inotify: fix issue1375, add a test. The biggest problem was the data structure, which did not allow changing a file into a directory or vice versa. This problem is fixed by b55d44719b47. The walk() method also had an issue in this case: - we know 'path' as a directory. inotify server sleeps. - 'path' is deleted - 'path' is recreated as a file - the server catches up here, and see the deletion. it instantiates a scan(), which in its turn calls for walk(repo, path). - walk() then assumes that 'path' is a directory and calls os.listdir on it, which raises an OSError(errno.ENOTDIR) Catch the error, and yield the file instead of the directory contents.

inotify: server: new data structure to keep track of changes. == Rationale for the new structure == Current structure was a dictionary tree. One directory was tracked as a dictionary: - keys: file/subdir name - values: - for a file, the status (a/r/m/...) - for a subdir, the directory representing the subdir It allowed efficient lookups, no matter of the type of the terminal leaf: for part in path.split('/'): tree = tree[part] However, there is no way to represent a directory and a file with the same name because keys are conflicting in the dictionary. Concrete example: Initial state: root dir |- foo (file) |- bar (file) # data state is: {'foo': 'n', 'bar': 'n'} Remove foo: root dir |- bar (file) # Data becomes {'foo': 'r'} until next commit. Add foo, as a directory, and foo/barbar file: root dir |- bar (file) |-> foo (dir) |- barbar (file) # New state should be represented as: {'foo': {'barbar': 'a'}, 'bar': 'n'} however, the key "foo" is already used and represents the old file. The dirstate: D foo A foo/barbar cannot be represented, hence the need for a new structure. == The new structure == 'directory' class. Represents one directory level. * Notable attributes: Two dictionaries: - 'files' Maps filename -> status for the current dir. - 'dirs' Maps subdir's name -> directory object representing the subdir * methods - walk(), formerly server.walk - lookup(), old server.lookup - dir(), old server.dir This new class allows embedding all the tree walks/lookups in its own class, instead of having everything mixed together in server. Incidently, since files and directories are not stored in the same dictionaries, we are solving the previous key conflict problem. The small drawback is that lookup operation is a bit more complex: for a path a/b/c/d/e we have to check twice the leaf, if e is a directory or a file.

merge with crew-stable

gitweb, monoblue: fix double-spacing in file view (issue1733) The error was introduced in 338412820a57.

util: use propertycache in opener instead of __getattr__

mq: use ui.status when pushing and popping patches

mq: qpop now tells which patches are popped When navigating through a patch series by qpushing and qpopping it was easy to get lost because qpop was silent. I found myself often running qnext after qpop to see which patch I just dropped - especially if I was about to qrm it. This patch makes qpop more symmetric to qpush by showing which patches are applied/"unapplied". I think it is a good change even though it changes output by adding a new line of output.

merge with crew-stable

url: fix use of non-int port in https connections (issue1725) Versions of Python before 2.6 cannot automatically convert a given port number to an integer, so we add a conversion to coerce the given input to an int.

doc/Makefile: let hg1.gendoc.txt depend on gendoc.py

keyword: use all caps for 'note:' in help string The diff, export, log, and status commands write it as 'NOTE:'.

notify: fix indentation in module docstring

notify: cleanup module docstring

convert: fix inconsistent indentation in help text The sort modes flags and convert.cvsps options were indented while the convert.hg, convert.svn, and convert.p4 options were not. I've now un-indented all of them.

Merge with stable

filectx: add a hex method

win32mbcs: also wrap windows.pconvert()

Use os.path.split() for MBCS with win32mbcs extension.

win32mbcs: look up modules using sys.modules (issue1729) globals()[module] doesn't work for modules inside of packages, such as os.path.

fix memory usage of revlog caches by limiting cache size [issue1639]

keyword: eliminate potential reference cycles from kwrepo - delete kwrepo.commitctx after using the tweaked version - prefer self.hook over repo.hook to avoid nesting Also pass arguments to commit as arbitrary list. Thanks to Simon Heimberg and Matt Mackall for guidance.

convert/cvs: improve error message on unexpected server output.

transaction: fix uncaught ENOENT (issue1724) The opener raises an IOError on errors where transaction expects an OSError.

branch heads: fix regression introduced in e67e5b60e55f (issue1726) For merge nodes it is not adequate to only check a single possible branch head for whether it is an ancestor of the latest head, but it needs to be done for each possible branch head.

subrepo: use hg.repository instead of creating localrepo directly this way, extensions' reposetup will be called, which allows for git subrepos to be handled by hg-git (and I believe the same goes for svn and hgsubversion)

ignore: separate pattern extraction from match compilation hgignore files have slightly different syntax from match objects, e.g. syntax: foo headers, regexp: forms, and re being relre. Put conversion from hgignore syntax into match syntax into a standalone function so that it is easier to validate hgignore hunks externally.

merge with crew-stable

util: remove unused bufsize argument Removed it correctly this time: the subprocess default is 0, not -1 and so we must pass -1 explicitly. Added a comment to that effect.

convert/cvs: improve error message on unexpected server output.

typos: "it's" -> "its"

convert: fixed typo in docstring

Merge OS X performance regression fix

Merge backed out change

Backed out changeset fce065538bcf: it caused a 5x performance regression on OS X

help: wrapped help strings at 78 characters

keyword: collect kwmaps using a generator expression Allowed for Python 2.4.

commands: avoid bad linebreak in addremove docstring

Makefile: wrap hg.pot strings at 82 chars By default, xgettext will wrap strings at 79 characters. This width includes the two double-quotes and the '\n' added. Our strings are now wrapped at 78 characters and a line can therefore be up to 82 characters wide in the .po file. Setting the width to 82 characters ensures that xgettext does no silly wrapping on our already-wrapped strings while still wrapping long one-liners at 82 characters.

zeroconf: wrapped docstrings at 78 characters

win32mbcs: wrapped docstrings at 78 characters

transplant: wrapped docstrings at 78 characters

share: wrapped docstrings at 78 characters

record: wrapped docstrings at 78 characters

rebase: wrapped docstrings at 78 characters

purge: wrapped docstrings at 78 characters

patchbomb: wrapped docstrings at 78 characters

parentrevspec: wrapped docstrings at 78 characters

pager: wrapped docstrings at 78 characters

notify: wrapped docstrings at 78 characters

mq: wrapped docstrings at 78 characters

keyword: wrapped docstrings at 78 characters

interhg: wrapped docstrings at 78 characters

highlight: wrapped docstrings at 78 characters

hgk: wrapped docstrings at 78 characters

hgcia: wrapped docstrings at 78 characters

graphlog: wrapped docstrings at 78 characters

fetch: wrapped docstrings at 78 characters

extdiff: wrapped docstrings at 78 characters

convert: wrapped docstrings at 78 characters

color: wrapped docstrings at 78 characters

churn: wrapped docstrings at 78 characters

children: wrapped docstrings at 78 characters

bugzilla: wrapped docstrings at 78 characters

bookmarks: wrapped docstrings at 78 characters

acl: wrapped docstrings at 78 characters

commands: wrapped docstrings at 78 characters We have always had a left margin of 4 characters -- probably just because that's how docstrings for top-level functions turn out by default, but it also looks nice in the built-in help. The docstrings were wrapped at 70 characters, which is the default for Emacs. However, this gives a right margin of 10 characters in a standard 80 character terminal. I've now wrapped the relevant docstrings at 78 characters, effectively killing the right margin. The asymmetric margins looked a bit odd and some of the text looked cramped with a right margin, so Dirkjan and I felt that it was best to remove it entirely. The two character gap was kept to have some space between the border of the terminal -- it will also make diffs involving the docstrings fit in a 80 character line.

extdiff: fix indentation and use gettext

filemerge, subrepo: correct indention

ui: extract choice from prompt avoid translating single characters (as l for _local or sym_link)

patchbomb: use local time for the Date: header

patchbomb: fix From_ in the message's envelope Localized date in the From_ prevents MUAs like mutt from parsing mbox files generated by patchbomb. Using a 24 characters date in asctime format instead.

merge with crew-stable

Bourne shells do not maintain $PWD; update tests accordingly

tests: remove more instances of export FOO=bar bashism

tests: remove more instances of export FOO=bar bashism

test-gpg: suppress secure memory warning Some platforms (Solaris for one) does not support secure memory and would give a warning.

