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

Merge with crew-stable

Test mercurial convert sink removes empty directories.

convert: fix missing import

Merge with crew-stable.

convert: delete empty directories if deleting a file (bug 754)

Merge with crew.

darcs2hg: Added support for darcs tags.

darcs2hg: Now detects and recovers from simple darcs conflicts.

darcs2hg: Now understands files that were explicitly renamed in darcs.

revlog: fix revlogio.packentry corner case We want to store version information about the revlog in the first entry of its index. The code in packentry was using some heuristics to detect whether this was the first entry, but these heuristics could fail in some cases (e.g. rev 0 was empty; rev 1 descends directly from the nullid and is stored as a delta). We now give the revision number to packentry to avoid heuristics.

Don't try to determine interactivity if ui() called with interactive=False. WSGI applications are not supposed to refer to sys.stdin. In ed6df6b1c29a, hgweb and hgwebdir were fixed to pass interactive=False to their ui()'s, but sys.stdin.isatty() was still called by the ui objects. This change makes sure only the ui.fixconfig() method will call ui.isatty() (by making the ui._readline() method, which is currently only called from ui.prompt(), private). ui.fixconfig() is changed to let config files override the initial interactivity setting, but not check isatty() if interactive=False was specified in the creation of the ui.

Merge with crew-stable

convert_git: avoid returning two entries for the same file in getchanges This could happen in merge changesets if the merged file was different from both parents.

mq: fix qrefresh -e with no patches applied

Push over HTTP: really tell the user the size of the bundle

hgwebdir: ignore hgrc parse errors while building the index page An error in the .hg/hgrc file from a repository would prevent the following repos from being shown in the index page. The IOError handling was unnecessary - it's already handled in readconfig. This should fix issue731. The error in the .hg/hgrc file will still prevent the repo from being exported with hgweb.

hgweb: don't raise an exception when displying empty repos The nullid node claims it's in the default branch, but the branch dict is empty. This fixes the main symptom from issue696, but we may want to set branchtags()['default'] = nullid somewhere for empty repos.

make identify an optionalrepo command Fixes issue726.

copy: if destination ends with "/", make sure it's a directory Fixes issue724.

manifest: minor performance tweak

dirstate: speed up read and write read: - single call to len(st) - fewer assignments for position tracking - don't split apart tuple from unpack - use a literal for the unpack spec write: - localize variables and functions - avoid copied function call - use % for string concatenation

dirstate: make dir collision logic faster - shortcircuit decpath if we haven't built the _dirs map - increment only for leafnodes of directory tree (this should make construction more like O(nlog n) than O(n^2))

Merge with crew

convert-cvs: update test to pass on Linux

verify: report first bad changeset - add linkrev and file arguments to err - track lowest reported linkrev in err - find linkrev in a couple cases - sort linkrevs when reporting a bunch of errors against a file - tidy up lots of messages - make more robust against damaged indexes - report first bad changeset if an error is encountered

revlog: more robust for damaged indexes - fix flag reporting - make broken length reporting smarter

debugindex: more robust for damaged indexes

revlog: fix inlined revision transaction extra data (issue 749)

Merge with crew-stable

Test renaming files brought by merge second parent (issue 746)

localrepo: fix undelete() on merge working directory (issue 746)

bisect: remove useless try/except

Merge with crew-stable

Test static-http calls on '/' URI (issue 747)

statichttprepo: fix calls on '/' URI (issue 747)

bisect: don't write data on failure

bisect: do silent init if necessary

debuginstall: stop looking for patch utility, just test it

Updated zsh completion from issue722.

Merge with crew-stable

convert: test cvs repository conversion

convert: workaround for cvsps.cache under Windows In non-cygwin environment, cvsps fails to create its cache directory and redirect its output to stderr. Just ignore the error and capture stderr as well.

convert: fix remote cvs file paths separator

convert: avoid interpreting Windows path as CVS connection strings. CVS connection strings regexp detect colons to separate protocols from path and login. Unfortunately, Windows paths contains colons and were interpreted as rsh connection strings.

convert: call popen2 in binary mode, with a command string.

hghave: detect cvs and cvsps availability "cvsps -h" was returning 1 in cygwin, probably because CVSROOT was unset, which does not prevent it to work correctly.

hghave: wrap command output matching

Merge with crew-stable

Test parents behaviour with files.

parents: make it match the doc when called on a file parents used to return the parent revision of the last revision updating selected file, instead of the updating revision itself.

Merge with crew-stable

Document pseudo-tags "null" and ".".

Merge with crew

mq: fix regression in 6fd953d5faea It caused qrefresh; qrefresh -X addedfile to leave the file as normal in dirstate. It should be marked added on the second pass.

