largefiles: simplify lfutil.writehash This was unnecessarily verbose: there is no need to unlink the file when we open it for write anyway, and there is no need to check if the file exists after we created it.

merge with stable

largefiles: copy files into .hg/largefiles atomically Copying from the user cache into .hg/largefiles could fail halfway though with a partially written file.

largefiles: write .hg/largefiles/ files atomically Before, it was possible to create a .hg/largefiles/hash file with truncated content, i.e., content where SHA-1(content) != hash This breaks the fundamental invariant in largefiles that the file content for files in .hg/largefiles hash to the filename.

largefiles: add comment about non-atomic working directory When updating the working copy with a normal 'hg update', we also don't use atomic writes.

tests: use an alias to make msys 'pwd' return paths with forward slashes

tests: skip cvs tests with msys on windows They will fail with weird 'ssh' errors.

tests: introduce 'hghave msys' to skip tests that would fail because of msys

merge with stable

convert: handle trailing slashes in filemap better (issue3124)

tests: remove test-casecollision-i18n.t The result of this test is quite suspect. Remove it for now as it now correctly breaks on HFS+.

posix: fix HFS+ percent-encoding folding We use 'ignore' rather than 'replace' so we don't have to deal with u+FFFD in UTF-8 round-trip.

mail: use quoted-printable for mime encoding to avoid too long lines (issue3075) Quoted-printable was already used for the more critical patch mails, so it should be fine for everything else as well.

notify: add option for writing to mbox This makes it possible to test how the mails that are sent _really_ look like.

mail: mbox handling as a part of mail handling, refactored from patchbomb

patchbomb: minor refactoring of mbox functionality, preparing for move This allows reuse of the mbox functionality but causes minor changes in the patchbomb console output when writing to a mbox.

tests: run check-code with warnings and maintain a whitelist check-code has fine warning checks, but they are a bit noisy and nobody used them. Now the warnings will be run in the test suite, where a list of accepted warnings will be maintained. Those who introduce or touch a line with a warning will now have to update the whitelist ... or fix the warning.

tests: convert test-check-code-hg.py to .t

test-keyword: use inline doctest syntax

merge with stable

tests: test-largefiles.t should not try to escape from $TESTTMP

largefiles: file storage should be relative to repo, not relative to cwd Revealed by the test in 1fd126cd2d91.

clone: don't save user's password in .hg/hgrc (Issue3122)

posix: add extended support for OS X path folding OS X does the following transformation on paths for comparisons: a) 8-bit strings are decoded as UTF-8 to UTF-16 b) undecodable bytes are percent-escaped c) accented characters are converted to NFD decomposed form, approximately d) characters are converted to _lowercase_ using internal tables Both (c) and (d) are done using internal tables that vary from release to release and match Unicode specs to greater or lesser extent. We approximate these functions using Python's internal Unicode data. With this change, Mercurial will (in almost all cases) match OS X folding and not report unknown file aliases for files in UTF-8 or other encodings.

dirstate: use util.normcase to build foldmap

merge with stable

largefiles: fix 'hg clone . ../foo' OSError abort Operating on a non-existant file can cause both IOError and OSError, depending on the function used: open raises IOError, os.lstat raises OSError. The largefiles code called dirstate.normal, which in turn calls os.lstat, so OSError is the right exception to catch here.

tests: test-patchbomb.t cleanup Let commit messages and file names describe the facts without making incorrect guess of the actual handling of it.

tests: show actual output in test-patchbomb.t Drop fixheader and decode base64 encoded content and show non-ascii content.

tests: test findexe() access check fixed in 58f96703a9ab

check-code: fix return code initialization Don't crash when nothing is tested, but return failure when other files than the last one fails.

merge with stable

tests: make test-commandserver.py independent of line ending and slash direction

tests: make test-commandserver.py output readable

tests: 'hghave unix-permissions' for test-import.t 'hg import' of a patch with mode changes requires unix permissions support in the file system.