test: fix for compatibilty with true Bourne /bin/sh 'export foo=bar' is an extension, replaced with 'foo=bar; export foo'

p4: simplify sort key

localrepo: removed unnecessary revkey sort helper

compat: remove unnecessary diamond-shaped multiple inheritance

split local and stdlib module imports (eases migration issues)

compat: don't reference an exception var inside a lambda

compat: use email in favor of mimetools

verify: fix scope issues with del statement

compat: use 'key' argument instead of 'cmp' when sorting a list

compat: use open() instead of file() everywhere

compat: can't compare two values of unequal datatypes

compat: use // for integer division

tests: add -3 switch to run-tests.py

keyword: remove unused import

merge with crew-stable

help: add #revision syntax to the example valid URLs.

context: improve arg-checking assert.

help: add #revision syntax to the example valid URLs.

merge wtih crew-stable

gendoc: fix localization of help topic When a topic provides a callable method for its text, most likely this text will be generated from different parts, so it does not make sense to apply gettext on the whole result, rather the method should provide translation by itself. This is the case with the extensions topic, which triggers a double gettext call, making the ASCII codec fail when it encounters 8 bit characters, and prevents the documentation from being built.

extensions: remove dead code enabled used to be a boolean, and somehow that bit of code inadvertently slipped through during a refactoring. Effectively dead code, as the condition can never be triggered.

merge with crew-stable

acl: read correct index into url for username (issue298) The index was inadvertedly off-by-one causing the username to be the remote host rather than the remote user when hosted in a http(s) session.

color: don't highlight the summary part of the qseries command

mq: align columns in verbose qseries output.

extensions: remove unused imports

test python hook configured with python:[file]:[hook] syntax Adds a test for python hooks configured with python:[file]:[hook] syntax in .hgrc. Increases the test coverage for mercurial/hook from 92% (79/85) to 96% (82/85).

keyword: break overlong line

bisect: fix format specifiers for integers

Added signature for changeset 196d40e7c885

Added tag 1.3 for changeset 196d40e7c885

Merge with i18n

i18n-pt_BR: synchronized with 59bb11a97fe5

i18n-jp: translation update for 09c6735e3880

i18n-fr: translation of part of the mq extension

i18n-fr: translation of the zeroconf extension

i18n-fr: translation of the purge extension

i18n-fr: translation of the parentrevspec extension

i18n-fr: synchronized with 7d39c76be4ee

i18n-da: synchronized with 09c6735e3880

Merge with mpm

i18n-da: use "lap" instead of "rettelse"

i18n-da: translated many one-liners

subrepo: audit subrepo paths

keyword: make repo.commit use a custom commitctx wrapper This avoids forcing the dirstate of overwritten files to normal during a commit. Thanks to Dan Villiom Podlaski Christiansen for the idea of a "double wrapper", so other extensions can still wrap repo.commitctx safely.

commit: lose some pointless 'nothing changed' messages Move message from localrepo to commands. This drops confusing messages in subrepo, import, and rebase.

commit: report modified subrepos in commit editor

verify: demote warning about nullid in copy to note

help: add more pointers to commit --close-branch

branches: add --closed flag for consistency with heads

addremove: build lists of already added and removed files too (issue1696)

addremove: pass unknown and deleted to findrenames directly again

addremove: normalize some variable names

findrenames: drop unused argument defaults

zsh completion: update bisect for builtin syntax

zsh completion: add hg branch

zsh completion: add hg branches

zsh completion: basic merge support

zsh completion: disable defaults and verbose output when completing If ui.verbose is set to True in hgrc, commands used to provide completions would yield bad results. For example, qpush completion would include the patch numbers and statuses as returned by qunapplied -v, instead of just the patch names. Defaults are also disabled when completing to prevent similar issues when an option is set that changes a command's output.

Merge with mpm

zsh completion: include guarded patches in qdelete

Merge with i18n

i18n-pt_BR: synchronized with eac360045ba4

i18n-fr: translation of the extensions help topic

i18n-fr: translation of help screen strings

i18n-fr: translation of the patchbomb extension

i18n-fr: add a glossary and recommendations

i18n: add French translation file

i18n-jp: translation update for 288ba6d6c5c7 this revision consists of contribution by: Muraoka Yusuke<yusuke.muraoka@gmail.com> Tadashi Jokagi <elf@elf.no-ip.org> UENISHI Kota <kuenishi@gmail.com> Yuki KODAMA <endflow.net@gmail.com> FUJIWARA Katsunori <foozy@lares.dti.ne.jp>

i18n-da: translated 50% This change translates many of the strings from the convert, inotify, fetch, gpg, and transplant extensions.

convert: do not ask for translation of "%s %s"

i18n-da: translated cmdline help strings

i18n-da: translated log and some mq commands

i18n-da: change "arkiv" to "depot"

i18n-da: sync with 288ba6d6c5c7

i18n-zh: sync with 288ba6d6c5c7(2009-06-24 04:43:09+08:00)

extensions: catch OSError when hgext is not accessible (issue1708) Temporary workaround for issue1708: on win32 with py2exe, hgext is distributed inside a zipped file (which anyway does not contain the py files from which we ought to extract the documentation strings), which raises a WindowsError (subclasses OSError). This means that on such platforms the list of disabled extensions won't be available. Real fix postponed for after Mercurial 1.3.

color: wrap qdiff/qseries after all extensions are loaded

atom: use urlescape in filelog's file link

Relicense zsh completion script under zsh license. Steve agreed some time ago.

Merge with crew

Update test output for e67e5b60e55f

commands: improve bundle compression methods description

keyword: improve help for kwfiles - just use "files" instead of "filenames" (analogous to "hg status -h") - reference the extension help wrt pattern configuration Kudos to timeless for helpful suggestions.

keyword: lowercase status flags of untracked files in kwfile output Document the flags and their meanings in the command help (thanks to timeless for bringing this to my attention).

help: branch heads have no descendants, not no children

Branch heads should not include "heads" that are ancestors of other heads. For example, given 1 (branch a) -> 2 (branch b) -> 3 (branch a) I expect "hg heads a" to show only 3. Discovered by running hg heads HEAD on the mutt repo, where older clients committed default on top of HEAD.

inotify: server: remove wpath method Only called once. And can be simplified to an easy oneliner.

inotify: return version to client even when not matching Old clients are expecting a version string even in case of mismatch, otherwise they get stuck.

windows: fix use of undefined exception (issue1707) This fixes the implied reliance on pywin32 and the win32 module. This also fixes a regression in ae275ad46bd0 that made Mercurial unusable without pywin32.

keyword: improve help for kwfiles

highlight: remove author name from docstring I don't feel this should be part of the documentation. The name of the original author can of course still be found with 'hg log'.

mq: standardize qfinish cmdline help string

inotify: OPT -> OPTION in cmdline help string

hgk: standardize cmdline help strings

color: don't blow up if configured with unknown color (just warn).

bookmarks: Change references to "repo" by references to "self" (issue1611) Using "repo" instead of "self" inside bookmark_repo methods was causing a circular reference and, thus, a memory leak. It has been detected because the method bundlerepository.__del__ is never called, therefore leaving dangling uncompressed bundles inside .hg subdirectory.

run-tests: use os.path.realpath() to find hg's setup.py. Needed for running out-of-tree test suites without --with-hg/--local.

Make RPM spec in contrib more aligned with the one from Fedora This is mostly syntactical changes which make the real differences more obvious when looking at a diff. contrib/mergetools.rc is now also installed and enabled by default.

addremove: drop some silly variable assignments

ui.prompt: Show prompt and selection in non-interactive mode ui.prompt was completely silent in non-interactive mode, unless in verbose mode. It is fine that it chooses the default automatically, but it is confusing that the message and prompt shown interactively can't be found in scripted tests. The prompt and selection is now .write'ed instead of .note'ed.

keyword: make kwdemo more translation friendly

commands: wrap short descriptions in 'hg help' The code for wrapping a single line of text with a hanging indent was duplicated in commands and help -- it's now moved to a new function called wrap in util. The function defaults to a line width is 78 chars, and this un-wraps some command line flag descriptions, hence the test output changes.

commands: add "--" to command line flag in branches docstring

change wiki/bts URLs to point to new hostname

extensions: change descriptions for hook-providing extensions Thanks to timeless and Dave Townsend for advice!

extensions: change descriptions for extensions providing a few commands

help: rewrap extension description line length at 78 instead of 70

extensions: fix up description lines some more

commands: add note about import retrieving patches from URLs

windows: import WinIOError from win32 module (issue1707) Quick fix pending further refactoring of windows error handling.