Merge with crew-stable

sshrepo: fix Windows command quoting

extdiff: avoid repr() doubling paths backslashes under Windows

merge with crew-stable

Prevent WSGI apps from touching sys.stdin by setting ui.interactive to False.

interhg: allow more flexible pattern specification (fixes 2/3 of issue699) - something else than "pat" followed by a number can be used as key - something else than "/" can be used as delimiter - "ilmsux" flags (e.g. "i" for case insensitive) can be used

Don't decode unicode strings. This does not work with python2.3 and is probably not useful anyway with 2.4.

Update test-record for 3ef190234b55

record: change wording of initial per-file prompt

hgk: add repo root to window title

Forgot to quote "6^" in test-parentrevspec (see 94e77a174f55)

test-convert-hg-source: enforce conversion order with --date

convert: rename a class and a function

convert: mercurial_source: also search for copies in modified files There are some corner cases where we may have a copy in a file that isn't in the added list: - the result of a hg copy --after --force - after a merge across a (local) rename

convert: clear the dirstate before a conversion, invalidate it afterwards Clearing it before the conversion protects us from whatever data were there (file copies in particular). Invalidating it after the conversion avoids writing a possibly inconsistent dirstate to disk.

convert: avoid dirstate checks; add a test During a conversion, the dirstate contents are not consistent - there are files that may be missing from the dirstate and there may be files that shouldn't be in the dirstate. While this is not fixed, don't mark files as added - put them directly in state 'n'ormal.

merge with crew

merge with crew-stable

fix the dummy 00changelog.i not being copied on clone

convert/subversion: Use util.set() instead of set() for python2.3 compatibility

Fix \ inside ` problem in test-convert-svn (avoids code duplication, too)

merge with crew-stable

Make test-convert-git compatible with other git versions (tested with 1.4.4.3)

hgweb: Show date of last change for each file in manifest

remove unneeded statement

kill lonely docstring

Atom support: replaced xhtml namespace prefix with default namespace

Added support for the Atom syndication format

Include . in PYTHONPATH (makes testing unbundled extensions easier)

Make run-tests.py work when invoked outside of tests. This makes it usable for unbundled extensions.

Merge with crew-stable

When reverting a file which was renamed, also revert the old name

Avoid a working dir walk while trying to detect copies for diff --git

Make hg diff --git -r revA:revB detect (inverted) copies if revA > revB

bundle: fix up lazy reads

Make bundlerepo lazier This allows the localbranch extension to override sopener in time.

Merge with crew

convert: new config variable hg.tagsbranch controls which branch tags are committed to

Push capability checking into protocol-level code.

Turn capabilities into a mutable set, instead of a fixed tuple.

Automated merge with http://hg.intevation.org/mercurial/crew

convert: document filemap.

Test svn repository conversion

hghave: detect subversion bindings availability

hghave: detect subversion client and admin tools availability

hghave: wrap command output matching

run-tests: append PYTHONPATH instead of overwriting it Optional dependencies like svn bindings not necessarily live in site-packages.

convert: svn: check for branch movement in any log entry, not just the first. This fixes, for example, r4151 D /branches A /project/branches (from /branches:4150) A /project/tags (from /tags:4150) A /project/trunk (from /trunk:4150) D /tags D /trunk

Merge with crew-stable.

correctly record file-level history when the local side renames a directory

Fix theoretical issue in filecommit. If the file was copied, we don't want to reuse the original entry. I think this is mostly a theoretical issue - when there are copies, fp1 == nullid, so it's very unlikely that the fl.cmp(fp1, t) would think the file was unmodified. In any case, if there was a copy, we should forcefully create a new entry.

Avoid adding to the changelog executable files added by the second parent. This was a regression introduced by b51a8138292a.

Merge with crew-stable

avoid double slash problem mentioned in issue695

clone: remove "file://" before making the path absolute This avoids writing bogus paths to .hgrc. Fixes issue695.

pull -u: if "url#rev" was given, update to rev

clone: if "url#rev" was given, update to rev

parseurl: also return the revision after the "#"; add a test

pull -r: pass the revisions as the heads argument of findincoming This can make a hg pull -r faster if the remote repo has many heads, and fixes an "abort: received changelog group is empty".

Merge with crew-stable

convert: fix /dev/null redirections under Windows

Merge with crew-stable

Test git repository conversion

hghave: detect git availability

convert: fix issue702 about GIT_DIR= construct unsupported under Windows.

convert: gitcmd wrapper for os.popen

Improve extdiff configuration. In addition to the old cmd.foo, opts.foo hgrc entries, allow a more simple alias = command [opts]... form. For example: [extdiff] cdiff = colordiff -uprN

cgi: sys.path.insert should be before importing mercurial thanks to R. Burke

merge with crew

improve error message for 'hg merge' when repo already at branchtip

imerge: add automerge flag to attempt to batch merge all conflicts

Merge with crew

imerge: fix ancestor calculation If wctx already has two parents, ancestor calculation is wrong. Normally merge is called before wctx gets the second parent, so we simulate this in imerge by temporarily popping the second parent before calling filemerge. Highly dirty. This patch also handles the ParseError move from cmdutil to dispatch.

merge with crew-stable

fix test-permissions output

merge with crew-stable

testcase for issue705 (fixed by 316ce5e85b3e)

merge with crew-stable

merge with -stable

check exec: return fallback in case of error during the check If there is any error while checking if exec is supported, we can return fallback. fix issue705

mq: some improvements to header rewriting (closes #690) Tested with qrefresh -m against the mm series.

