merge with crew-stable

largefiles: don't query the dirstate for key None lfutil.splitstandin(f) can be None, and we query the dirstate for that without checking if it is. This will cause problems with the upcoming move to critbit- based dicts, since they only support strings as keys.

merge

bookmarks: moving the active bookmark deactivates it After this change, moving the active bookmark somewhere other than the current changeset (i.e., with --rev) deactivates it. Previously it would remain in .hg/bookmarks.current, which seems like a bug.

bookmarks: allow (re-)activating a bookmark on the current changeset Allow a bookmark that points to the current changeset to be made the active bookmark without requiring --force. Previously, this would've aborted with: abort: bookmark 'Z' already exists (use -f to force)

applyupdates: assign variable before we try to use it (issue3855) The variable 'fd' was getting used with a value left over from a prior iteration, causing a KeyError: '.hgsubstate'.

revset: don't abort when regex to tag() matches nothing (issue3850) This makes the tag("re:...") revset consistent with branch("re:...").

rebase: fix --collapse when a file was added then removed When a series of commits first adds a file and then removes it, hg rebase --collapse prompts whether to keep the file or delete it. This is due to it reusing the branch merge code. In a noninteractive terminal it defaults to keeping the file, which results in a collapsed commit that is has a file that should be deleted. This bug resulted in developers accidentally commiting unintentional changes to our repo twice today, so it's fairly important to get fixed. This change allows rebase --collapse to tell the merge code to accept the latest version every time without prompting. Adds a test as well.

Merge

grep: use re2 if possible Since re2 is enormously faster than Python's re module, this should help performance, ceteris paribus. grep currently has other huge performance problems that mask any gain :-(

util: add flag support to compilere

bookmarks: fix test broken by 0bba1ff2ac7b The added test reflects an error in printing the origin node of the bookmark that I fixed before committing, but I forgot to update the test (whoops).

bookmarks: allow moving a bookmark forward to a descendant Allow 'hg bookmark MARK', with an existing bookmark MARK, to move the bookmark forward to the current or specified revision, if the target revision is a descendant of the revision the bookmark currently points to. Prints a status message including the revision the bookmark was formerly at: $ hg bookmark Z moving bookmark 'Z' forward from 663762316562 Test coverage is added.

hgweb: change manifest archive links to only archive the current directory When the web server shows the manifest for a single, non top directory, append the path to the directory to the archive links. This makes the web server generate archive files that only include the current directory (and its subdirectories). Note that archive links in other pages (e.g. changeset) or at the top of the manifest are unchanged. Directory archive links have an extra "/" at the end which does not impact the result of the archive operation. Keeping it there made the implementation of this feature simpler.

hgweb: teach archive how to download a specific directory or file The archive web command now takes into account the "file" request entry, if one is provided. The provided "file" is processed as a "path" corresponding to a directory or file that will be downloaded. With this change hgweb can to process requests such as: http://mercurial.selenic.com/hg/archive/tip.zip/mercurial/templates This will download all files on the mercurial/templates directory as a zip file. It is not possible to specify file patterns ('glob', 'relglob', 'path', 'relpath', 're', 'relre' nor 'set'). The server will reject those with a 403 HTTP error response. Note that this is a first step to add support for downloading directories from the web interface. A following patch will modify the archiveentry map entry on the different templates so that it adds the current folder path to the archive links.

test-archive: gracefully handle HTTPErrors on get-with-headers This avoids pritting out a traceback when a get-with-headers call causes hgweb to respond with an HTTPError code.

bash_completion: tell an editor what type of file this is

bash_completion: allow remove to complete normal files Previously, we only completed files that had already been manually deleted. That behaviour made no sense. We now complete unmodified, modified, and deleted files.

bash_completion: match more narrowly This greatly helps completion performance for most commands that deal with files. In a working dir with 150,000 files, where we want to complete the name of a modified file under a path beginning with "a", from the root of the working dir: (old) hg status -nm . 1.7 sec (new) hg status -nm "glob:a**" 0.3 Even "hg add" becomes a little faster, in spite of being the worst case (matching untracked files).

blackbox: add backup bundle paths to blackbox logs Writes the backup bundle paths to the blackbox so it's easy to see which backup bundle is associated with which command when you are debugging an issue. Example output: 2013/03/13 10:39:56 durham> strip tip 2013/03/13 10:39:59 durham> saved backup bundle to /data/users/durham/www-hg/.hg/strip-backup/e5fac262363a-backup.hg 2013/03/13 10:40:03 durham> strip tip exited 0 after 7.97 seconds

tests: fix test-profile to not depend on HGPROF environment variable The test-profile test would fail if the user had HGPROF set to another profiler in their environment. This fix makes the test independent of that environment variable. Reverts the previous attempt to fix this, which was not cross platoform.

merge with crew-stable

tests: fix test-profile to not depend on HGPROF environment variable The test-profile test would fail if the user had HGPROF set to another profiler in their environment. This fix makes the test independent of that environment variable.

check-code: do not prepend "warning" to a failure message The prefix has not been removed when this check changed from a warning to a failure.