mq: compact & rewrap documentation for qdelete command

test-globalopts: broken by output change in 62e3b9466700

Merge with i18n

i18n-pt_BR: synchronized with 3ad0b5ddae58

merged with crew

i18n-pt_BR: minor rewording and formatting fixes

i18n-pt_BR: synchronized with d9f4c182aeca

i18n-pt_BR: synchronized with 708938509732

i18n-pt_BR: synchronized with fccdf5ca5065

merged Greek translation

Add the start of a new translation of Mercurial to Greek (UTF-8)

i18n-da: sync with hg.pot - translated fuzzy strings - removed old unused strings

commands: removed redundant line from push docstring

merged with crew

i18n-da: merged translation with TortoiseHg

i18n-da: translate fuzzy strings

i18n-da: sync with hg.pot

i18n-da: add missing newline

translated a bunch of strings to danish

interhg: escape backslashes in docstring

interhg: remove unnecessary escaping

Fix warning: Seen unexpected token "%"

Add (pre)outgoing hooks for local clones.

contrib/buildrpm: force en_US locale during changelog's creation Running this script using a different locale (i.e., Italian) returns an error: $ ./contrib/buildrpm sto aggiornando la directory di lavoro 1060 file aggiornati, 0 file merge, 0 file rimossi, 0 file non risolti errore: data errata in %changelog: mar giu 23 2009 Stefano Tortarolo <stefano.tortarolo@gmail.com> 1.2.1-1034

contrib/buildrpm: set python requirement to 2.4

mq: no longer mention the deprecated qdelete's --revision option The option will no longer show up in the list, unless --verbose is used. 'qdel -r' has been deprecated in favor of qfinish since the 1.1 release. It is still fully functional, but new users need not learn about it.

help: describe bundle compression methods (issue1523)

Add a forget command for easily untracking files. This command does exactly what 'hg remove -Af [FILES]' does. The reason for creating a new command is that the options for 'hg remove' are confusing (-A removes only deleted files, -f forces deletion, and using both means *the exact opposite of both*). [mpm: simplified help text, code, and updated tests]

test merge-tools internal:dump and internal:prompt

hgrc.5: remove reference to DEFAULT section

hgrc.5: mark more literal strings

hgrc.5: markup compromis compatible with 7.1.2 and 8.4.5

test-grep: fixed comment

extensions: remove import rendered unnecessary by e4e22a310b62

help: fixing non-matching example texts

extensions: improve the consistency of synopses Trying as much as possible to consistently: - use a present tense predicate followed by a direct object - verb referring directly to the functionality provided (ie. not "add command that does this" but simple "do that") - keep simple and to the point, leaving details for the long help (width is tight, possibly even more so for translations) Thanks to timeless, Martin Geisler, Rafael Villar Burke, Dan Villiom Podlaski Christiansen and others for the helpful suggestions.

acl: help improvements Thanks to timeless for the review.

bookmarks: help improvements - bookmarks are obviously Mercurial's, no need to specify it - more explicit about the behavior without track.current - typos Note: the extension's help is shadowed by that of the command of the same name and as such will never appear to the user. Thanks to timeless for the fixes.

patch: use new style class in linereader

convert: fix builtin cvsps when no branchpoints are present

help: wrap extension descriptions

convert/svn: drop undocumented and untested revision 'blacklist'

convert/svn: test svn tags encoding

convert/svn: stop returning unicode revision identifiers

convert/svn: remove confusing unicode variable

convert/svn: remove dead code from entry deletion code path It was made obsolete when multi-branches commit were split into multiple Mercurial commits.

convert/svn: remove a useless recode() call in deletion path

convert/svn: remove useless encoding/decoding calls (issue1676)

convert/svn: remove dead code and obsolete comments

help: improve grammar/wording of 'extensions' topic

help: more improvements for the extensions topic - improve help text English (thanks to timeless for corrections) - rename and simplify functions a little bit, improved comments

extensions: use _order to determine if an extension is enabled

extensions: check for path existence only when necessary

extensions: remove now-useless pathdirs() function

mq: fix error message for qpush inexistent-patch (issue1702)

highlight: was broken since 580a79dde2a3 (encoding)

help: add/fix docstrings for a bunch of extensions

extensions: simplify by selecting primary hgext

extensions: move extensions listing functions from mercurial.help Includes some small fix-ups to comments in enabled() and disabled().

serve: obey the --encoding option

contrib/buildrpm: Don't require installed hg, use local hg with pure extensions

contrib/buildrpm: Don't use perl

contrib/buildrpm: Support python 2.4 and 2.6

help: remove per-extension paragraph on how to enable it The central place to find this information is the "extensions" help topic.

help: expand the extensions topic

help: refactor extensions listing, and show enabled ones in the dedicated topic

help: adding a new help topic about extensions Currently listing non-enabled extensions and a short introductory text. Thanks to Dan Villiom Podlaski Christiansen for the preliminary proof-of-concept code for listing available extensions.

bookmarks: add appropriate locking (issue1691) - make updates of bookmark state locked and atomic - wrap commit so that commit and bookmarks happen under the same lock

filemerge: fix internal:dump

hgweb: deny cloning a subpath of a repo

hgweb: web.encoding should override encoding.encoding (issue1183)

bdiff: fix compile with GCC -ansi (issue1690)

tags: silence warning about unknown tags This is mostly a nuisance and can happen legitimately with pull -r, etc.

rollback: minor clarification (issue828)

update: add --check option

tags: drop nested function

tags: generate contexts directly

tags: reverse and simplify head-walking

tags: simplify rev handling

tags: fold in _hgtagsnodes

compare grep result between target and its parent I found that typical case is that grep target is added at (*) revision in the tree shown below. +--- 1(*) --- 3 0 +--- 2 ------ 4 Now, I expect 'hg grep --all' to show only rev:1 which is first appearance of target line. But 'hg grep --all' will tell: target line dis-appeared at 3 => 4 target line appeared at 2 => 3 target line dis-appeared at 1 => 2 target line appeared at 0 => 1 because current 'hg grep' implementation compares not between target revision and its parent, but between neighbor revisions in walkthrough order. I checked performance of this patch by "hg grep --follow --all walkchangerevs" on whole Mercurial repo, and patched version could complete as fast as un-patched one.

url: let host port take precedence when connecting to HTTPS Fixes use of HTTPS connections on non-standard ports.

url: support client certificate files over HTTPS (issue643) This extends the httpshandler with the means to utilise the auth section to provide it with a PEM encoded certificate key file and certificate chain file. This works also with sites that both require client certificate authentication and basic or digest password authentication, although the latter situation may require the user to enter the PEM password multiple times.

acl: support for getting authenticated user from web server (issue298) Previously, the acl extension just read the current system user, which is fine for direct file system access and SSH, but will not work for HTTP(S) as that would return the web server process user identity rather than the authenticated user. An empty user is returned if the user is not authenticated.

hgweb: escape REMOTE_HOST when passing url for addchangegroup If DNS lookups are turned off on the web server, REMOTE_HOST may be populated with REMOTE_ADDR, which, if the remote is an IPv6 hosts will contain colons, thus interfering with the separator character. This is solved by URL quoting the REMOTE_HOST string.

tests: fix test for convert detection of p4 repos

kill trailing whitespace

graphmod/webcommands: use generic DAG walks Changes graph() to colorededges(), which operates on the new generic DAG walks and adds color and edge information needed by the web graph. This is in preparation of adding DAG walk filters, like the linear run collapser in the next patch. The idea is to have a bunch of changelog walkers that return basic data. Then we can filter this data. Finally we add edge and formatting info suitable for the output media we want to target (glog, hgweb).

graphmod: rename a bunch of vars in graph()

graphmod/graphlog: make dag walks carry data as type, payload

graphlog: rename grapher to asciiedges

graphlog: move and rename graphabledag to asciiformat

graphmod/graphlog: extract nodelistwalk

graphmod/graphlog: move log walks to graphmod

graphmod: code cleanup and doc fix

merge: rename -S/--show option to -P/--preview

mq: unify code for qdel -r and qfin

mq: warn about finalizing patches without cset message

filemerge: add internal:dump This create foo.{local,other,base} files for people to manually merge files while littering their working directory.

filemerge: add internal:prompt target

convert: Make P4 conversion cope with keywords, binary files and symbolic links. Convert now handles errors from p4 during conversion more gracefully. If keyword expansion is enabled in a P4 file then keywords will be unexpanded in hg. Added testcase for p4 filetypes and keyword (un)expansion. This testcase ignores UTF and Apple files to avoid binary data. Edited by pmezard: fixed collation issue on OSX

i18n-zh_TW: Traditional Chinese translation

patchbomb: update test output to match 2aff285b902f

patchbomb: do not assume the presence of angle brackets around msg-id RFC 5322 states: "Semantically, the angle bracket characters are not part of the msg-id; the msg-id is what is contained between the two angle bracket characters." Hence it should be correct to pass a message Id with no angle brackets to --in-reply-to. Adding them if missing.

keyword: concise help title

interhg: upgrade comments to online help and improve them Some inspiration taken from the wiki page of the extension.

churn: improve description Thanks to Martin Geisler and timeless for the suggested improvements.

doc: fix regexp for determining the man page section Preceding a .* pattern with an other dot serves little purpose, better anchor the regexp to the start of the file name.

