i18n-ja: synchronized with 5246fd4781d8

i18n-pt_BR: synchronized with 224e96078708

test-import-git.t: work around check-code hating on ^ and $( in tests

patch: add support for git delta hunks When creating patches modifying binary files using "git format-patch", git creates 'literal' and 'delta' hunks. Mercurial currently supports 'literal' hunks only, which makes it impossible to import patches with 'delta' hunks. This changeset adds support for 'delta' hunks. It is a reimplementation of patch-delta.c from git : http://git.kernel.org/cgit/git/git.git/tree/patch-delta.c

pathcomplete: remove ambiguous entries for sole completion on a directory Previously, directories were added with the trailing slash and, if there was only one completion, then another ambiguous entry was created using '.', as follows: $ hg rm mer<TAB> mercurial/./ mercurial// This was added in fa6d5c62f3bd (though, some logic existed before that) to work around bash completion adding a space after the sole entry because we treated directories and files the same. We no longer do that now so we remove this unneeded code. Tests have been updated to match this new behavior.

bash_completion: add global support for -b|--branch Previously, only -r|--rev was parsed globally which meant 'hg push -b <tab>' would try to complete a path instead of a branch.

bash_completion: add global support for -B|--bookmark Previously, only -r|--rev was parsed globally which meant 'hg push -B <tab>' would try to complete a path instead of a bookmark.

bash_completion: change --rev if-else block into a case Again, this doesn't change behavior but does make it easy to add cases in the next patch.

bash_completion: simplify if-else logic There is no change in functionality here but this makes refactoring this later easier.

bash_completion: add _hg_branches for list of branches

bash_completion: add -B|--bookmark support for strip

bash_completion: add completion for deleting a shelve

bash_completion: use correct command for listing shelves Previously, _hg_shelves used the unshelve command to list current shelves. This is actually the wrong command. The correct usage is 'hg shelve -l' so we use that instead.

bash_completion: remove restriction on bookmark completion Previously, we only completed a bookmark when deleting or renaming a bookmark. Since we can now move a bookmark forward, we drop this restriction.

bash_completion: fix issue with subdirectories not being completed Previously, if there was a directory between the file and first-level directory (e.g. 'bar' in foo/bar/file), then bash_completion would only list 'foo/file' instead of 'foo/bar/file'. This behavior was introduced in 80f3ae36f908 to fix spaces in file names. In this patch, we keep that behavior while also fixing subdirectory completion by reverting 80f3ae36f908 and instead add backslashes to whitespace manually. This approach means adding the completion option 'nospace' since we do this manually now.

bash_completion: determine root executable in case of alias This solves an issue for users that have "alias hg='hg --some_opts'" where they would see wrong entries in the completion output (such as color codes).

bash_completion: expand aliased commands

bash_completion: untabify

dispatch: turn off custom debugger for HGPLAIN mode Some debuggers, such as ipdb, load escape codes and color codes even when later turned off. This will affect scripts that do simple parsing and can't handle escape codes. Therefore, we only load a custom debugger if ui.plain() is false.

Makefile: temporary copy of hg.pot is hg.pot.tmp instead of hg.pot~ The file ending of temporary files for generating XX.po has been changed to .tmp (from ~). Use the same ending when generating hg.pot.

Makefile: fix po file updating by using same file name everywhere The update is done on a temporary file since ce5d711475a3. The name of this file was first XX.po~ and has been changed to XX.po.tmp in all except of one place. Do this now.

mq: prefer a loop to a double-for list comprehension The [x for y in l for x in y] syntax is nigh-incomprehensible, and this is a particularly easy case to expand into a loop since there's no 'if' condition in the list comprehension.

graphlog: mark as deprecated

tests: don't load unnecessary graphlog extension Since graphlog is in core, we can use 'hg log -G' instead.

hg: remove outdated comment The graphlog functionality has been in core since Mercurial 2.3.

tests: use strip extension instead of mq where it makes sense When a test only uses 'hg strip', it no longer needs to load MQ.

tests: fix `hghave root` on windows This lets tests\test-hghave.t pass on windows where geteuid does not exist.

keyword: wlock while setting branch in kwdemo

tests: add missing glob

Merge

mercurial/parsers.c: fix compiler warning When try to compile on x64 OS X, I get this warning: mercurial/parsers.c:931:27: warning: implicit conversion loses integer precision : 'long' to 'int' [-Wshorten-64-to-32] ? 4 : self->raw_length / 2; The patch verifies if value of self->raw_length falls bellow INT_MAX; if not, it raises the ValueError exception. If value of self->raw_length is greater than 4, it's casted to int type, to eliminate the warning.