Merge with crew-stable

Do not lose previous tags when adding a tag in raw mode

merge: forcefully mark files that we get from the second parent as dirty After a hg merge, we want to include in the commit all the files that we got from the second parent, so that we have the correct file-level history. To make them visible to hg commit, we try to mark them as dirty. Unfortunately, right now we can't really mark them as dirty[1] - the best we can do is to mark them as needing a full comparison of their contents, but they will still be considered clean if they happen to be identical to the version in the first parent. This changeset extends the dirstate format in a compatible way, so that we can mark a file as dirty: Right now we use a negative file size to indicate we don't have valid stat data for this entry. In practice, this size is always -1. This patch uses -2 to indicate that the entry is dirty. Older versions of hg won't choke on this dirstate, but they may happily mark the file as clean after a full comparison, destroying all of our hard work. The patch adds a dirstate.normallookup method with the semantics of the current normaldirty, and changes normaldirty to forcefully mark the entry as dirty. This should fix issue522. [1] - well, we could put them in state 'm', but that state has a different meaning.

gendoc: use demandimport

Merge with -crew

Merge with -stable

match: handle large regexes Some Python versions don't handle large regexes, so when we hit an overflow, split our regex in two.

make hg add of a removed file unremove the file if a file was removed, hg add shouldn't mark the file as added, but as normal see issue683

abort when adding a file in merged state

add import from url capability

convert: load parent commits on-demand

convert: wrap cached commits author remapping

path_auditor: cache names of audited directories We use a separate cache to avoid problems with audit = path_auditor(repo.root) audit("subrepo") audit("subrepo/file") whitelisting "subrepo" (which is fine) and then using the same whitelist with "subrepo/file" (which is not fine). Since we create a separate path_auditor for every path on the command line, a "hg add dir/a dir/b dir/c" will still lstat dir 3 times just to audit the paths.

Quote ^ and ~ chars in test-parentrevspec. At least ^ causes problems on some sh implementations.

Remove demandimport.enable from dispatch.py

Enable demandimport only in scripts, not in importable modules (issue605) This way other applications can choose if and when they want this feature, because it might be problematic if those applications rely on ImportError.

Merge with crew-stable

Fix sshrepo.unbundle We weren't reading all the data sent by the server. Depending on the system, the remote hg (actually, the remote python) could send a "close failed: [Errno 32] Broken pipe", making some tests fail.

run-tests.py: pass --install-scripts to setup.py This avoids possible problems when the user has a ~/.pydistutils.cfg. Fixes issue638. Patch provided by Ralf Schmitt.

recognize user:password in URLs that include port numbers. when giving user/pwd in an URL, eg. hg pull http://user:pwd@host.com:666/hg/something, hg would still ask for user/pwd in interactive mode (or fail in non-interactive)

convert: fix mercurial_sink.putcommit Changeset 4ebc8693ce72 added some code to putcommit to avoid creating a revision that touches no files, but this can break regular conversions from some repositories: - conceptually, since we're converting a repo, we should try to make the new hg repo as similar as possible to the original repo - we should create a new changeset, even if the original revision didn't touch any files (maybe the commit message had some important bit); - even if a "regular" revision that doesn't touch any file may seem weird (and maybe even broken), it's completely legitimate for a merge revision to not touch any file, and, if we just skip it, the converted repo will end up with wrong history and possibly an extra head. As an example, say the crew and main hg repos are sync'ed. Somebody sends an important patch to the mailing list. Matt quickly applies and pushes it. But at the same time somebody also applies it to crew and pushes it. Suppose the commit message ended up being a bit different (say, there was a typo and somebody didn't fix it) or that the date ended up being different (because of different patch-applying scripts): the changeset hashes will be different, but the manifests will be the same. Since both changesets were pushed to public repos, it's hard to recall them. If both are merged, the manifest from the resulting merge revision will have the exact same contents as its parents - i.e. the merge revision really doesn't touch any file at all. To keep the file filtering stuff "working", the generic code was changed to skip empty revisions if we're filtering the repo, fixing a bug in the process (we want parents[0] instead of tip).

Add parentrevspec extension

Move enabling of demandimport to dispatch.py This avoids loading many modules before enabling demandimport.

