mercurial.el: mark-marker() returns non-nil with no active mark on GNU Emacs

mq.el: declare variable mq-top

mercurial.el, mq.el: remove runtime dependencies on CL package * (require 'cl) at compile-time * don't use map()

hgk: colorize commits by authors Color commits by author. Colors will be assigned to authors by using a configurable list of colors. The line corresponding to current id (the version currently in the working tree) is highlighted in a special font which is also configurable. There is provision to assign fixed colors to authors. See .gitk comment for details.

osutil: use fdopendir instead of dirfd

convert: svn -- fix tags handling They were simply not detected.

convert: svn -- fix 'exists' Previously 'exists' erroneously returned False for empty dirictories. This is wrong since we want to detect even empty 'branches/' or 'tags/'.

merge with goto

mpatch: Define Py_ssize_t for old pythons and use it instead of ssize_t. See also PEP 353. NOTE: Microsoft compilers (8 or earlier) does not have ssize_t.

mpatch: use int rather than ssize_t We might need to be more clever than this to make Py2.5 happy.

osutil: improve portability - manually inline mode_to_kind - remove unused alloca include - remove fstatat and associated bits It's not obvious that there's an advantage to using fstatat in terms of performance. The race-avoidance properties of fstatat aren't terribly useful to us either. So best to avoid it until we figure out how to use it portably.

fix typo

import gettext since '_' is used

explicitely use integer division

fix UnboundLocalError, refactor a bit bin wasn't defined in all branches (bug introduced in 82b4ff3abbcd

bdiff: tweaks for large files - adjust the common line threshold to .1% this speeds up a delta of 7M lines of source from 10m to 40s - adjust the scaling of the hash array down a bit as it was raising the peak memory usage significantly

revlog: break up compression of large deltas Python's zlib apparently makes an internal copy of strings passed to compress(). To avoid this, compress strings 1M at a time, then join them at the end if the result would be smaller than the original. For initial commits of large but compressible files, this cuts peak memory usage nearly in half.

revlog: fix caching of buffer objects

chunkiter: handle large reads more efficiently - for large reads, don't attempt to read more than necessary - if we've gathered the exact number of bytes needed, avoid a string copy

revlog: avoid large yields in group() Split large yields so that the downstream consumer (chunkbuffer) will avoid some pain when reading ahead.

chunkiter: simplify iter logic

chunkbuffer: removed unused method and arg

revlog: reduce memory usage in addgroup - use a buffer to extract the delta from a chunk - avoid concatenating to a compressed delta - use a buffer to directly extra full text from a trivial delta - delete chunk and delta objects after use

mpatch: allow buffer objects for input

Merge with crew Those crew folks are getting lazy about pulling from upstream before committing.

Merge with crew

Merge with -stable

findcopies: fix rename bug We've fiddled with this line several times, and an old bug has reappeared from it. Let's take a peek at the history. The original "or" (rev 3674, in 0.9.2 and 0.9.3): http://www.selenic.com/hg/rev/9103dab96093 Then I changed it to an "and" to fix a bug (rev 4304): http://www.selenic.com/hg/rev/4787e2b0dd03 Then for reasons now lost in the mists of time, I dropped half (rev 4399): http://www.selenic.com/hg/rev/93652499bed3 Then we added back the "or" (rev 4416, in 0.9.4): http://www.selenic.com/hg/rev/bb1800a7d7e1 So it seems it ought to be "and".

mq: clarify queue top message

test-convert-darcs: don't let $HOME/.darcs pollute the test environment

convert: refactor sink initialisation, to remove hardcoding of hg We also introduce options to explicitly set the source and destination repository types. Improve testing of corner cases a little.

convert: add default constructor for converter_sink

convert: make contents of "extra" dict available from sources, for sinks. This breaks hash preservation for hg->hg conversion, as each converted change gets a convert_revision item added to its extra dict. Ugh.

convert: rename convert_svn to svn_source

convert: report errors more meaningfully if run with --traceback

osutil: Solaris build fix

Fix build error with Sun C compiler.