Fix test-convert-cvsnt-mergepoints so it works reliably. Specifically, always run 'cvs commit' with -f option to force commit; add one strategic sleep which seems to be necessary for post-merge clobber-and-commit (-f doesn't force a commit there?).

test-convert-cvsnt-mergepoints: make it work on OS X. - don't create file 'cvs' where directory 'CVS' already exists

test-convert-cvsnt-mergepoints: ignore differences between CVS 1.11 and 1.12. - factor out cvsci function (similar to other test-convert-cvs* scripts) - add filterpath function (also similar to other scripts) - generally munge the output of CVS - add lots of output to make it easier to follow when things go wrong This doesn't make the test pass reliably under CVS 1.11; it just makes it behave the same as under CVS 1.12, i.e. sometimes it passes and sometimes it fails. Failure is more frequent with faster hardware.

Rename CVSNT "mergepoints" test and make it executable. - rename test-convert-cvs-builtincvsps-cvsnt-mergepoints (and related files) to test-convert-cvsnt-mergepoints - this ensures that the test will be run, but does NOT make it pass: in particularly, it fails regularly for me due to the inconsistent behaviour of CVS itself - expect "Branchpoints:" in debugcvsps output

Fixed patch.eol bug that truncated all patched lines to one character Also updated the import-eol test to test with lines longer than one character, and also empty lines.

subrepo: add a test

subrepo: basic push support

subrepo: add update/merge logic

commit: recurse into subrepositories

subrepo: introduce basic state parsing

Make mq, record and transplant honor patch.eol

Add patch.eol to ignore EOLs when patching (issue1019) The intent is to fix many issues involving patching when win32ext is enabled. With win32ext, the working directory and repository files EOLs are not the same which means that patches made on a non-win32ext host do not apply cleanly because of EOLs discrepancies. A theorically correct approach would be transform either the patched file or the patch content with the encoding/decoding filters used by win32ext. This solution is tricky to implement and invasive, instead we prefer to address the win32ext case, by offering a way to ignore input EOLs when patching and rewriting them when saving the patched result.

gpg: add test

gpg: unbreak 02a12835b17e. We need a test.

share: allow dest to default to the basename of source

Merge with crew

bisect: improve --command output - Display tested revisions without --verbose - Display revision number

dirstate: fix typo introduced by 3507f6c7715c

test-bisect: test --command option

bisect: fix traceback when command executable is not found

share: add experimental share extension

add helper function to create shared repos

repo: add internal support for sharing store directories set .hg/sharedpath to point to the .hg to share with

remoteui: properly create dst with copy()

repo: set up ui and extensions earlier

localrepo: remove 'closed' argument to heads(...) function - repository heads are not associated with the closed attribute, so remove it making the code in line with the concept. - Fix functions that were calling heads with the parameter. - Adjust webcommands.branches to include the concept of inactive as well as open and closed branches - Fix code and docstrings in commands to make the correct use of closed branches & branch heads clearer - Improve grammar of 'hg heads' help text (2nd submission) this does not alter the cli for hg branches, that work is still to be done

mq: eliminate warning on qpush with empty series Warning was "DeprecationWarning: use lock.release instead of del lock"

inotify: repowatcher: don't use a watches attribute to count watches The encapsulated watcher alread has a __len__ method for this purpose. Also, repowatcher.watches was never decremented on watch deletions.

inotify: put the "while True: poll()" loop in pollable class

inotify: refactor (un)register methods into pollable object repowatcher.master becomes unnecessary

inotify.server: the decorator eventaction() shouldn't be a method of repowatcher

inotify: add log config option redirect inotify server output to a file

cmdutil: service: logfile option to redirect stdout & stderr in a file

inotify: raise QueryFailed when the server crash When the server crashes, it does not write back on the socket. The message which is then read from the socket is '', which raises a TypeError. Catching TypeError to raise QueryFailed instead.

inotify: server: move split() out of server split() has nothing to do with the server logic, it does not need to be in the class. Move it on top, next to join() which does the opposite.

inotify: remove unused variables

atomictempfile: fix exception in __del__ if mktempcopy fails (self._fp is None) Make what is going on more obvious by explicitely getting the 'closed' attribute from _fp, instead of from the gettatr proxy.

Merge with crew-stable

bzr convert: restore paths to source encoding. Closes issue1692.

doc: wrap man pages after fixing quoting

hgrc.5: fix quoting of literal strings The quotes would go wrong in many places due to differences between asciidoc version 8.2.7 used by Benoit and 8.4.5 used by me. Between those versions asciidoc stopped interpreting the content of `quoted strings`, and so `*` would start bold text in the old version, but do nothing in the new version. To complicate things further, `\*` would escape the bold tag in the old version, but in the new version the backslash was inserted literally into the output (because the backtick quotes it). I've now replaced backticks with non-quoting plusses and escaped backslashes as appropriate.

hgignore.5: remove extra quotes The so-called "unconstrained quotes" didn't really work, so I reverted them to normal quotes. Tested with asciidoc version 8.2.7 and 8.4.5.

Improve English for help text of many core hg commands. co-author: Greg Ward <greg-hg@gerg.ca>

use new style classes

doc: correctly escape *

doc: correctly escape *

hg.1: highlight references to commands and man pages

hg.1: simplify synopsis Now that arguments can be mixed with options we can simplify the synopsis considerably. Also, highlighting the command name in bold (instead of italics) seem to be the standard convention.

hgrc.5: spell check

man pages: add meta data This removes the fixme's that were otherwise inserted.

man pages: updated copyright years

hgrc.5: use NOTE admonition paragraph

hgrc.5: reformatted to avoid big chunks of monospaced text The man page used to have lines longer than 80 characters, even though all lines in the hgrc.5.txt file were wrapped nicely. The problem turned out to be that the indented paragraphs started literal blocks instead of normal paragraphs. The literal blocks were of course not wrapped when displayed by man. In short, the asciidoc rules require lists to be formatted like this: foo:: Some description of foo. + Another paragraph in the description of foo. It *must* start flush left and the plus is necessary to indicate that this is a list item continuation. Lists with nested lists can be formatted correctly using something called "open blocks". These blocks are used to group the list items and are marked by a line above and below with two dashes. See the asciidoc user guide for the gory details...

hgrc.5: mark filename patterns as monospaced text

hgrc.5: single quotes means italic in asciidoc Use double quotes instead and mark the * as monospaced text.

man pages: mark filenames and env vars as fixed text

test-convert-svn-*: no longer need the url building code

convert: default to file protocol when no :// found for svn repo url Edited by pmezard: add path separator normalization

keyword: improve English

bookmarks: improve English

Generally replace "file name" with "filename" in help and comments.

Spell Mercurial as a proper noun

win32text: drop disabling portion of test disabling extension leaves dirstate confused about whether affected files are clean or not, not much point testing it.

i18n-de: Synchronised with hg.pot translated by Fabian Kreutz <fabian.kreutz@qvantel.com> Tobias Bell <tobias.bell@gmail.com>

i18n-pt_BR: synchronized and updated - synchronized with hg 6019e6517f95 - translated remaining messages - many spelling and formatting fixes

convert: better support for CVS branchpoints (issue1447) This records the branches starting at individual CVS file revisions, using the symbolic names map rather than just the branches information. This information is used to generate Mercurial changesets. Despite the changes, the CVS conversion still suffers heavily from cvsps' deficiencies in generating a correct representation of the CVS repository history.