parse_index2: fix crash on bad argument type (issue4110) Passing a non-string to parsers.parse_index2() causes Mercurial to crash instead of raising a TypeError (found on Mac OS X 10.8.5, Python 2.7.6): import mercurial.parsers as parsers parsers.parse_index2(0, 0) Thread 0 Crashed:: Dispatch queue: com.apple.main-thread 0 parsers.so 0x000000010e071c59 _index_clearcaches + 73 (parsers.c:644) 1 parsers.so 0x000000010e06f2d5 index_dealloc + 21 (parsers.c:1767) 2 parsers.so 0x000000010e074e3b parse_index2 + 347 (parsers.c:1891) 3 org.python.python 0x000000010dda8b17 PyEval_EvalFrameEx + 9911 This happens because when arguments of the wrong type are passed to parsers.parse_index2(), indexType's initialization function index_init() in parsers.c leaves the indexObject instance in a state that indexType's destructor function index_dealloc() cannot handle. This patch moves enough of the indexObject initialization code inside index_init() from after the argument validation code to before it. This way, when bad arguments are passed to index_init(), the destructor doesn't crash and the existing code to raise a TypeError works. This patch also adds a test to check that a TypeError is raised.

merge with stable

bookmarks: consider successor changesets when moving bookmark (issue4015) Previously, this required -f because we didn't consider obsolete changesets (and their children ... or successors of those children, etc.). We now use obsolete.foreground to calculate acceptable changesets when advancing the bookmark. Test coverage has been added.

util: url keeps backslash in paths Backslashes (\) in paths were encoded to %C5 when converting from url to string. This does not look nice for windows paths. And it introduces many problems when running tests on windows.

tests: deal with new gits sending status messages to stderr git-1.8.4.2 will send messages like Cloning into 'X'... done. to stderr. Mute stderr.

subrepo: sanitize non-hg subrepos

shelve: fix bad argument interaction with largefiles (issue4111)

strip.stripcmd: remove redundant wlock acquire/release Now that we acquire and release a wlock in the outer scope, this is redundant.

merge with stable

bookmarks: make setcurrent with None an error

strip: use bookmarks.unsetcurrent instead of setcurrent with None

bmstore.write: use unsetcurrent instead of setcurrent with None

commands.bookmark: use unsetcurrent instead of setcurrent with None There are currently two different ways we can have no active bookmark: .hg/bookmarks.current being missing and it being an empty file. This patch and upcoming ones make an empty file the only way to represent no active bookmarks. This is the right choice because it matches the state that a new repository without bookmarks will be in.

strip: hold wlock for entire duration Previously, we'd acquire and release the wlock several times. This meant that other hg processes could come in and change state. Instead of that, retain the wlock for the entire duration of the strip.

merge with stable

bisect: report "both good and bad" as such, not as "not directly related"

phase: better error message when --force is needed When trying to turn a draft changeset into a secret changeset, I was told: % hg phase -s . cannot move 1 changesets to a more permissive phase, use --force no phases changed That message struck me as being backwards -- the secret phase feels less permissive to me since it restricts the changesets from being pushed. We don't use the word "permissive" elsewhere, 'hg help phase' talks about "lower phases" and "higher phases". I therefore reformulated the error message to be cannot move 1 changesets to a higher phase, use --force That is not perfect either, but more in line with the help text. An alternative could be cannot move phase backwards for 1 changesets, use --force which fits better with the help text for --force.

hgweb: ignore non numeric "revcount" parameter values (issue4091)

lock: take both vfs and lock file path relative to vfs to access via vfs This patch makes "lock.lock.__init__()" take both vfs and lock file path relative to vfs, instead of absolute path to lock file. This allows lock file to be accessed via vfs.

vfs: add "makelock()" and "readlock()"

hg: rewrite "copystore()" with vfs This patch rewrites "copystore()" with vfs, because succeeding patch requires "lock.lock()" invocation with vfs. This patch uses 'dstbase + "/lock"' instead of "join()" with both elements even on Windows environment but it should be reasonable, because target files given from "store.copyfiles()" already uses "/" as path separator. "util.copyfiles()" between two vfs-s may have to be rewritten in the future.

localrepo: import "lock" module as "lockmod" for readability Before this patch, "localrepo.py" has many methods defining local variable "lock", even though it imports "lock" module as "lock". This ambiguity decreases readability.

transaction: take journal file path relative to vfs to use file API via vfs

vfs: add "chmod()"

vfs: add "isfile()"

transaction: unlink target file via vfs Before this patch, unlink target file is once opened before unlinking, because "opener" before vfs migration doesn't have "unlink()" function. This patch uses "vfs.unlink()" instead of "open()" and "fp.name".