osutil.c: use strncpy instead of strncat

findcopies: fix rename bug We've fiddled with this line several times, and an old bug has reappeared from it. Let's take a peek at the history. The original "or" (rev 3674, in 0.9.2 and 0.9.3): http://www.selenic.com/hg/rev/9103dab96093 Then I changed it to an "and" to fix a bug (rev 4304): http://www.selenic.com/hg/rev/4787e2b0dd03 Then for reasons now lost in the mists of time, I dropped half (rev 4399): http://www.selenic.com/hg/rev/93652499bed3 Then we added back the "or" (rev 4416, in 0.9.4): http://www.selenic.com/hg/rev/bb1800a7d7e1 So it seems it ought to be "and".

osutil: more tidying - do_stat -> keep_stat - all_kinds -> !need_stat - simplify main error logic - reorder declarations

osutil: move file list loop to its own function

osutil: simplify DT_REG support

osutils: pull file stat loop into its own function

osutil: fold stat paths together - simplify st/py_st logic - use stp to point to stat buffer - combine stat paths

osutil: more cleanups - eliminate a level of nesting - untab targets of gotos

osutil: eliminate alloca call - make full_path a PATH_MAX + epsilon local buffer - use strncpy and strncat

osutil: cleanups - use tabs - eliminate old-style function pointer calls - eliminate weird scoping - eliminate assignment-in-iff - use !foo rather than foo == NULL

Execution bit detection fixes for VFAT on Linux On Linux VFAT execution mode can be modified, but changes don't persist a filesy stem remount. The current test can be trickled by this. We can help with the det ection of VFAT checking whether new files get created with the execution bits on (as usually these partitions are mounted with the exec option, for convenience) .

churn: simplify code to get terminal width

patch: fix git sendmail handling without proper mail headers

hgk: allow any extdiff command for visual diff

osutil.c: style fix - delete trailing end-of-line spaces

convert: do not output when trying to load svn bindings

convert: split converters into sink/source

Test darcs converter

convert: fix darcs_source.pull() under windows

convert: fix darcs_source._run() under windows

hghave: detect darcs client

hghave: reorder check functions and entries

Test workingctx exec/link bit for copies. Also, make the test runnable on platforms without symlinks.

Fix workingctx exec/link bit of copies on non-supporting systems

test-manifest: test symlink and exec bit display

Display symlink or executable bit with manifest -v New output looks like: 644 a 755 * b/a 644 @ l

convert: quote "^" to avoid windows using it as an escape char.

Always copy the necessary files before applying a git patch This patch removes the "copymod" attribute from the gitpatch class. AFAICS, that attribute was only used to delay the copying of renamed/copied files if there are no other changes to the target, but in this case, if there are changes to the source, we'll end up copying the wrong version. This should fix issue762.

mercurial_sink: regrab locks in setbranch

convert --filemap: reduce memory usage

convert: change SKIPREV to 'SKIP'

Fix bad lambda prototype in workingctx.fileflags()

osutil.c: use readdir instead of readdir64 Some systems (e.g. *BSD) don't have a readdir64 function - the regular readdir already uses 64-bit types. On other systems (Linux, Solaris, ...), if Python was compiled with large file support, Python.h will define _LARGEFILE_SOURCE and _FILE_OFFSET_BITS=64, so that any call to readdir will actually be a call to readdir64. If Python was not compiled with large file support, we probably don't want to define these macros to avoid ABI problems.

osutil.c: include Python.h before the other headers This is recommended by the Python documentation, since Python.h defines some constants.

Add osutil module, containing a listdir function. This is similar to os.listdir, only it returns a sorted list of tuples.

hgk: add basic usage and configuration documentation

hgk: add context menu visual diff action * * *

hgk: add debug-config command to pass configuration options

hgk: fix right-mouse button handling under macosx

Test *_ISLINK merge environment vars

merge: provide *_ISLINK environment vars to merge helper Sets HG_MY_ISLINK, HG_OTHER_ISLINK, HG_BASE_ISLINK in environment. Without these variables, it's impossible for the merge application to know whether the 'other' and 'base' files were symlinks in their original contexts. For the purposes of the merge they are always emitted as small text files.

context: add fileflags() to avoid rebuilding manifests

run-tests.py: allow a different temporary directory to be specified

Update test for new output produced since 17ed9b9a0d03

Fix accidentally broken test.