Merge with stable

Merge with crew

merge: refactor manifestmerge init to better report effective ancestor

merge: simplify file revision comparison logic

merge: make locally-added file test more correct

merge: drop unused diverge initialization

merge: refactor some initialization, drop backwards var

merge: combine a copy and move case

merge: drop a flag update case

merge: combine three identical 'remote is newer' cases

merge: drop 'remote deleted' case

merge: simplify 'other deleted' case

merge: simplify backwards revert test, ready to combine with other cases

merge: allow merging going backwards New behavior is generally superior and more correct, except possibly with regards to missing files. hg up . is now effectively a no-op, which is probably the desired behavior for people expecting to move to tip, but may surprise people who were expecting deleted files to reappear. case 1: update to . a-w -> a-w classic: ancestor a missing recreated right? rmed recreated WRONG added forgotten WRONG changed preserved RIGHT conflicted can't happen backward merge: ancestor a (NO EFFECT) missing missing wrong? rm'ed rm'ed RIGHT added preserved RIGHT changed preserved RIGHT conflicted can't happen case 2: update to ancestor of . a-b-w -> b-w \ a classic: ancestor a missing recreated right? rmed recreated wrong? added forgotten wrong? changed preserved RIGHT conflicted preserved wrong? backwards merge: ancestor b missing missing or conflict right? rm'ed missing or conflict right? changed preserved RIGHT conflicted merge RIGHT added preserved right?

merge: reorder remote creation tests

merge: drop recreating case

merge: drop an overwrite test

merge: simplify a delete case

merge: reorder get cases for future simplification

merge: fix prompt keep

merge: remove a flags case

merge: pretend ancestor is local to simplify backwards and overwrite

merge: simplify flag merging code slightly

dirstate: eliminate reference cycle from normalize Bound methods hold a reference to self, so assigning a bound method to an instance unavoidably creates a cycle. Work around this by choosing a normalize method at walk time instead. Eliminate default arg while we're at it.

cleanup: removed unused imports

hgrc.5: wrap lines at 70 chars (whitespace cleanup)

hgignore.5: wrap lines at 70 chars (whitespace cleanup)

hg.1: wrap lines at 70 chars (whitespace cleanup)

hgrc.5: language fixes

hgrc.5: wrap manpage at 80 chars Checked with asciidoc 8.4.5

bookmarks: update docstring

Unbreak run-tests support for out-of-tree extensions

Merge with crew-stable

Some platforms lack the PATH_MAX definition (eg. GNU/Hurd). Thanks to ronny for making it cleaner.

test-fncache: Use find instead of ls -R. Different platforms implement -R differently (and it produces unneccessarily verbose output in this case). find is just as good and more consistent. Unbreaks test on OpenBSD. Edited by pmezard: added 'sort' call

Make test-verify SUSv3 compatible. 'head -c' isn't supported on (at least) OpenBSD and is not part of SUSv3. Instead, use dd. See also: http://www.opengroup.org/onlinepubs/009695399/utilities/head.html

Support branch names in contrib/bash_completion Anywhere a revision is needed a branch name can be used to specify the tip of that branch, so branch names should be tab completed.

Support merge in contrib/bash_completion

context: add a dirty method to detect modified contexts

dirstate: don't complain about 0-length files

commit: move some setup outside the lock

Add cp950 as problematic encoding which is used in chinese windows.

hgweb: allow distinction between open/closed branches on branches page

commit: rename wctx to cctx

mq: fix commit prototype

commit: trade O(n^2) file checks for O(n^2) dir checks

commit: move explicit file checking into repo.commit

dirstate: more accurate use of match.dir callback

commit: editor reads file lists from provided context

commit: drop the now-unused files parameter

tag: use match.exact for commit

commands: drop files arg for commit calls

transplant: use match object rather than files for commit

rebase: remove unneeded status call before commit

gpg: use match.exact rather than files for commit

mq: drop unneeded files args for commit

commit: apply force flag without files

fetch: no need to pass files list to commit

templater: retain author's full name if no email is supplied (issue1685)

tests: delete parallel coverage files after combining

heads: show closed heads only when --closed is passed Add a --closed (-c) option to 'hg heads' to show all heads and change the default behavior to refrain from showing fully closed branches. Enhance 'hg heads <branch>' so that: * default: displays normal & inactive heads, not closed heads * --closed: displays normal, inactive & closed heads * --active: displays only normal heads * both --closed and --active: displays normal & closed heads only

localrepo: set heads and branchheads to be closed=False by default The heads(...) and branchheads(...) functions will now only return closed heads when explicitly asked for them. This will cause 'hg merge' to have better behavior in the presence of a branch that has closed heads when no explicit rev is passed.

convert: rewrite tags when converting from hg to hg

convert: default revisions order depends on source When converting Mercurial repositories you expect the revision numbers to be preserved, while other sources conversions focus on efficiency.

convert: fail fast if source does not support --sourcesort

convert: add --sourcesort option for source specific sort Only supported by Mercurial source for now.

convert: parse sort mode sooner

convert: split toposort() into subfunctions for readability

run-tests: don't replace PYTHONPATH, just augment it. (Needed at least for Subversion bindings on OS X, which are in /opt/subversion. Useful for other external libraries installed in non-standard places too.)

merge with mpm

color: yet another typo

match: fold plan cases down to two special cases - merge always and match with patterns - make always and match with patterns the default - invert dostep3 to skipstep3 - move dirignore test inside exact case

walk: refactor walk plan - never is gone - reorder tests more cleanly - rename nomatches to exact for clearer semantics

match: remove match.never Only one user, can be translated to match.exact()

walk: use match.bad callback for filetype messages

match: ignore return of match.bad All users returned false, return can now be dropped

add: use match.bad callback more effectively

match: document bad callback semantics

walk: simplify logic for badfn clause - matchfn redundant - call badfn always rather than fwarn - use for/else rather than keep var

walk: we always have a badfn

walk: simplify check for missing file

run-tests: redefine --with-hg so it takes the 'hg' script to run. - in parseargs(), check that --with-hg value is valid - add handy --local option for "--with-hg=../hg" - ensure that we always set PATH and PYTHONPATH (not just when doing a temporary install) - override any existing PYTHONPATH, so test success does not depend on whatever happens to be in the caller's environment - give tests a little more control by exporting $PYTHON to the environment; needed by test-convert and test-mergetool when they run hg with a stripped-down $PATH Also, add a big comment explaining all the corner cases to test for the next person who tries to modify this script.

run-tests: fix _checkhglib() so it's correct when using --with-hg. Since this makes the warning less likely, upgrade it to a "real" warning to stderr.

run-tests: factor out _checkhglib() to check import path of 'mercurial'. - rename _hgpath() to _gethgpath() and move it down next to _checkhglib() - change _gethgpath() so it caches its result: replaces global 'hgpkg' that anyone can use with global '_hgpath' that is only for _gethgpath() - no need to pass 'expecthg' to runchildren() or runtests() anymore Also: do not change global PYTHON when doing coverage; this seems to be unnecessary, since we create a dummy 'hg' script that handles coverage for us, and it made things brittle. (E.g. the rest of this patch makes the call to _hgpath() come later, and it was broken by enabling coverage.)

run-tests: show PID if running in parallel mode with -v. (Extremely handy when you break parallel operation and need to figure out what's going on. The assumption is that if you run with -v, you want more details.)

i18n-da: synchronized and updated slightly

commands: typo in bundle abort message This error trigger if one calls bundle with the wrong parameters and it is thus not an error scripts will want to look for (they could and should ensure that they call bundle with the correct parameters).

fixed typos found in translatable strings This is from a spell-check of hg.pot.

graphlog, win32mbcs: capitalize ASCII

highlight: consistently capitalize Pygments

win32mbcs: capitalize Unicode

commands: expand "arg" -> "argument" in showconfig help strings It is easier to translate full sentences without abbreviations, or said differently, I don't know a short Danish word for "argument".

wrap string literals in error messages

convert/gnuarch: wrap long line, format kwargs without spaces

convert/cvsps: wrap long lines

convert/subversion: wrap long lines in comments

test-hybridencode: break long string literals

revlog: make triple-quoted string a real comment

posix: do not use fstat in isowner The fstat function was undefined, but never used since a stat object was always passed in the optional st argument. Passing st is now mandatory. This bug crept in when util was split up into posix and windows modules. The fstat function is still defined in util, but importing it into posix would create an import cycle which seems unnecessary.

removed unused imports

Move alias into core

mq: rename setheader to updateheader and fix comment

mq: initializing patchheader class directly from patch content The patch header needs only be read in order to instantiate the class, and as such it makes more sense to do it within the class.

