util' instead of the
hardworking 'error'. In a spirit of equity, we break the cycle of injustice and
give back to 'error' the respect it deserves. And screw that 'util' poser.
For great justice.
Python 2.6 introduced the "except type as instance" syntax, replacing
the "except type, instance" syntax that came before. Python 3 dropped
support for the latter syntax. Since we no longer support Python 2.4 or
2.5, we have no need to continue supporting the "except type, instance".
This patch mass rewrites the exception syntax to be Python 2.6+ and
Python 3 compatible.
This patch was produced by running `2to3 -f except -w -n .`.
We retain the forwards because it helps code be more ignorant of
implementation details, but use forwards instead of our own method
definitions since we don't need any truly custom behavior for now.
I've done this as its own step so that it's easy to see that the
posixpath implementations pass the doctests in this package. In a
future patch I'll just make these pure forwards of the methods so that
things using pathutil can be oblivious to the posix nature of these
functions.
This adds dirname and join functions to pathutil which are explicitly for
handling '/' delimited paths. The implementations are basically copy paste from
python's posix os.path.dirname and os.path.join functions.
Given that this path is going to abort, it seems OK to spend the time to do an
alternate lookup to better inform the user. The path returned by util.pathto()
ends with '/' at least in some cases, so os.path.relpath() is used instead,
which requires python 2.6.
Before this patch, "reporelpath()" uses "rstrip(os.sep)" to trim
"os.sep" at the end of "parent.root" path.
But it doesn't work correctly with some problematic encodings on
Windows, because some multi-byte characters in such encodings contain
'\\' (0x5c) as the tail byte of them.
In such cases, "reporelpath()" leaves unexpected '\\' at the beginning
of the path returned to callers.
"lcalrepository.root" seems not to have tail "os.sep", because it is
always normalized by "os.path.realpath()" in "vfs.__init__()", but in
fact it has tail "os.sep", if it is a root (of the drive): path
normalization trims tail "os.sep" off "/foo/bar/", but doesn't trim
one off "/".
So, just avoiding "rstrip(os.sep)" in "reporelpath()" causes
regression around issue3033 fixed by e3dfde137fa5.
This patch introduces "pathutil.normasprefix" to normalize specified
path in the specific way for problematic encodings without regression
around issue3033.