Merge with crew.

Allow tests to run in parallel.

run-tests.py: reorder options alphabetically.

convert_svn: add --filemap support

convert_cvs: add --filemap support

convert_git: add --filemap support

mercurial_source: add --filemap support

convert: add a mode where mercurial_sink skips empty revisions. The getchanges function of some converter_source classes can return some false positives. I.e. they sometimes claim that a file "foo" was changed in some revision, even though its contents are still the same. convert_svn is particularly bad, but I think this can also happen with convert_cvs and, at least in theory, with mercurial_source. For regular conversions this is not really a problem - as long as getfile returns the right contents, we'll get a converted revision with the right contents. But when we use --filemap, this could lead to superfluous revisions being converted. Instead of fixing every converter_source, I decided to change mercurial_sink to work around this problem. When --filemap is used, we're interested only in revisions that touch some specific files. If a revision doesn't change any of these files, then we're not interested in it (at least for revisions with a single parent; merges are special). For mercurial_sink, we abuse this property and rollback a commit if the manifest text hasn't changed. This avoids duplicating the logic from localrepo.filecommit to detect unchanged files.

convert: readd --filemap To handle merges correctly, this revision adds a filemap_source class that wraps a converter_source and does the work necessary to calculate the subgraph we're interested in. The wrapped converter_source must provide a new getchangedfiles method that, given a revision rev, and an index N, returns the list of files that are different in rev and its Nth parent. The implementation depends on the ability to skip some revisions and to change the parents field of the commit objects that we returned earlier. To make the conversion restartable, we assume the revisons in the revmapfile are topologically sorted.

convert: move filemapper class to a separate file

convert: disable current --filemap support The --filemap support in hg convert doesn't handle merges correctly. (And after 33015dac5df5 I managed to break it even for simple cases where we don't want the first revision.)

convert: allow the converter_source to say "skip this revision" If getchanges returns a string, it's assumed to be the id of an already converted revision. We map the current revision to the same revision this converted revision was mapped to. To allow skipping a root revision, getchanges can return the special string 'hg-convert-skipped-revision' (a.k.a. common.SKIPREV), which hopefully won't clash with any real id. The converter_source is responsible for rewriting the parents of the commit objects to make sure the revision graph makes sense.

convert: pass the order of the revmapfile to the converter_source The entries in this file are supposed to be topologically sorted and this may be useful for a converter_source.

Merge with crew

merge: add debug diagnostics for findcopies

changegroup: avoid large copies - handle chunk headers separately rather than prepending them to (potentially large) chunks - break large chunks into 1M pieces for compression - don't prepend file metadata onto (potentially large) file data

revlog: generate trivial deltas against null revision To avoid extra memory usage and performance issues with large files, generate a trivial delta header for deltas against the null revision rather than calling the usual delta generator. We append the delta header to meta rather than prepending it to data to avoid a large allocate and copy.

Merge with crew

Merge with crew

bdiff: switch to lyhash lyhash is a very simple and fast hash function that had the fewest hash collisions on a 3.9M line text corpus and 190k line binary corpus and should have significantly fewer collisions than the current hash function.

bdiff: use INT_MAX to avoid some inner loop comparisons

bdiff: simple splitlines optimization

I have spotted the biggest bottleneck in "bdiff.c". Actually it was pretty easy to find after I recompiled the python interpreter and mercurial for profiling. In "bdiff.c" function "equatelines" allocates the minimum hash table size, which can lead to tons of collisions. I introduced an "overcommit" factor of 16, this is, I allocate 16 times more memory than the minimum value. Overcommiting 128 times does not improve the performance over the 16-times case.

Merge with crew-stable

test-convert-git: support older git client (1.4.4.4)

Merge with crew-stable

test-convert-cvs: handle older cvs client

Merge with crew

convert: fix a few residual bugs in darcs importer

Merge with crew.

convert: support darcs as a source repo

convert: fail properly if we can't read a source hg repository

util: add default argument to strdate

convert: add before/after hooks for converter sources

merge with -stable

set_exec: do not chmod a symlink

Merge with crew-stable

Test debugindexdot