Merge with mpm

run-tests: clarify timeout log message.

revlog: refactor chunk cache interface again - chunk to _chunk - _prime to _chunkraw - _chunkclear for cache clearing - _chunk calls _chunkraw - clean up _prime a bit - simplify users in revision and checkinlinesize - drop file descriptor passing (we're better off opening fds lazily

clone: fall back to pull source repo cannot be locked, 937ee88da3ef was a noop 937ee88da3ef was a noop, it was missing the wait keyword

fixed 0xc0150004 error building Mercurial under Python 2.6 for Windows

localrepo: move comment

localrepo: use lock.release for single lock

dirstate: fixed typo in comment

changelog: make delayopener less intrusive

revlog: report indexfile rather than datafile for integrity check

ui: fix two bugs in %% warning

revlog: move stat inside lazyparser

dirstate: notice truncated parents read

lookup: check for dirstate damage on failure

keyword: rename matcher() to match() mimicking changes in main

tests: test for dispatch on [defaults]: more clearly differing output Using '-r null' instead of '-v' as the overriden command default. The latter did not have any effect on output, thus not giving much indication on whether the modified defaults were really in use or not.

merge: whitespace cleanup

revlog: add test for 7659eecd9da2 (issue1678)

changegroup: the node list might be an empty generator (fix issue1678)

store encoding: .i/.d encoding for non-store repo (broken by 810387f59696)

replace "i in range(len(xs))" with "i, x in enumerate(xs)" The remaining occurrences should be the ones where "xs" is mutated or where "i" is used for index arithmetic.

util: simplify range expression The n index variable was unused. Every iteration would pop one element off of parts, so the for loop can be replaced with a while loop.

color: cleanup extra commas

setup: execute hg in C locale

setup: handle trust warnings when determining version

setup: renamed l, e vars to out, err

hggettext: ensure correct Mercurial is imported

Makefile: do not create i18n/, it is already there

replace xrange(0, n) with xrange(n)

color: replace re.split with ui.configlist

color: use lists instead of tuples for effects

hgweb: extract config values after reading webdir-config

tests: make coverage run in parallel mode, clean up coverage code

revlog: fix undefined variable introduced in 5726bb290bfe

parentrevspec: remove a trailing colon

purge: fix spelling error

zsh-comp: explain how to use for non-global install

use ui instead of repo.ui when the former is in scope

cmdutils: Take over glob expansion duties from util

match: fix _patsplit breakage with drive letters

statichttprepo: handle remote not supporting Range headers - If remote does not support Range header, 200 is answered instead of 206. The HTTPRangeHandler left these responses unchanged, so the data has to be sliced by the receiver. - httprangereader file pointer was not updated.

convert: better feedback when filtering out empty revisions Original patch by Herbert Griebel <herbertg@gmx.at>

inotify: server: use a common 'pollable' interface for server & repowatcher Mainly for documentation purposes: it easily explains the role of handle_event and handle_timeout, and why both server & repowatcher implement those methods.

inotify: process all inotify events in one batch When several inotify events happen, we don't have to process each event separately, calling everytime repowatcher.read_events() to fetch events from the underlying watcher: it is sufficient to call once read_events, to fetch all the events from the watcher.

inotify: rename handle_event to handle_pollevent to avoid confusion event here refers to poll events, and are different from events read in server.read_events for example, where those events are inotify events.

inotify: handle_event: do not use event and fd parameters. event is particularly confusing given the context (is it an inotify event? a polling event?) and is never used. Remove it. fd has very little use, and it gives the false impression that event handling depends on fd. It's wrong: the same behavior is triggered, for all events.

inotify: use a decorator instead of dispatching calls

inotify: do not defer inotify events processing Doing a part of the event processing and deferring the rest is a bad habit: it complexifies the code, and it does not respect event ordering! Moreover, there is already a timeout handling, so that inotify events are only processed when a treshold is exceeded: there is no requirement to delay anymore the events processing.

inotify: do not recurse in handle_timeout(): call it explicitely, not in scan() When in handle_timeout, scan() is called when a repertory is created/modified. But the first line of scan calls handle_timeout. This had the consequence of calling recursively handle_timeout: * several calls to read_events (but only the first one retrieves events) * every time that an event is queued for a deferred action, the next time that scan() is called, handle_timeout is called, the event queue is treated, even if all the events haven't been read/queued yet. This could lead to inconsistencies

i18n-da: typo

merge with crew

inotify: adding test for issue1556

inotify: proper fix for issue1542 (partially reverting 67e59a9886d5) issue1542 description: Unknown files (?) placed in a directory are still marked as present and unknown when the containing directory is moved out of the repository scope. Why 67e59a9886d5 was bad: * When the problem we're addressing only deals with unknown files, the fix to updatestatus applies for all statuses * The only reason to move the call schedule_work(wpath, 'd') seems to be that it allowed an updatestatus call on the deleted directory, in deleted(). But deleted() should not be called on directories in the first place. * After fixing an independant issue (1371), test-inotify-issue1542 was failing Fix: When processing a deletion of a directory, walk the tree of the unknown files and remove the entries from repowatcher. This step does not need to be added in the generic scan() routine: it is only necessary on a directory deletion.

inotify: server: refactor updatestatus() * Instead of one entry point, use two entry points, updatefile() and deletefile(), both internally calling the helper function _updatestatus * Do not rely on TypeError to detect the type of oldstatus: use isinstance * The call updatestatus(wpath, None) in deleted() was a bit particular: because no osstat and no newstatus was given, the newstatus was determined using the data stored internally. To replace this exact behavior with the new code, one would use: root, fn = self.split(wpath) d = self.dir(self.tree, root) self.filedeleted(wpath, d.get(fn)) This, however, duplicates code with _updatestatus(), which led us to an interesting question: why are we basing ourselves on repowatcher data to update the status, where everywhere else, we are comparing against dirsate? There is no reason to do this, which is why the new code is: self.filedeleted(wpath, self.repo.dirstate[wpath]) Incidentally, after this, the test for issue1371 passes again.

convert: deprecate external cvsps, to be removed in 1.4

osutil: silence uninitialized variable warning

convert/hg: update documentation

i18n-da: better word for "check"

i18n-da: fixed use of plural form for parent

i18n-da: better Danish word for "entry"

runtest: do not start testing when there is no test

templatefilters: add filter to convert date to local date (issue1674) Issue1674 suggests the localdate filter be applied as default to log, but this patch only introduces the filter, not the changed default behaviour.

url: use CONNECT for HTTPS connections through HTTP proxy (issue967) urllib2 and httplib does not support using CONNECT proxy requests, but only regular requests over the proxy. This does not work with HTTPS requests as they typically require that the client issues a CONNECT to the proxy to give a direct connection to the remote HTTPS server. This is solved by duplicating some of the httplib functionality and tying it together with the keepalive library such that a HTTPS connection that need to be proxied can be proxied by letting a connection be established to the proxy server and then subsequently performing the normal request to the specified server through the proxy server. As it stands, the code also purports to support HTTPS proxies, i.e. proxies that you connect to using SSL. These are extremely rare and nothing is done to ensure that CONNECT requests can be made to these as that would require multiple SSL handshakes. This use case is also not supported by most other contemporary web tools like curl and Firefox3.

dirstate: skip step 3 in walk if nothing new will match nothing will ever match on match.never nothing new will match on match.exact (all found in step 1) nothing new will match on match.match when there is no pattern and there is no direcory in pats

dirstate: set more states in step 1 of walk

match: fold match into _match base class

match: add exact flag to match() to unify all match forms

match: redefine always and never in terms of match and exact

match: fold _globprefix into _roots

match: optimize escaping in _globre - localize re.escape - fastpath escaping of non-special characters

match: remove head and tail args from _globre

match: fold _matcher into match.__init__

match: rename _matchfn to _buildmatch

match: optimize _patsplit

match: tweak some names

match: simplify _matcher - get rid of special case - simplify anypats logic - fold inckinds and exckinds

match: split up _normalizepats

match: optimize _globprefix

match: unnest functions in _matcher

match: kill unused defaults on _globre

match: kill test in matchfn

match: refactor matchfn generation

match: move util match functions over

util: privatize globre

match: refactor patkind add patkind(pat) to match change external users change util.patkind to _patsplit

match: add some default args

match: change all users of util.matcher to match.match

tests: add tests for new pre-push logic (issue736) Co-contributor: Henrik Stuart <henrik.stuart@edlund.dk>

named branches: improve pre-push logic (issue736) Each named branch is considered separately, and the push is allowed if no new branch heads are created for any named branch to be pushed. Due to some tests's use of --debug, their output will change after this addition. This has been fixed as well. Co-contributor: Henrik Stuart <henrik.stuart@edlund.dk>

named branches: client branchmap wire protocol support (issue736) Co-contributor: Sune Foldager <cryo@cyanite.org>