relink: abort earlier when on different devices (issue3916) Add a first check of the devices before collecting candidate files. This is much quicker when big repos are on different devices. Keep the existing check in prune. It checks for same device of the files. This could probably be different in a special repo store (with symlinks).

localrepo: prevent to copy repo local config, copy baseui instead Copying a repos local configuration to another repo is a bad idea because the 2nd repo gets the configuration of the first. Prevent this by really calling repo.baseui.copy when repo.ui.copy is called. This requires some changes in commandserver which needs to clone repo.ui for rejecting temporary changes. This patch has its roots back in the topic "repo isolation" around f0564402d059 and was suggested by mpm.

doc: show details of command options in pages generated by docutils Before this patch, HTML/man pages generated by docutils don't show details of each command options, whether it should take argument or not for example, even though "hg help" does. This patch shows details of command options as same as "hg help" shows. This patch uses "--option <VALUE[+]>" style instead of "--option <VALUE> [+]" used in output of "hg help", because docutils requires that option argument strings starts with "<" and ends with ">".

Makefile: intermediate steps work with temporary copy of hg.pot In case of abortion or a failing step, the target file remains untouched instead of being in an intermediate state.

merge with stable

help: drop help for branches template keyword The {branches} keyword dates to pre-1.0 Mercurial's tag-like branch scheme which allowed changesets to be on multiple branches. This is the last visible vestige of that scheme, users should instead be using {branch}, possibly with if().

test-command-template.t: fix failure due to trailing whitespace

templater: makes branches work correctly with stringify (issue4108)

manifest: increase lrucache from 3 to 4 During a commit amend there are 4 manifests being handled: - original commit - temporary commit - amended commit - merge base This causes a manifest cache miss which hurts perf on large repos. On a large repo, this fix causes amend to go from 6 seconds to 5.5 seconds.

strip: add faster revlog strip computation The previous revlog strip computation would walk every rev in the revlog, from the bottom to the top. Since we're usually stripping only the top few revs of the revlog, this was needlessly expensive on large repos. The new algorithm walks the exact number of revs that will be stripped, thus making the operation not dependent on the number of revs in the repo. This makes amend on a large repo go from 8.7 seconds to 6 seconds.

revlog: return lazy set from findcommonmissing When computing the commonmissing, it greedily computes the entire set immediately. On a large repo where the majority of history is irrelevant, this causes a significant slow down. Replacing it with a lazy set makes amend go from 11 seconds to 8.7 seconds.

merge with stable

histedit: hold wlock and lock while in progress Currently, histedit acquires and releases lock and wlock several times during its run. This isn't great because it allows other hg processes to come in and change state. With this fix, lock and wlock are acquired and released exactly once. The change to test-histedit-drop.t is a minor implementation one -- the cache is still correctly invalidated, but it just happens a little later and only gets printed out because of the unrelated --debug flag.

merge with i18n

i18n-pt_BR: fix typos in notify docs

i18n-pt_BR: synchronized with 06e118c097ff

templater: only recursively evaluate string literals as templates (issue4103)

templater: fix escaping in nested string literals (issue4102) Before the templater got extended for nested expressions, it made sense to decode string escapes across the whole string. Now we do it on a piece by piece basis.

merge with stable

shelve: unshelve using an unfiltered repository when evolve is enabled and a hidden obsolete changeset exists in the repository, the strip during unshelve will fail due to filtered revs. we use an unfiltered repository like to repair.strip to strip the proper nodes.

largefiles: update in two steps, handle interrupted updates better An update would try to fetch any missing largefiles after having updated normal files and standins. That could fail or be interrupted and would leave the working directory in a state where the largefiles not only were missing but also were scheduled for remove ... and where the old largefile was left in place. Instead we now remove old largefiles before starting to download and update missing largefiles.

largefiles: inline _updatelfile, prepare for further refactorings

largefiles: cache largefiles for update, also without printmessage

largefiles: cleanup of printmessage handling - the printed flag was redundant

test-module-imports: skip on Python < 2.6, since ast is new in 2.6

Merge with stable.

convert: fix svn crash when svn.ra.get_log calls back with orig_paths=None get_log started calling back with orig_paths=None on Fedora 20 with subversion-1.8.3. That broke test-convert-svn-source.t . There used to be some handling of that situation until e8f4e40f285a apparently broke it. This patch restores what seems to be the most obvious handling of the situation.