merge with crew-stable

strip: make --keep option not set all dirstate times to 0 hg strip -k was using dirstate.rebuild() which reset all the dirstate entries timestamps to 0. This meant that the next time hg status was run every file was considered to be 'unsure', which caused it to do expensive read operations on every filelog. On a repo with >150,000 files it took 70 seconds when everything was in memory. From a cold cache it took several minutes. The fix is to only reset files that have changed between the working context and the destination context. For reference, --keep means the working directory is left alone during the strip. We have users wanting to use this operation to store their work-in-progress as a commit on a branch while they go work on another branch, then come back later and be able to uncommit that work and continue working. They currently use 'git reset HARD^' to accomplish this in git.

sshpeer: store subprocess so it cleans up correctly When running 'hg pull --rebase', I was seeing this exception 100% of the time as the python process was closing down: Exception TypeError: TypeError("'NoneType' object is not callable",) in <bound method Popen.__del__ of <subprocess.Popen object at 0x937c10>> ignored By storing the subprocess on the sshpeer, the subprocess seems to clean up correctly, and I no longer see the exception. I have no idea why this actually works, but I get a 0% repro if I store the subprocess in self.subprocess, and a 100% repro if I store None in self.subprocess. Possibly related to issue 2240.

blackbox: fix exception when logging commands with format characters When running commands like 'hg export -o mypatch-%N.patch', the blackbox would throw an exception because it tried to format %N. This change prevents it from trying to format the command string.

localrepo: always write the filtered phasecache when nodes are destroyed (issue3827) When the strip command is run, it calls repo.destroyed, which in turn checks if we read _phasecache, and if we did calls filterunknown on it and flushes the changes immediately. But in some cases, nothing causes _phasecache to be read, so we miss out on this and the file remains the same on-disk. Then a call to invalidate comes, which should refresh _phasecache if it changed, but it didn't, so it keeps using the old one with the stripped revision which causes an IndexError. Test written by Yuya Nishihara.

merge with stable

rebase: restore active bookmark after rebase --continue When a rebase has conflicts and the user uses rebase --continue, the previously active bookmark was not being made active once again. With this change that bookmark is made active again, just as if the rebase had never been interrupted. This changes the rebasestate file format, but should handle old formats correctly. Since the file is transient, this is even less of a problem. Adds a test to verify the new behavior. I manually tested continuing rebases with and without an active bookmark, and with and without being on the bookmark being rebased.

setup.py: ignore warnings from obsolete Fix version detection with recent hg versions.

setup.py: add metadata to register package to PyPI This allows updating PyPI with just 'setup.py register' (assuming the user is an owner or maintainer of the PyPI package).

minirst: optimize HTML table generation a bit avoid a couple of array copies and string interpolations

minirst: HTML formatter tweaks output table rows on distinct lines don't make the first row a table header

minirst: CGI escape strings prior to embedding it in the HTML

minirst: don't generate TH elements for the first row in table We mainly use them for option tables -- browsers bolding and centering them looks wrong.

help: use a full header for topic titles ...for prettier HTML!

hgweb: generate HTML documentation It's generated from the raw ReST source, as returned from help.help_().

help: move the majority of the help command to the help module We move the logic for generating the unformatted ReST source to the help module, in order to eventually avoid calling commands.help_() from hgweb. No functionality change.

hgweb help: split up long lines (in generated output)

hgweb help: add tests

merge with stable

merge with crew

Added signature for changeset 5b7175377bab

Added tag 2.5.2 for changeset 5b7175377bab

setparents: drop copies from dropped p2 (issue3843)

httppeer: avoid large dumps when we don't see an hgweb repo When we don't get an hgweb protocol response, we dump the response to the user for diagnostic purposes (it might be a cgitb message, for instance). But if we try to clone a bundle, we don't want to show the entire bundle in the error message. Also, we don't want fetch the full bundle multiple times during fallback. So we only fetch 1k here.

httppeer: improve protocol check Pre-0.6c hgweb used text/plain for protocol responses. This meant that a web server could serve a static file and confuse a client into generating a nasty traceback. Now we insist that text/plain protocol responses not include a Content-Length, which older hgweb didn't generate but will typically be produced for static files.

util: add a timed function for use during development I often want to measure the cost of a function call before/after an optimization, where using top level "hg --time" timing introduces enough other noise that I can't tell if my efforts are having an effect. This decorator allows a developer to measure a function's cost with finer granularity.

util: generalize bytecount to unitcountfn This gives us a function we can reuse to count units of other kinds.

help: new SHA-1 fingerprint of hg.intevation.org in hostfingerprints example The certificate was updated in March 2012. You can verify the certificate by using the Root CA certificate downloadable from https://ssl.intevation.de/ The intermediate CA is sent by https://hg.intevation.org/

tests: work around msys bash mangling of paths in test-mq.t

largefiles: remove wrong comment on standin matcher and code for creating dirs The existence of these directories could not be relied on anyway and all dependencies on them should have been removed.

merge with stable

templatefilters: add missing import of _

largefiles: updatelfiles should use working dir standins, not standins from p1 This makes a difference when working directory is dirty, especially when merging with a revision for which we don't have largefiles.

largefiles: simplify cachelfiles - don't spend a lot of time checking hashes cachelfiles jumped through loops to handle merges and modified files ... but it did apparently no longer have a valid reason to do so. It should just always make sure that the largefiles referenced from the standins are present - no matter which actual largefile is stored in the working directory. If there is no standin then there is nothing to fetch. The old code usually verified the hash of all largefiles every time this function was invoked - for examply by 'update'. This change makes a trivial noop update 5-10 seconds faster on our repo (with the other 50% spent doing another unnecessary hashing of all largefiles).