Move cmdtable and reposetup handling out of extensions.py A new function (extensions.extensions) allows the code that is interested in those attributes to handle them directly. This allows some cleanups of extensions.py. Notably, we can remove the extensions.commandtable hack. It also makes it easier to add standard extension attributes, like a "hgwebsetup" function or a "helptable" dict that augments the data in help.py, etc.

Adjust contrib/hg-ssh for moved dispatch() function.

Merge with crew

verify: report expected linkrev

dispatch: move command dispatching into its own module - move command dispatching functions from commands and cmdutil to dispatch - change findcmd to take a table argument - remove circular import of commands in cmdutil - privatize helper functions in dispatch

move parseurl from cmdutil to hg

fix-up references to repo.RepoError

remove unneeded imports of mdiff

Merge with stable

merge with -stable

clone: do not delete the target if only the update fails

Merge with crew-stable

Merge with hg-stable

test-grep: use lowercase hg command

Merge with crew-stable

mq: truncate patch just before rewriting header

convert: raise Abort instead of NoRepo when CVS pserver auth fails. At this point we know the source is CVS, so we should not go through the rest of the converters.

convert: hg: optionally create branches as clones If convert.hg.clonebranches is set, branches will be created as clones of their parent revisions. All clones will be subdirectories of the destination path.

Use smerge minor mode in emacs to resolve conflicts. The condition-case is used to suppress errors with emacs versions before 21.

Merge with crew-stable

test-grep: test issue 685

Fix issue 685: trackback in grep -r after rename

merge with crew-stable

The >| redirection does not exist in plain sh.

Merge with crew-stable

hghave: test hotshot dependencies (debian does not provide profile)

Merge with crew-stable

test-profile: fix grep, check hotshot availability

merge with -stable

Return 0 as 'hg bisect reset' is successful

fix bundlerepo broken by 4205f626dc05 since we now use struct in case of lazy parser, this have to be an integer instead of None see issue681

Merge with crew-stable

Test --time, --profile and --lsprof

demandimport: ignore resource module, not available under Windows.

demandimport: ignore pwd and grp. Both are unavailable under Windows and tarfile detects them via ImportError.

imerge: simplify 1d5ebb0d366f

imerge: fix status file lookups

demandimport: ignore pwd and grp. Both are unavailable under Windows and tarfile detects them via ImportError.

util: ignore invalid path errors in path_auditor.

merge with crew-stable

merge: fix a copy detection bug (issue672) When merging rev1 and rev2, we want to search for copies that happened in rev1 but not in rev2 and vice-versa. We were starting the search at rev1/rev2 and then going back, stopping as soon as we reached the revno of the ancestor, but that can miss some cases (see the new test-issue672). Now we calculate the revisions that are ancestors of rev1 or rev2 (but not both) and make sure the search doesn't stop too early. Simplified test provided by mpm, based on a test case provided by Edward Lee.

merge with crew-stable

Added test for commands.dispatch (especially 88803a69b24)

test-context.py: Don't create second repo instance (fixed since 7b5723c95a82)

fancyopts: Copy list arguments in command table before modifying. Before this, executing commands.dispatch(['log', '-r', '0']) commands.dispatch(['log', '-r', 'tip']) would look like: hg log -r 0 hg log -r 0 -r tip Reported by TK Soh, patch by Alexis S. L. Carvalho

Print less scary message if the system supports symlinks: "Skipping test-no-symlinks: system supports symbolic links" instead of "Skipping test-no-symlinks: unexpected feature: symbolic links"

Hide absolute path from test-no-symlinks output. And add missing eol and remove trailing space which where introduced by the patches adding this test.

Make a few portability improvements to path auditing code.

Make audit_path more stringent. The following properties of a path are now checked for: - under top-level .hg - starts at the root of a windows drive - contains ".." - traverses a symlink (e.g. a/symlink_here/b) - inside a nested repository If any of these is true, the path is rejected. The check for traversing a symlink is arguably stricter than necessary; perhaps we should be checking for symlinks that point outside the repository.

mq: teach qpop about symlinks

debugstate: print symlinks as 'lnk', not '777'

manifest: accept -r for rev specification

record: improve docs, improve prompts

Merge with crew-stable

hghave: prefix temporary files with "hg-hghave-"

test-no-symlinks: rename .bundle into .hg for consistency.

Merge with crew-stable

fix up previous commit for stable

Merge with crew-stable

Cache extension load failures. hg commands call extensions.loadall twice, once during dispatch and once when the repository is instantiated. Without this change, load caches successful loads, but not unsuccessful, causing errors to be displayed twice.

merge with -stable

unobfuscate part of f94dbc6c7eaf

Merge with crew-stable

Test symlink handling on platform not supporting them

hghave: feature absence can be checked by prefixing with 'no-'

merge with -stable