named branches: server branchmap wire protocol support (issue736) The repository command, 'branchmap', returns a dictionary, branchname -> [branchheads], and will be implemented for localrepo, httprepo and sshrepo. The following wire format is used for returning data: branchname1 branch1head2 branch1head2 ... branchname2 ... ... Branch names are URL encoded to escape white space, and branch heads are sent as hex encoded node ids. All branches and all their heads are sent. The background and motivation for this command is the desire for a richer named branch semantics when pushing changesets. The details are explained in the original proposal which is included below. 1. BACKGROUND The algorithm currently implemented in Mercurial only considers the graph theoretical heads when determining whether new heads are created, rather than using the branch heads as a count (the algorithm considers a branch head effectively closed when it is merged into another branch or a new named branch is started from that point onward). Our particular problem with the algorithm is that we'd like to see the following case working without forcing a push: Upsteam has: (0:dev) ---- (1:dev) \ `--- (2:stable) Someone merges stable into dev: (0:dev) ---- (1:dev) ------(3:dev) \ / `--- (2:stable) --------´ This can be pushed without --force (as it should). Now someone else does some coding on stable (a bug fix, say): (0:dev) ---- (1:dev) ------(3:dev) \ / `--- (2:stable) ---------´---------(4:stable) This time we need --force to push. We allow this to be pushed without using --force by getting all the remote branch heads (by extending the wire protocol with a new function). We would, furthermore, also prefer if it is impossible to push a new branch without --force (or a later --newbranch option so --force isn't shoe-horned into too many disparate functions, if need be), except of course in the case where the remote repository is empty. This is what our patches accomplish. 2. ALTERNATIVES We have, of course, considered some alternatives to reconstructing enough information to decide whether we are creating new remote branch heads, before we added the new wire protocol command. 2.1. LOOKUP ON REMOTE The main alternative is to use the information from remote.heads() and remote.lookup() to try to reconstruct enough graph information to decide whether we are creating new heads. This is not adequate as illustrated below. Remember that each lookup is typically a request-response pair over SSH or HTTP(S). If we have a simple repository at the remote end like this: (0:dev) ---- (1:dev) ---- (3:stable) \ `--- (2:dev) then remote.heads() will yield [2, 3]. Assume we have nodes [0, 1, 2] locally and want to create a new node, 4:dev, as a descendant from (1:dev), which should be OK as 1:dev is a branch head. If we do remote.lookup('dev') we will get [2]. Thus, we can get information about whether a branch exists on the remote server or not, but this does not solve our problem of figuring out whether we are creating new heads or not. Pushing 4:dev ought to be OK, since after the push, we still only have two heads on branch a. Using remote.lookup() and remote.heads() is thus not adequate to consistently decide whether we are creating new remote heads (e.g. in this situation the latter would never return 1:dev). 2.2. USING INCOMING TO RECONSTRUCT THE GRAPH An alternative would be to use information equivalent to hg incoming to get the full remote graph in addition to the local graph. To do this, we would have to get a changegroup(subset) bundle representing the remote end (which may be a substantial amount of data), getting the branch heads from an instantiated bundlerepository, deleting the bundle, and finally, we can compute the prepush logic. While this is backwards compatible, it will cause a possibly substantial slowdown of the push command as it first needs to pull in all changes. 3. FURTHER ARGUMENTS IN FAVOUR OF THE BRANCHMAP WIRE-PROTOCOL EXTENSION Currently, the commands incoming and pull, work based on the tip of a given branch if used with "-r branchname", making it hard to get all revisions of a certain branch only (if it has multiple heads). This can be solved by requesting the remote's branchheads and letting the revisions to be used with the command be these heads. This can be done by extending the commands with a new option, e.g.: hg pull -b branchname which will be turned into the equivalent of: hg pull -r branchhead1 -r branchhead2 -r branchhead3 We have a simple follow-up patch that can do this ready as well (although not submitted yet as it is pending the acceptance of the branch patch). 4. WRAP-UP We generally find that the branchmap wire protocol extension can provide better named branch support to Mercurial. Currently, some things, like the initial push scenario in this mail, are fairly counter-intuitive, and the more often you have to force push, the more it is likely you will get a lot of spurious and unnecessary merge nodes. Also, restricting incoming and pull to all changes on a branch rather than changes on the tip-most head would be a sensible extension to making named branches a first class citizen in Mercurial. Currently, named branches sometimes feel like a late-coming unwanted step-child. We have run it in a production environment for a while, with fewer multiple heads occurring in our repositories and fewer confused users as a result. Also, it fixes the long-standing issue 736. Co-contributor: Sune Foldager <cryo@cyanite.org>

Merge with -stable

Merge with crew-stable

windows: make openhardlinks work Despite the intention, openhardlinks would always evaluate to False.

revlog: fix reading of larger revlog indices on Windows

inotify: Removing the unnecessary "inotifyserver" class variable.

inotify: set a flag so a failed inotify query doesn't get repeated. If, for some reason, we can't get the inotify server to start, it's better to disable inotify queries for the instance to avoid trying over and over to start the server, which takes time. Just fall back on repo.status()

inotify: introduce debuginotify, which lists which paths are under watch

inotify: put STAT-specific query answer generation part in its own method

inotify: change protocol so that different query types can be supported.

inotify: Separate query sending logic from Server starting. Use a decorator around the public statusquery method of Client: start_server(query_to_server): try: query_to_server() except QueryFailed: [error recovery, inotify Server (re)starting] query_to_server() This way, introducing a new xxxquery Client method is easy: one has only to code the protocol part of xxxquery, ignoring errors, and decorating it using start_server to handle server recovery and (re)starts

inotify: modular architecture for inotify clients Put the socket init, query generation and response analysis in a more generic client class.

setup: another newline if version could not be established.

setup: add missing newline in warning

setup: read .hg_archival.txt for version info (issue1670) Previously, setup.py was enhanced to identify the Mercurial version from either .hg/ or mercurial/__version__.py. When archives are created using 'hg archive' or via hgweb, neither of those options are available. However, there is a .hg_archival.txt file in the root of the archive that has the information. This patch enhances setup.py to identify the Mercurial version from the .hg_archival.txt file when there is no .hg/ or mercurial/__version__.py available.

setup.py: subprocess instead of os.popen, sys.stderr.write instead of print Also check that .hg is a directory.

gendoc: add missing space in command synopsis

merge: give hint as to how to discover uncommitted changes Many users will try 'hg diff' here, and it will not show them missing files.

commands: standardize parents command meta data

i18n-da: synchronized

i18n: accurately generate hg.pot

hgcia, convert: escape backslashes in docstrings

commands: remove unnecessary quotes in backout help

test-merge-tools: windows fixes - /bin/cat does not exist, expect 'cat' - non-interactive mode is not correctly detected in pysh, force it

ui: honor interactive=off even if isatty()

Merge with crew-stable

hgwebdir: fix [collections] evaluation under Windows Virtual and real path separators are not the same under Windows.

cmdutil: mark string for translation

util: use "is" for True/False/None comparisons

remove: warn if unversionned files are specified (issue1454)

fetch: allow -r for remote repos

filelog encoding: move the encoding/decoding into store the escaping of directories ending with .i or .d doesn't really belong to filelog. we put the encoding/decoding in store instead, for backwards compat, streamclone and the fncache file format still uses the partially encoded filenames.

store: refactor the fncache handling put all the fncache file writing and reading in the same place.

hgweb: make hgwebdir handle dict/list paths the same as config paths Before this patch, the only way to get hgwebdir to honor the recursive paths was to create a config object or a config file with the recursive paths in it. This patch makes hgwebdir treat paths the same whether passed in as a list, tuple, config or however hgwebdir supports passing paths.

workingfilectx: always use the same filelog, even for renames workingfilectx() was using the "src" filelog in case the file was renamed in the working copy. For consistency, stop special-casing it. This allows us to remove some duplication between filectx and workingfilectx.

use 'x is None' instead of 'x == None' The built-in None object is a singleton and it is therefore safe to compare memory addresses with is. It is also faster, how much depends on the object being compared. For a simple type like str I get: | s = "foo" | s = None ----------+-----------+---------- s == None | 0.25 usec | 0.21 usec s is None | 0.17 usec | 0.17 usec

patch: simplify Boolean expression slightly The context variable is either True, False or None. Abbreviate it as C and we get the following truth table where the second column is the original expression and the third column is the new expression: C | C or C == None | C is not False True | True | True False | False | False None | True | True

mq: new method invalidate Introduce method invalidate (similar to dirstate and repo).