largefiles: don't let update leave wrong largefiles in wd if fetch fails Situations where a largefile for some reason wasn't available sometimes caused wrong largefile content and state. It has mostly been seen when interrupting download of largefiles ... and when introducing programming errors. Instead we now make sure to delete the old and wrong largefile. A missing file is a well-known error condition and much more reasonable way to handle the situation.

largefiles: missing largefiles should not be committed as removed Largefiles can easily become missing - for example if it simply isn't available or the download fail. It might even be convenient to be able to work that way in some cases. But commiting missing largefiles as if they had been 'hg remove'd is plain wrong.

largefiles: don't assume that .hg/largefiles/ still exists It might not have been created and it might have been removed.

largefiles: getstandinmatcher should not depend on existence of directories Looking for a (potentially empty) directory was not reliable - both because it is a reasonable assumption that empty directories can be removed and because it wasn't created in all cases ... such as when pulling to an existing repository.

tests: don't rely on broken behaviour in test-largefiles-cache.t The test relied on the bug that 'pull largefiles from branchheads' didn't pull any largefiles from tip revision when it seemed like no largefiles had been checked out before.

largefiles: fix download of largefiles from an empty list of changesets The empty list was interpreted as all revisions - just like None is. The empty list is now handled explicitly.

largefiles: fix off-by-one error on pull --all-largefiles Test output is changed in a case where one revision was pulled, but because of the off-by-one error it thought that 0 revisions were pulled ... and because of another bug it thus (tried to) fetch largefiles for all revisions. After this change it no longer reports failure when it failed while trying to fetch largefiles it shouldn't fetch. Largefiles that it shouldn't fetch but managed to fetch anyway will now correctly be missing later on. This change thus resolves some of unexplained test output introduced in 1e4eb1faba6e.

tests: remove glob lines which unnecessary match / for \ on windows This lines were reported as unnecessary when running the tests on windows because the path was already printed with a slash and not a backslash.

tests: append glob to filename output when required for windows The test failed on windows before this patch.

convert: stabilize cvsps commitid sort order

pager: catch ctrl-c on exit (issue3834)

merge with crew

templater: properly handle file_copies with %

localrepo: iterate over manifest key/value pairs in status This saves us a couple of dict lookups in the common case, and improves the performance of the status method by 5% (measured with util.timed) in a repo with a large manifest.

match: more accurately report when we're always going to match This improves the performance of log --patch and --stat by about 20% for moderately large manifests (e.g. mozilla-central) for the common case of no -I/-X patterns.

scmutil: fix NameError on windows Forgot to import util in mercurial.scmwindows. Error was introduced in: changeset: 18862:4c6f7f0dadab user: Kevin Bullock <kbullock@ringworld.org> date: Tue Feb 12 11:36:21 2013 -0600 summary: scmutil: split platform-specific bits into their own modules

commands: exit from the log loop at the right time Previously, we'd run the iterator one more time than necessary, potentially doing a lot of extra work in the process.

cmdutil: use a small initial window with --limit In a large repo, running a command like "log -l1 -p" was expensive because it would always traverse 8 commits, as 8 was the initial window size. We now choose the lesser of 8 or the limit, speeding up the "log -l1 -p" case by a factor of 5.

worker: handle worker failures more aggressively We now wait for worker processes in a separate thread, so that we can spot failures in a timely way, wihout waiting for the progress pipe to drain. If a worker fails, we recover the pre-parallel-update behaviour of failing early by killing its peers before propagating the failure.

worker: fix a race in SIGINT handling This is almost impossible to trigger due to the tiny time window involved.

worker: on error, exit similarly to the first failing worker Previously, if a worker failed, we exited with status 1. We now exit with the correct exit code (killing ourselves if necessary).

merge with stable

merge with main

largefiles: don't cache largefiles for pulled heads by default After discussion, we've agreed that largefiles for newly pulled heads should not be cached by default. The use case for this is using largefiles repos with multiple remote servers (and therefore multiple remote largefiles caches), where users will be pulling from non-default locations on a regular basis. We think this use case will be significantly less common than the use case where all largefiles are stored on the same central server, so the default should be no caching. The old behavior can be obtained by passing the --cache-largefiles flag to pull.

merge with stable

merge with crew

bundle: treat branches created newly on the local correctly (issue3828) Before this patch, "hg bundle --branch foo other" fails to create bundle file, if specified "foo" branch is created newly on the local repository. "hg bundle" uses "hg.addbranchrevs(repo, other, ...)" to look branch names up, even though other outgoing-like implementation uses "hg.addbranchrevs(repo, repo, ...)". In the former invocation, "other" repository recognizes such branches as unknown, so execution is aborted. This patch uses "hg.addbranchrevs(repo, repo, ..)" in "hg bundle" to bundle revisions on such branches correctly.

merge with stable

mergetools: refine vimdiff warning message We explicitly redraw before echoing the message so that it simply displays at the bottom of the window. Also simplifies the message printing by using 'echomsg' (which uses 'echohl' internally) and adds the names of the software involved for improved Googleability.

mergetools: vimdiff issue a warning explaining how to abort Adds a message displayed at each vimdiff invocation: merge conflict detected, type ":cq" to abort Vimdiff is very confusing for non-vim user (not to speak about vim user confused anyway. However it is very likely that vimdiff is picked as the mergetool of choice when using the default config: - vim is available on all UNIX system. - Its one of the rare non graphical merge tools.