share: fix unshare calling wrong repo.__init__() method When running the unshare command, if there's other code that tries to use the repo after the command is finished, it'll end up with a ui object for repo.unfiltered(). This change fixes an erroneous call to repo.__init__() that could be on the repoview proxy class--now it's always done on the unfiltered repo.

statichttprepo.httprangeheader: implement readlines bookmarks.readcurrent() requires readlines() on file objects returned from repo.vfs. It isn't used right now but will be in upcoming patches.

statichttprepo: pass in True to splitlines, not 1 splitlines actually takes a boolean.

mq: don't add '* * *' separators when there is no commit message

bisect: avoid confusing use of variables with same names in nested local scopes

discovery: make note messages for new heads more readable

discovery: tweak error message for multiple branch heads

test-obsolete-checkheads.t: remove todo comment that was fixed in be0fcbb1c92f

convert: readability and test of rpairs function

tests: make doctest test runner less verbose

tests: python hash seed is only relevant after failures - otherwise keep quiet

scmutil.filecache: support watching over multiple files

scmutil: introduce a filecacheentry that can watch multiple paths

scmutil: rename filecacheentry to filecachesubentry Upcoming patches will allow the file cache to watch over multiple files, and call the decorated function again if any of the files change. The particular use case for this is the bookmark store, which needs to be invalidated if either .hg/bookmarks or .hg/bookmarks.current changes. (This doesn't currently happen, which is a bug. This bug will also be fixed in upcoming patches.)

scmutil.filecacheentry: make stat argument to constructor mandatory There's no real upside to making stat optional -- this constructor has just two callers.

test-filecache.py: add markers to the output for each event Previously it was possible that a different, incorrect set of events might print out 'creating' the same number of times.

test-filecache.py: make setbeforeget test clearer '0' and 'None' as outputs tripped me up. Make the distinction between values set externally and values computed by calling the decorated function clearer.

test-module-imports.t: new test to use the import cycle detector

import-checker: try a little harder to show fewer cycles This makes sure that all cycles begin with the lexicographically first module, so that we're less likely to show overlapping cycles in the final analysis.

import-checker: ignore nested imports

contrib: add an import checker This checks for cycles in the module graph and verifies that imports of stdlib modules are not on the same line as relative imports of mercurial modules.

makememctx: move from patch to context to break import cycle

cleanup: move stdlib imports to their own import statement There are a few warnings still produced by my import checker, but those are false positives produced by modules that share a name with stdlib modules.

pathutil: tease out a new library to break an import cycle from canonpath use

subsettable: move from repoview to branchmap, the only place it's used This is a step towards breaking an import cycle between revset and repoview. Import cycles happened to work in Python 2 with implicit relative imports, but breaks on Python 3 when we start using explicit relative imports via 2to3 rewrite rules.

obsolete: add __eq__ and __hash__ to marker to make set() deduplication work

obsolete: do not accept duplicated marker during exchange Before this patch, duplicated obsolescence markers could slip into an obstore if the bookmark was unknown locally and duplicated in the incoming obsolescence stream. Existing duplicate markers will not be automatically removed but they'll stop propagating. Having a few duplicated markers is harmless and people have been warned evolution is <blink>experimental</blink> anyway.

Merge with stable.

obsolete: stop doing membership test on list According to the Surgeon General, computer should not use list for membership testing because of the risk of being slow.

bookmarks: rewrite pushing bookmarks in "localrepository.push()" by "compare()" This patch adds "updateremote()", which uses "compare()" to compare bookmarks between the local and the remote repositories, to replace pushing local bookmarks in "localrepository.push()".

bookmarks: rewrite pushing local bookmarks in "commands.push()" by "compare()" This patch adds "pushtoremote()", which uses "compare()" to compare bookmarks between the local and the remote repositories, to replace pushing local bookmarks in "commands.push()".

bookmarks: rewrite "updatefromremote()" by "compare()" To update entries in bmstore "localmarks", this patch uses "bin(changesetid)" instead of "repo[changesetid].node()" used in original "updatefromremote()" implementation, because the former is cheaper than the latter.

bookmarks: add function to centralize the logic to compare bookmarks This patch adds "compare()" function to centralize the logic to compare bookmarks between two repositories.

hgweb: use semantically suitable filelog.revs in filelog The functions are equivalent in behaviour, so no behavior change.

hgweb: always compute all entries and latestentry in filelog This is the same thing which was done for changelog earlier, and it doesn't affect performance at all. This change will make it possible to get the first entry of the next page easily without computing the list twice.

hgweb: remove unused argument of entries function in filelog This doesn't change the behavior as the argument wasn't used anyway.