mq: only read files when needed use util.propertycache as in dirstate

tests: replace #...# syntax with {...}

match: use self.exact instead of lambda self.exact uses a set and does not need an extra copy of the files

dirstate: use quicker matchfn() instead of match() everywhere matchfn does the same as match(), but a direct "link"

patchbomb: don't localize Date headers

alias: a0104303f400 did not correctly handle whitespace in the args

update --clean: do not unlink added files (issue575)

templater: lowercase error message Changing this messages should be safe: automated scripts ought to have debugged their templates and wont grep for this error message.

util: use set instead of dict

localrepo: update commit*() docstrings

patchbomb: with --in-reply-to, still thread message under first in series When specifying --in-reply-to for a message M, have [M] [0/2] [1/2] [2/2] instead of [M] [0/2] [1/2] [2/2] which is more consistent with the way messages are being threaded when --in-reply-to is not used.

merge with mpm

patchbomb: minor typo and language fixes

patchbomb: clarify diffstat usage

patchbomb: fix inexact description of the messages threading Each message refers only to the first of the series.

coal/paper: minimize whitespace differences in CSS files

coal/paper: equal heights for file and annotate views

spartan: ensure equal heights for file and annotate views The display used to jump up and down when switching from the file view to the annotate view. It now stays still: when switching from the file view to the annotate view it looks like the file content is simply right-shifted to make room for the author names.

coal/paper: simplify margin and padding CSS rules

spartan: simplify CSS colors (#xxxxxx -> #xxx)

coal/paper/spartan: remove extra line break in annotate view

commit: tidy up mergestate slightly

commit: drop unneeded dirstate invalidate logic We no longer touch the dirstate in the middle of a commit, so a failing commit doesn't require invalidating the dirstate.

commit: some tidying - simplify handling of 'close' - kill silly wlock=None - sort/uniq files later

commit: remove unused lock var

commit: move description trimming into changelog

commit: simplify manifest commit

commit: explain how to abort commit in editor

commit: move editor outside transaction The commit editor is now invoked before files and manifest are committed. The editor is now run with only the wlock held and aborting an edit no longer requires rolling back a transaction. Changes to files during a commit still result in undefined behavior. (This is preliminary work for committing subrepositories)

convert: directly use nullid

Merge with stable

Fix how setup.py identifies the Mercurial version. There is a problem with setup.py where it will not identify the Mercurial version properly when not being ran in within a repository even if mercurial/__version__.py exists. To fix, use mercurial.__version__.version when available before defaulting to "unknown". (Using mercurial.util.version() is not an option due to a dependency issue where osutil can be referenced before it is built.)

templater: replace eval with closure

win32mbcs: remove silly eval

merge with crew

addremove/findrenames: find renames according to the match object (issue1527) Instead of only finding similarities in the added/removed files found by the addremove step, follow the match object: hg addremove -s80 foo -> add and removes files in foo + find similarities between files in foo hg addremove -s80 -> add and removes files in the whole repo + find similarities between files in the whole repo hg import --similarity will still work correctly (only find similarities between files found in the patch).

addremove: mapping isn't really needed, simplify

gitweb, monoblue: break long line in map file

coal, paper: add newlines when displaying a file in hgweb The newlines make the HTML somewhat readable and gives much smaller, readable diffs when the tests change.

test-highlight: decouple test from get-with-headers.py The test copied get-with-headers.py from $TESTDIR and committed it to a test repository. The test output therefore depended unnecessarily on the exact content of get-with-headers.py. It has now been replaced with another small Python script.

Merge with crew-stable

convert/cvsps: use set.update for bulk update

hbisect: use set.update for bulk updates

localrepo: use set.update for bulk updates

store: create set directly from iterable

repair: bulk update sets Use a single set.update and set.difference_update call instead of many set.add and set.discard calls.

ui: use set instead of dict

alias: honor the [defaults] section, fix issue1642

templater: keep a cache of Python functions for filter expressions

templater: replace regex complexity by simple str containment checks

templater: clean up the process method, separating code paths

patchbomb: make diffstat usage clearer Making it clear that diffstat output does not only depend on diffstat being installed - it also needs to be explicitely requested.

patchbomb: describe --attach and --inline options in help

patchbomb: fix help to reflect actual operation In the body, the description is shown in its entirety, unstripped from what was already in the subject.

convert/bzr: warn when source is a lightweight checkout (issue1647)

localrepo: use set instead of dict

copies: use set instead of dict

store: use set instead of dict

verify: use set instead of dict

ancestor: use set instead of dict

revlog: use set instead of dict

bisect: use set instead of dict

repair: use set instead of dict

patch: use set instead of dict

record: use set instead of dict

hgk: use set instead of dict

convert: cleanups in monotone converter

merge with crew

convert: use set instead of dict

bugzilla: use set instead of dict

rebase: use set instead of dict

revlog.missing(): use sets instead of a dict

tests: add missing interpreter lines

cvsnt-mergepoints test: use sh instead of bash

tests: removed unnecessary execute bit on Python tests

tests: renamed Python tests to .py

tests: renamed helper to avoid clash with stdlib repr module

tests: fix doc string in get-with-headers.py

tests: use repr.py to avoid binary characters in test output

tests: new helper for tests with binary output

convert: improve docstrings, comments.

changelog: turn {de,en}code_extra methods into functions The methods were not really methods -- they didn't use 'self'. Having them as functions in the module it useful for other modules (like the commitsigs extension) that want to recompute the changeset hash and thus want to encode dicts the same way as changelog does it. Removed the underbars from their names at the same time.

merge with crew-stable

contrib/win32: add simple style sheet

contrib/win32: updated title of the hgbook

contrib/win32: format file names and env vars with <tt>

contrib/win32: indent and word wrap consistently

contrib/win32: add missing <p> tags

contrib/win32: fix HTML tag nesting

contrib/win32: license is GPLv2 only according to Matt Mackall

Merge with crew-stable

mq: qpush shouldn't complain if working dir is at a non-tip head

contrib: add tmplrewrite.py script to help rewrite old templater syntax

spartan: use newer template syntax everywhere

monoblue: use newer template syntax everywhere

raw: use newer template syntax everywhere

atom/rss: use newer template syntax everywhere

git: use newer template syntax everywhere

Add tests for qfold

mq: abort qfold upon local changes (issue1662)

changelog: refuse to add revisions with empty usernames An empty username or a username with a "\n" will make the revision text contain two "\n\n" sequences -> corrupt repository. The problem is that changelog.read expects to find exactly one "\n\n" separator and thus cannot unpack the revision.

convert/bzr: fix symlinks target (issue1626/2)

changelog: removed bad default arguments in add method The arguments defaulted to None, but - user cannot be None since it is immediately stripped. - p1 and p2 cannot be None since they are passed directly to revlog.addrevision, where they are mandatory.

posixfile: remove posixfile_nt and fix import bug in windows.py The posixfile_nt class has been superseded by posixfile in osutils.c, which works on Windows NT and above. All other systems get the regular python file class which is assigned to posixfile in posix.py (for POSIX) and in the pure python version of osutils.py (for Win 9x or Windows NT in pure mode).

util: make atomictempfile saner if mktempcopy fails

merged with crew-stable

util: import random, missing since ea82a23cf887

clone: try updating to the actual changeset specified in options When cloning with the -r option or # url format from a tag the destination repo most likely won't have the tag. We can save the lookup result to get to the correct parent anyway. Similar to issue1306, but for tags.

commit: hoist the rest of the dirstate manipulation out of commitctx

commit: hoist up dirstate invalidate

commitctx: use contexts more fully

context: add new manifestnode method

commitctx: eliminate some variables

commit: move lots of commitctx outside of the repo lock

commit: combine _commitctx and commitctx, drop unused force argument

editor: move HG: filtering from ui to commiteditor

grep: make cache LRU rather than unlimited grep could cache an unbounded number of revlogs, limit to 20 with an LRU cache.

commit: move commit editor to cmdutil, pass as function

context: add p1 and p2 methods

commit: push repo lock down into _commitctx

commit: move 'nothing changed' test into commit()

commit: drop unused p1 and p2 args

tag: drop unused use_dirstate and parent from _tag()

filecommit: swallow some bits from _commitctx, add _

commitctx: replace wctx with ctx

commitctx: replace two dirstate vars with working

commitctx: simplify locking (spotted by Simon Heimberg)

remove deprecated rawcommit

patch: copy copies dict before changing it (issue1651)

status: check cmp list in order

verify: detect manifest revs not in any changeset

localrepo: use more direct vars in addchangegroup