httppeer: improve protocol check Pre-0.6c hgweb used text/plain for protocol responses. This meant that a web server could serve a static file and confuse a client into generating a nasty traceback. Now we insist that text/plain protocol responses not include a Content-Length, which older hgweb didn't generate but will typically be produced for static files.

httppeer: avoid large dumps when we don't see an hgweb repo When we don't get an hgweb protocol response, we dump the response to the user for diagnostic purposes (it might be a cgitb message, for instance). But if we try to clone a bundle, we don't want to show the entire bundle in the error message. Also, we don't want fetch the full bundle multiple times during fallback. So we only fetch 1k here.

mergetools: refine vimdiff warning message We explicitly redraw before echoing the message so that it simply displays at the bottom of the window. Also simplifies the message printing by using 'echomsg' (which uses 'echohl' internally) and adds the names of the software involved for improved Googleability.

mergetools: vimdiff issue a warning explaining how to abort Adds a message displayed at each vimdiff invocation: merge conflict detected, type ":cq" to abort Vimdiff is very confusing for non-vim user (not to speak about vim user confused anyway. However it is very likely that vimdiff is picked as the mergetool of choice when using the default config: - vim is available on all UNIX system. - Its one of the rare non graphical merge tools.

dispatch: also a separate warning message on aliases with --config As mentioned in bug 2043, --config is also not supported in an alias. So report this the same way as the other "early" options. Example with alias.broken = stat --config a.config=1 Before: $ hg broken abort: Option --config may not be abbreviated! After: $ hg broken error in definition for alias 'broken': --config may only be given on the command line

extensions: remove erroneous comment We actually -do- use the 'ui' argument to print a debug statement.

blackbox: do not translate the log messages User 'timeless' in irc mentioned that having the blackbox be translated would result in logs that: - may be mixed language, if multiple users use the same repo - are not google searchable (since searching for english gives more results) - might not be readable by an admin if the employee is using hg in his native language And therefore we should log everything in english.

scmutil: split platform-specific bits into their own modules This parallels what's done for the util module, which imports either mercurial.posix or mercurial.windows as 'platform' and then slurps the appropriate functions into its own namespace.

backout: call cmdutil.commit directly instead of commands.commit This cleans up the messiness of having one command call another, and makes the backout command robust against changes to the commit command.

commit: factor out status printing into a helper function We create a new function commitstatus() in cmdutil that handles printing the status message(s) after a commit. This will allow other commit-like commands to use it, and in particular is step 2 towards removing backout's call to commands.commit.

backout: remove unnecessary dict copy This is step 1 to remove backout's call to commands.commit. We don't use the options again anywhere below except for backout's own purposes, specifically choosing a merge tool, so we just write the commit options in directly.

backout: remove unnecessary frobbing of addremove option There's no way for addremove to show up in backout's opts dictionary. It was being set manually because cmdutil.commit expected it to be there (and would throw an exception if it wasn't). This was fixed waaaaaaay back in: changeset: 5829:784073457a0f user: Kirill Smelkov <kirr@mns.spb.ru> date: Thu Jan 10 12:07:18 2008 +0300 summary: cmdutil.commit: extract 'addremove' from opts carefully

backout: use cmdutil.revert directly instead of commands.revert Before this change, backout would explicitly set the options it passed to commands.revert in order to fall thru most of its logic and call cmdutil.revert. This change makes it clearer what backup is trying to accomplish and makes it robust against changes to the revert command.

merge with stable

check-code: warn about line glob match with no glob character (?*/)

tests: remove glob from output lines containing no glob character

tests: inform on Windows about unnecessary glob lines When glob lines directly match on windows, "/" (and not "\") was output in the path on the line. No glob matching is necessary in this case. The test output will look like this (when 5 tests have passed and no 4 has an unnecessary glob): ... Info, unnecessary glob: info about some/thing (glob) ..

tests: quickly check if the glob line already matches the output This happens when a path with "/" as only glob char is matched on a non windows platform. (Currently one third of all glob matches.) The slowdown on windows and the speedup on other os are neglectable.

tests: append glob to filename output when required (windows) This lines were introduced in cd403d6d96ef and made the test fail on windows.

util: make ensuredirs safer against races

blackbox: only show new heads on incoming The blackbox was logging every head after every incoming group. Now we only log the heads that have changed. Added a test. Moved the hooks test to the bottom of the file since the hooks interfer with the tests after it.

blackbox: fix copyright

blackbox: fix a failing pyflakes test

blackbox: tests for the blackbox extension A few tests to cover the blackbox extension. Covers commands, hooks, and incoming changes.

blackbox: adds a 'blackbox' command for viewing recent logs Adds a 'hg blackbox' command for viewing the latest entries in the blackbox log. By default it shows the last 10 entries, but -l allows the user to specify.

blackbox: log incoming changes via ui.log() Logs incoming changes to a repo to ui.log(). Includes the number of changes and the hashes of the heads after the new changes. Example log line: 2013/02/09 08:35:19 durham> 1 incoming changes - new heads: cb9a9f314b8b Currently the blackbox logs the unix user that is performing the push/pull. It would be nice to log the http authorized user as well so it works with hgweb, but that's outside the scope of this commit.