tests: skip color test on platforms without tic

update: don't clobber untracked files with wrong casing

merge with i18n

merge with i18n

i18n-pt_BR: synchronized with 00276525e2b7

merge with stable

help: Backed out changeset e99facd2cd2a, description of file urls This help entry didn't try to describe the 'localhost' notation. It described a non-standard host-less notation where 'local' just was the first part of a sample relative path. It describe "urls" with relative and absolute paths like: file://file.txt file:///tmp/file.txt file://c:/tmp/file.txt

revset: follow(nosuchfile) should give an empty set (issue3114)

subrepo: fix git status false positive (issue3109) It turns out there's a separate command to force git to look at file contents if the stat has changed.

mdiff: replace wscleanup() regexps with C loops On my system it reduces: hg annotate -w mercurial/commands.py from 36s to less than 8s, to be compared with 6.3s when run without whitespace options.

mdiff: split lines in allblocks() only when necessary These are only required to handle the --ignore-blank-lines case

annotate: support diff whitespace filtering flags (issue3030) splitblock() was added to handle blocks returned by bdiff.blocks() which differ only by blank lines but are not made only of blank lines. I do not know exactly how it could happen but mdiff.blocks() threshold behaviour makes me think it can if those blocks are made of very popular lines mixed with popular blank lines. If it is proven to be wrong, the function can be dropped. The first implementation made annotate share diff configuration entries. But it looks like users will user -w/b for annotate but not for diff, on both the command line and hgweb. Since the latter cannot use command line entries, we introduce a new [annotate] section duplicating the diff whitespace options.

setup: fix py2exe generation broken by c3a6ec304055 (issue3116) Py2exe is patching default distutils Distribution class. This patch ensures that the right one is taken when calling hgdist class.

mdiff: make diffblocks() return all blocks, matching and changed Annotate uses matching blocks not changed ones.

mdiff: extract blocks whitespace normalization in diffblocks() We want to reuse it in annotate for whitespace normalization.

merge with crew

setup: backout 8504699d1aa6 Change wasn't compatible with Python 2.x

tests: don't use stdout redirection for test data Stdout redirection is not binary safe everywhere - for example not on windows with msys.

tests: hide 'No such file or directory' messages Windows will use a different and localized message.

tests: add missing '(glob)'s to match '\' in paths in test output on windows

tests: add 'hghave system-sh' to skip tests on windows

tests: launch hghave with python interpreter without relying on hash-bang

tests: use the specified shell for running old fashioned sh tests

tests: add missing 'hghave unix-permissions'

tests: remove hacks for testing if file or directory exists

tests: don't leave temporary directories without write permission behind

merge with stable

hook: flush stdout before redirecting to stderr When hook output redirection is enabled (e.g. when cloning over ssh), hook output on stdout is redirected to stderr, to prevent the repository data on stdout from being corrupted. In certain cases, the redirection could cause part of the repository data to end up on stderr as well. In case of a clone, this causes: "abort: consistency error in delta!" This was seen with a clone over ssh, an outgoing hook present (any non-python type, e.g. 'pwd'), on certain repositories only, probably depending on the distribution of the sent data) This patch updates the hook redirection code to flush stdout before redirecting, removing the problem.

import: fix parent selection when importing merges With "wp1" and "wp2" the current working directory parents, "p1" and "p2" the patch parents and "parents" the resulting commit parents, the current behaviour is: --bypass --exact p2 parents 0 0 0 [wp1, wp2] 0 0 1 [wp1, wp2]/buggy 0 1 0 [p1] 0 1 1 [p1, p2] 1 0 0 [wp1, wp2] 1 0 1 [p1, p2] 1 1 0 [p1] 1 1 1 [p1, p2] The original behaviour before f53dc0787424 was: --bypass --exact p2 parents 0 0 0 [wp1, wp2] 0 0 1 if p1 == wp1 then [p1, p2] otherwise [wp1, wp2] 0 1 0 [p1] 0 1 1 [p1, p2] This patch restores the previous behaviour when --bypass is not set, and align --bypass behaviour when --exact is not set with merge diffs.