blackbox: logs python and extension hooks via ui.log() Logs python and extension hooks to ui.log() for viewing in the blackbox. Example log lines: 2013/02/09 08:35:19 durham> pythonhook-preupdate: hgext.eol.preupdate finished in 0.01 seconds 2013/02/09 08:35:19 durham> exthook-update: echo hooked finished in 0.02 seconds

blackbox: log the commands that are run Uses ui.log to log which commands are run, their exit code, the time taken, and any unhandled exceptions thrown. Example log lines: 2013/02/09 08:35:19 durham> add foo 2013/02/09 08:35:19 durham> add exited 0 after 0.02 seconds Updates the progress tests because they use a mocked time.time() which these changes affect.

blackbox: adds a blackbox extension Adds a blackbox extension that listens to ui.log() and writes the messages to .hg/blackbox.log. Future commits will use ui.log() to log commands, unhandled exceptions, incoming changes, and hooks. The extension defaults to logging everything, but can be configured via blackbox.track to only log certain events. Log lines are of the format: "date time user> message" Example log line: 2013/02/09 08:35:19 durham> 1 incoming changes - new heads: d84ced58aaa

scmutil: create directories in a race-safe way during update With the new parallel update code, it is possible for multiple workers to try to create a hierarchy of directories at the same time. This is hard to trigger in general, but most likely during initial checkout. To deal with these races, we introduce a new ensuredirs function whose contract is to ensure that a directory hierarchy exists - it will ignore a failure that implies that the desired directory already exists.

Merge

test-atomictempfile: convert to unit test

tests: add a test runner utility that prints nothing when all tests pass This will be used to run tests through run-tests, which will expect no output for a unit test that passes successfully. The motivation for using unit tests instead of the current Python tests is that they don't require an output file for comparison and that they're easier to write because of the available tools from unittest (setup, asserts).

hgweb: consistent author name width Paper & coal use a sligtly shorter name length of 12em in the log view than gitweb's 15em. Names too long for 12em is not entirely unheard of, and hence increasing the length is appropriate; hgweb should remain usable on 1024x768.

dirstate: fix generator/list error when using python 2.7 util.statfiles returns a generator on python 2.7 with c extensions disabled. This caused util.statfiles(...) [0] to break in tests. Since we're only stat'ing one file, I just changed it to call lstat directly.

mq: comply with filtering when injecting fake tags (issue3812) mq was injecting fake tags whenever the revisions were accessible to the filtering level. This issue impacts hgweb since it's common to have "secret" mq patches. As secret changesets are filtered by hgweb, the tags computation could break.

commit: factor out post-commit cleanup into workingctx This pulls some of the logic for the cleanup that needs to happen after a commit has been made otu of localrepo.commit and into workingctx. This is part of a larger refactoring effort that will eventually allow us to perform some types of merges in-memory.

localrepo: use workingctx for validation in commit This changes localrepo.commit to use the workingctx it creates form the munged output of localrepo.status while running some precommit validation. Specifically, it uses functions that were already present on the workingctx. I believe this is a net readabilty improvement, and that this makes these lines consistent with the refactoring in a subsequent patch that pulls some of the validation logic into workingctx so that it can be reused elsewhere.

localrepo: create context used for actual commit earlier localrepo.commit creates a workingctx, calls self.status, does some munging on the changes status returns, does some validation on those changes, and then creates a new workingctx from the changes. This moves the creation of the new workginctx ahead of some validation, with the intention of refactoring some of that validation logic into the workingctx, so that it can be reused elsewhere.

merge with crew-stable

debugobsolete: improve command help The behavior without argument was not documented.

import: don't rollback on failed import --exact (issue3616) The checkexact() helper function was calling repo.rollback() from inside an open transaction. In addition to being insane, this is unnecessary because import will release the transaction on an exception. It turns out that this has been broken since the feature was first introduced, first released in v1.0: changeset: 4263:47ba52121433 user: Brendan Cully <brendan@kublai.com> date: Thu Mar 22 10:44:59 2007 -0700 files: mercurial/commands.py mercurial/patch.py description: Add import --exact. When this option is set, import will apply the patch (which must be generated by export) to the parents specified in the patch, and check that the node produced by the patch matches the node ID in the patch.

Merge

Merge

merge crew and main

merge crew and main

manifestmerge: fix order in which manifests are fetched If the manifest of an earlier revision on the same delta chain is read before that of a later revision, the revlog remembers that we parsed the earlier revision and continues applying deltas from there onwards. If manifests are parsed the other way round, we have to start over from the fulltext. For a fresh clone of mozilla-central, updating from 29dd80c95b7d to its parent aab96936a177 requires approximately 400 fewer zlib.decompress calls, which results in a speedup from 1.10 seconds to 1.05.

merge: run _forgetremoved after manifestmerge _forgetremoved can trigger manifest construction, but we only want it to happen after manifestmerge, so that our attempt to read the manifests in the right order in an upcoming patch actually works.

dirstate: disable gc while parsing the dirstate This prevents a performance regression an upcoming patch would otherwise introduce because it indirectly delays parsing the dirstate a bit.

export: show 'Date' header in a format that also is readable for humans 'export' is the official export format and used by patchbomb, but it would only show date as a timestamp that most humans might find it hard to relate to. It would be very convenient when reviewing a patch to be able to see what timestamp the patch will end up with. Mercurial has always used util.parsedate for parsing these headers. It can handle 'all' date formats, so we could just as well use a readable one. 'export' will now use the format used by 'log' - which is the format described as 'Unix date format' in the templating help. We assume that all parsers of '# HG changeset patch'es can handle that.

factotum: fix urllib2 import so it no longer relies on a demandimport bug demandimport will do that urllib2 can be imported 'from mercurial' even though it doesn't exist there.

hgweb: make the test suite use hgweb in a more WSGI compliant way - as checked by wsgiref.validate. This makes sure that we don't optimize hgweb for invalid use cases.

hgweb: simplify internal staticfile return codes

spelling: fix some minor issues found by spell checker

bundlerepo: replace basemap with the base field in the index Bundle revisions had some info in their fake revlog intries and some info in a dict with all the bundle revisions. This dict was used to get the stored data and to distinguish repo revisions from bundle revisions. Real repo revisions and bundle revisions will now be distinguished by comparing with the tip revision of the original repo. This reintroduces something similar to disktiprev which was unused and removed in a928865b4a4f and let that replace the O(reposize) dict.

profiling: replace '+' markup of nested lines with indentation The display of nested lines for hg --profile was very non-obvious and made it look like sort didn't work. The '+' immediately before CallCount was not related to the CallCount and did not mean plus in any integer sense. The '+' before module looked like a part of the module name and not like ascii art. Instead we now indent the subordinate module names to clearly show the structure.

Merge crew and main.

merge: don't fiddle with name lookups or i18n in hot loops We perform attribute dereferences and i18n lookups before looping.

merge: apply non-interactive working dir updates in parallel This has a big effect on the performance of working dir updates. Here are the results of update from null to the given rev in several repos, on a Linux 3.2 system with 32 cores running ext4, with the progress extension enabled. repo rev plain parallel speedup hg 7068089c95a2 0.9 0.3 3 mozilla-central fe1600b22c77 42.8 7.7 5.5 linux-2.6 9ef4b770e069 31.4 4.9 6.4

worker: allow a function to be run in multiple worker processes If we estimate that it will be worth the cost, we run the function in multiple processes. Otherwise, we run it in-process. Children report progress to the parent through a pipe. Not yet implemented on Windows.

worker: partition a list (of tasks) into equal-sized chunks

worker: estimate whether it's worth running a task in parallel Not implemented for Windows yet.

worker: count the number of CPUs This works on the major platforms, and falls back to a safe guess of 1 elsewhere.

tests: getremove test output changes (fold into previous patch)

merge: report non-interactive progress in chunks Instead of a monotonic count, getupdates yields the number of files it has updated since it last reported, and its caller sums the numbers when updating progress. Once we run these updates in parallel, this will allow worker processes to report progress less often, reducing overhead.

merge: handle subrepo merges and .hgsubstate specially In an upcoming patch, we will update .hgsubstate in a non-interactive worker process. Merges of subrepo contents will still need to occur in the master process (since they may be interactive), so we move that code into a place where it will always run in what will become the master process.

tests: update test output (will be folded into parent)

merge: split out mostly-non-interactive working dir updates In a later patch, we'll run these updates in parallel.

extensions: obsolete and remove interhg extension With the addition of the websub filter extension this extension is no longer needed. We maintain a sort of backwards compatibility by reading the [interhg] section and using it as we would use the [websub] section.

hgweb: apply the websub filter to revision descriptions In order to use this, add a [websub] section to your configuration and add websub expressions such as: italic = s/\b_(\S+)_\b/<i>\1<\/i>/ bold = s/\*\b(\S+)\b\*/<b>\1<\/b>/ issues = s|issue(\d+)|<a href="http://bts.example.org/issue\1">issue\1</a>|i bugzilla = s!((?:bug|b=|(?=#?\d{4,}))(?:\s*#?)(\d+))!<a href="http://bz.selenic.com/\2">\1</a>!i This also adds documentation (proofed by Kevin!) to the config help section.

hgweb: add websub template filter The purpose of this new filter is to make it possible to partially replace the functionality of the interhg extension. The idea is to be able to define regular expression based substitutions on a new "websub" config section. hgweb will then be able to apply these substitutions wherever the "websub" filter is used on a template. This first revision just adds the code necessary to load the websub expressions and adds the websub filter, but it does not add any calls to the websub filter itself on any of the templates. That will be done on the following revisions.

addremove: don't audit the path for paths already in the dirstate Now that dirstate.walk returns None for paths under symlink directories, addremove doesn't need to validate each path it sees to look for files under symlinks. On a large repository this brings addremove from 6.3 seconds down to 3.65 (42%) since addremove no longer has to stat every directory of every file to determine if the file is inside a symlink directory. I put it through our benchmark and see no perf hit to any other commands.

dirstate: walk returns None for files that have a symlink in their path Previously dirstate.walk would return a stat object for files in the dmap that have a symlink to a directory in their path. Now it will return None to indicate that they are no longer considered part of the repository. This currently only affects walks that traverse the entire directory tree (ex: hg status) and not walks that only list the contents of the dmap (ex: hg diff). In a situation like this: mkdir foo && touch foo/a && hg commit -Am "a" mv foo bar ln -s bar foo 'hg status' will now show '! foo/a', whereas before it incorrectly considered 'foo/a' to be unchanged. In addition to making 'hg status' report the correct information, this will allow callers to dirstate.walk to not have to detect symlinks themselves, which can be very expensive.

pathauditor: add check() method The pathauditor currently throws exceptions when it encounters an invalid path. This change adds a method to allow people to treat it as a boolean. This is currently used by scmutil.addremove and in a subsequent patch it will be used by dirstate.walk

summary: add missing space for updated active bookmark display

summary: show active bookmark even if not at current changeset Before this change, 'hg summary' would not show the active bookmark unless it pointed to the working directory parent. After this change, it will show it in parentheses, like so: parent: 18581:f0ff45fe6700 tip summary: simplify handling of active bookmark branch: default bookmarks: [crew] commit: (clean) update: (current)

summary: simplify handling of active bookmark

summary: test that current bookmark isn't shown This exposes the current behavior in a test. A later change will make summary show when the active bookmark has moved out from under us.

merge crew heads

merge with stable

outgoing: fix possible filtering crash in outgoing (issue3814) If there is no outgoiing changesets but we have filtered revision in outgoing.excluded We run into a filtering related crash. The excluded revision should not be there in the first place but discovery need cleanup in default, not stable.

test: display used python hash seed We keep using a random seed for each run, but we "compute" it ourself to be able to reproduce a failed test run.

outgoing: fix possible filtering crash in outgoing (issue3814) If there is no outgoiing changesets but we have filtered revision in outgoing.excluded We run into a filtering related crash. The excluded revision should not be there in the first place but discovery need cleanup in default, not stable.

parsedate: understand "now" as a shortcut for the current time

export: clobber files with -o (bc) (issue3652) This violated user expectation. Updated the code to clobber files, but preserve the behavior of appending multiple patches requested in a single export. Includes tests.

merge: don't call copies.mergecopies unless we need to This reduces the amount of time we spend calculating when doing a clean non-merge update. In a large repo, the time dropped from 10.1 seconds to 3.4.

merge: rename p1 to wctx in manifestmerge This is always a workingctx, and this name is more in line with other functions in this module.

largefiles: fix test and check code

histedit: always abort with same message when working dir is wrong. Using only one message makes live easier for translators. The clearer error message was supposed by Augie Fackler.

histedit: report when revisions to edit are not ancestors of working dir Editing the history only is possible when the working dir is a descendant of the revisions to edit. When this happens explain it by writing abort: %s is not an ancestor of working directory

Merge

manifestmerge: handle abort on local unknown, remote created files This replaces the _checkunknown call in calculateupdates with a more performant version. On a repository with over 150,000 files, this speeds up an update by 0.6-0.8 seconds, which is up to 25%. This does not introduce any UI changes. There is existing test coverage for every case, mostly in test-merge*.t.

manifestmerge: pass in branchmerge and force separately This will be used in an upcoming patch.

manifest: use a size 3 LRU cache to store parsed manifests Previously, the manifest cache would store the last manifest parsed. We could run into situations with operations like update where we would try parsing the manifest for a revision r1, then r2, then r1 again. This increases the cache size to 3 to avoid that bit of performance fragility.

util: add an LRU cache dict In certain cases we would like to have a cache of the last N results of a given computation, where N is small. This will be used in an upcoming patch to increase the size of the manifest cache from 1 to 3.

tests: remove last two check-code warnings about killdaemons Removes the last of the warnings in test-check-code-hg.t introduced in 667063b22a69.

tests: guard against obsolete markers in the hg repo itself If obsolete markers appear in the hg repo (because of enabling evolve), then tests that run hg against the hg repo itself will see warnings like: obsolete feature not enabled but 4 markers found! As far as I can tell, this only occurs in test-check-code-hg.t -- in particular, it will -not- show up on tests that run against test-created repos, as most of the test suite does.

largefiles: make caching largefiles message more explicit In some cases, caching largefiles may take a long time (if the user has pulled a lot of new heads). This patch makes it more clear what is happening, by showing the number of heads we are caching largefiles for.

largefiles: document behavior of caching largefiles for new heads

check-code: add Makefile target to run check-code

byterange: remove now-unused sys import

test-obsolete.t: migrate to killdaemons from kill `cat pidfile`

test-inotify.t: migrate to killdaemons from kill `cat pidfile`

test-inotify-lookup.t: migrate to killdaemons from kill `cat pidfile`

test-inotify-issue1556.t: migrate to killdaemons from kill `cat pidfile`

test-inotify-issue1542.t: migrate to killdaemons from kill `cat pidfile`

test-inotify-issue1371.t: switch to killdaemons from kill `cat pidfile`

test-inotify-debuginotify.t: migrate to killdaemons from kill `cat pidfile`

test-hgweb-raw.t: use killdaemons instead of kill `cat pidfile`

test-https.t: stop using kill `cat $pidfile`

changelog: hexlify node when throwing a LookupError on a filtered node The non-hexlified node was leaking all the way out to the web interface, and wasn't consistent with the behavior for nonexistent nodes.

byterange: remove old two-arg raise trick We don't care about the original traceback, and this was frustrating any efforts to make the code work on Python 3.3.

revlog: document v0 format

merge crew heads

merge with mpm stable

templater: add get() function to access dict element (e.g. extra)

templates: export extra as a dict to templates Currently only useful with the json filter.

Added signature for changeset 7511d4df752e

Added tag 2.5.1 for changeset 7511d4df752e

merge with i18n

i18n-de: fix duplicated shortcut key b for record (edit and quit) Use e for edit instead.

check-code: fix platform-specific error code variance test-check-code-hg.t uses xargs to invoke check-code.py on every file in 'hg manifest'. The return code from xargs varies between BSD xargs and GNU xargs: BSD will return 1 if any invocation exits with an error code; GNU xargs will return 123 in this case. This normalizes the exit code back to 1.

check-code: warn to use killdaemons instead of kill `cat PIDFILE` We have a bunch of tests that still use kill `cat hg.pid` or worse, kill `cat hg.pid`; while kill -0 `cat hg.pid`; sleep 0; done Cleaning these up to use tests/killdaemons.py is non-trivial, so for now we just add a warning.

largefiles: docstrings for verify methods

largefiles: fold oddly named _verify into remotestore.exists

git convert: some versions of git use fatal: instead of error: I saw this behavior with git 1.7.12 on my Mac.

test-convert-git: stabilize corruption of test git repo This error would show up only intermittently since the test depended on the order of the directories returned by os.walk. The damage repository test would delete the first object file it came across. However, the order of the directory listing is arbitrary (it seems to depend on the filesystem). This meant that sometimes a commit object was deleted, sometimes a blob object and sometimes a tree object. So, fix by hardcoding which object to delete. Delete a commit object, a blob object and a tree object in three separate tests.

convert/git: catch errors from modern git-ls-remote (issue3428) Since git v1.7.8.2-327-g926f1dd (the change was first released in git 1.7.10), git does not return non-zero when "git ls-remote --tags ..." is run and the repository is damaged. This causes the "damaged repository with missing commit" test in test-convert-git.t to unexpectedly succeed. Fix by aborting if git outputs any lines beginning with "error:", which required adding some subprocess use in convert/git.py.

merge with stable

incoming: fix incoming when a local head is remotely filtered (issue3805) In its current state discovery may return (remotely) filtered elements in "common". This has usually no impact as "missing" is kept clear of filtered elements. However when the "remote" repo is a local repo (disk accessible, and directly created in memory) the incoming code takes a shortcut and directly uses the "remote" repo to generate the incoming output. When some common elements are filtered this led to a crash. We now ensure we use an unfiltered repository to generate the incoming output. This does not change the behavior as missing is clear of filtered revision. Now that we have proper low level filtering, incoming code needs a deeper cleanup but it is already planned.

dirstate: move pure python dirstate packing to pure/parsers.py

bookmark: don't allow integers as bookmark/branch/tag names Bookmarks/branches/tags shouldn't be allowed to be integers because that overlaps with revision numbers. Right now if a user created one they can't use it anyway because the revision numbers take precedence. The check only happens when creating a new bookmark/etc from a command so it shouldn't affect existing bookmarks/branches/tags or importing branches from git. This fix was prompted by us having a user create a bookmark named "404" then accidentally checkout a very old version of our repository.

run-tests: do not fail on empty tsttest file Initialize n for not failing on empty tsttest files.

merge with stable

hgweb: make 'summary' work with hidden changesets (issue3810) Since the 'summary' view used by e.g. gitweb and monoblue shows both a changelog and a bookmarks list, the same changes are needed here as were made to the 'changelog' and 'bookmarks' web commands (56ca4443a343 and 886936ecc21b, respectively).

merge with stable

hgk: support the old way of getting the current Ttk theme (issue3808) It seems like the API has changed somewhere around 8.5.7, so the preferred way of getting the current theme is now [ttk::style theme use], while the deprecated (but still working) is $::ttk::currentTheme.

addremove: don't perform m.exact/rel until needed Moves the m.exact and m.rel calls within the conditionals they are used in. On a large repo this brings addremove from 7.1 seconds down to 6.3 (13%).

addremove: don't call lexists, isdir, and islink The dirstate walk results contain the stat information for each path, so we don't need to query it again. On a large repo this makes addremove go from 8.35 seconds to 7.1 (15%).

addremove: only query dirstate once per path Previously the addremove code queried the dirstate 4 times per path. Now it only does so once. On a large repo this brings addremove from 9.5 seconds to 8.35 seconds (12%).

tests: fix windows test failure with largefiles

merge with stable

tests: fix windows test failure in test-subrepo.t

merge with stable

largefiles: don't crash when trying to find default dest for url without path b32e55e6c3c7 introduced a crash when cloning a url without path - where util.url().path would be None. This None will now be handled as ''. clone will thus abort with 'repository / not found' as before.

hgweb.cgi: fix internal WSGI emulation (issue3804) The internal WSGI emulation in wsgicgi.py was not fully WSGI compliant and assumed that all responses sent a body. With 3fbdbeab38cc that caused a real bug when using hgweb.cgi. wsgicgi.py will now make sure headers always are sent, using the pattern from PEP 333 and similar to how it is done in c007e5c54b16.

bdiff: simplify overflow checks Rather than check that each delta start, end, and length is within 32 bits, we simply check that the input strings are under 4GB.

Merge crew-stable into crew

rebase: derive node from target rev (issue3802) dest.rev() is the same as target when a new rebase is run, but dest isn't set when rebase --continue is run. Bug introduced in 2a1fac3650a5, which fixed issue3685.

profiling: introduce limit configuration option The number of output lines was hardcoded to 30. There was a 'nested' configuration options that controlled something else related to counting the number of output lines. This introduces the profiling.limit configuration option for controlling the number of profiling output to show.

largefiles: make verify --lfa and --lfc work without --large The slightly obscure --lfa and --lfc only worked as modifiers to --large and could be combined. The documentation was however not clear what they did. Instead they now imply --large and the description is updated.

largefiles: verify status should be written as status, not as write Make 'hg verify --large -q' quiet when no errors.