merge with stable

grep: only catch re.error when compiling regular expressions

doc: clarify smtp.password usage

merge with stable

log: include unmodified-in-merge files in log diff/stat (issue2383) f786fc4b8764 assumed that walkchangerevs called prep with all relevant matched filenames, but actually it only contains the names of files changed in the relevant changeset. That meant that log diff/stat of merges missed the diff for files only changed in the other branch. This is a minimal fix for making sure we only use fns when we are following and thus will have problems with merges anyway ...

tests: fix unzip -l variability fix 01-01-1980 slipped through the grep.

merge with stable

tests: fix unzip -l variability

mq: always require --force when pushing patches (issue2363) --force was not necessary when passing --rev since 55578a8d7e84, but this behaviour is usually harmful when branch names are passed instead of explicit revisions.

tests: various fixes for new unified test pattern format

tests: add glob matching for unified tests This adds a " (glob)" marker that works like a simpler version of (re): "*" is converted to ".*", and "?" is converted to ".". Both special characters can be escaped using "\", and the backslash itself can be escaped as well. Other glob-style syntax, like "**", "[chars]", or "[!chars]", isn't supported.

tests: require regexes in unified tests to be marked with " (re)" Consider this test: $ hg glog --template '{rev}:{node|short} "{desc}"\n' @ 2:20c4f79fd7ac "3" | | o 1:38f24201dcab "2" |/ o 0:2a18120dc1c9 "1" Because each line beginning with "|" can be compiled as a regular expression (equivalent to ".*|"), they will match any output. Similarly: $ echo foo The blank output line can be compiled as a regular expression and will also match any output. With this patch, none of the above output lines will be matched as regular expressions. A line must end in " (re)" in order to be matched as one. Lines are still matched literally first, so the following will pass: $ echo 'foo (re)' foo (re)

tests: ensure regexes match to the end of the string Regular expressions in the test suite are currently written assuming that you need a trailing ".*" to avoid matching to the end. Instead of matching regular expressions using "^pattern", this patch makes matching more restrictive by matching "^pattern$".

tests: don't match blank output lines as regexes in unified tests Currently, the following unified test will pass: $ echo foo A blank output line (a line containing just two spaces) will match any output. The patch modifies the unified test runner to ignore empty strings strings when do regular expression matching.

tests: improve regexes in unified tests

templater: add hex filter.

tests: unify test-convert-svn-*

merge: delay writing the mergestate during until commit is called This speeds up merges of lots of files considerably.

tests: fix touch -t with RHEL

tests: compatibility fix. 'touch -d *' doesn't work on Mac OS X. In order to prevent this from occuring again, a check for the case is added to 'check-code.py'.

tests: drop a bunch of sed calls from unified tests

tests: cleanup exit code handling in unified tests

check-code: add some basic support for unified tests

test-subrepo-recursion: sort archive output for portability

Merge with crew

Merge with stable

context: fix filectx.undelete() (issue2388)

tests: fix hgweb template change

Merge with crew

Restore lexists() changes lost in e0ee3e822a9a merge

tests: fix hgweb template change

merge with stable

merge with crew

hgweb: be consistent with the background highlighting on closed branches (Issue2386) The coal and paper web themes will highlight closed branches as a visual cue to the user. This patch makes the other two themes constistent. For users with difficulty differentiating colors, it would be better to split out closed branches to a new table or optionally hide them.

mq: fix the deprecation comment for qsave & qrestore. Changeset bcf90e712dc3 deprecated qsave and qrestore. In the deprecating comment, users were referred to 'rebase --mq' which -- at the time the message was written -- didn't exist. Currently, on the default branch, rebase *does* take a '--mq' option, but it probably doesn't do what Dirkjan expected it to do when he wrote the message. In the original, deprecating commit, little context was provided as to why this change was made. Based on my recollection, concensus at the Paris Sprint in February 2010 was that one of the problems with MQ was that it exposed far too many commands. Notable among these were qsave & qrestore: very few core developers understood what they did and even fewer (none, IIRC) actually used them. However, they couldn't be removed; not only do the usual backwards compatibility reasons apply, but the hg book refers to them.

revsets: reduce cost of outgoing in the optimizer

test-convert-hg-startrev: enable convert.hg.saverev.

transplant: fix var name conflict introduced by 2912881c2a98

bundle: fix shrink-revlog bundle usage

bundlerepo: restore close() method

tests: drop final true command from unified tests

Merge with stable

Use lexists() instead of exists() where appropriate

convert/svn: fix broken symlink renames in svn sink

rename: do not overwrite existing broken symlinks

patch: do not overwrite broken untracked symlinks

patch: fix target when patching broken symlinks (issue2368)

tests: fixup tests using unzip -l

tests: add hack to avoid problem with graphlog in unified tests

bundle: encapsulate all bundle streams in unbundle class

bundle: move chunk parsing into unbundle class

bundle: get rid of chunkiter

bundle: refactor progress callback This lets us eliminate the chunk iterator so we can have the bundle class parse headers.

bundle: make getchunk() a method

bundlerepo: use bundle objects everywhere

bundlerepo: drop unused close() method

bundle: make unbundle object seekable This is only for uncompressed bundles used by bundlerepo.

bundle: push chunkbuffer down into decompress We replace the fixup iterator with a file-like object so that uncompressed file streams can be passed end to end through the stack.

tests: fix a bunch of pointless #s in unified tests

tests: exit 0 no longer needed on unified tests

tests: update debugcomplete output

merge with abuehl

tests: unify some of test-mq*

subrepo: add support for 'hg archive'

subrepo: introduce files and filedata methods for subrepo classes

merge with stable

revset: handle re.compile() errors in grep() Raise error.ParseError instead of allowing re.error to bubble up.

archive: set date to 1980 for very old zip files The zip file format stores the date using "MS-DOS format" which apparently means that they use 1980 as their epoch. Python's zipfile module emits deprecation warnings of this form /usr/lib/python2.6/zipfile.py:1108: DeprecationWarning: struct integer overflow masking is deprecated self.fp.write(zinfo.FileHeader()) /usr/lib/python2.6/zipfile.py:1108: DeprecationWarning: 'H' format requires 0 <= number <= 65535 self.fp.write(zinfo.FileHeader()) /home/mg/src/mercurial-crew/mercurial/archival.py:169: DeprecationWarning: struct integer overflow masking is deprecated self.z.close() /home/mg/src/mercurial-crew/mercurial/archival.py:169: DeprecationWarning: 'H' format requires 0 <= number <= 65535 self.z.close() when it is given such old timestamps. This fixes this by silently clamping the date to 1980.

Merge with stable

test-eol-update: record new results as correct Changeset 0852da25a31b changed the result of this test. The 'hg update 0' command, which causes a merge of modified a.txt, now leaves a.txt in the EOLN format specified by .hgeol as it was committed in revision 0. Previously, it used the .hgeol contents from the working directory before the update.

tests: add exit codes to unified tests

merge with stable

merge: suggest 'hg up -C .' for discarding changes, not 'hg up -C' Without specifying the parent revision of the working copy, users will update to tip, which is most likely the other head they were trying to merge, not the revision they were at before the merge.

transplant: maintain list of transplants in dict

tests: unify test-hgignore

win32: update build instructions with correct CRT version As stated in http://msdn.microsoft.com/en-us/library/cc664727.aspx, when you compile an application with MSVC 2008 SP1 it is bound by default to the original CRT version (9.0.21022.8). This is the case for Python 2.6 up to 3.1. If the wrong CRT version is embedded in the Inno Setup installer, with a PC that does not have the MSVC 2008 redistributable package installed, hg will refuse to launch with an error: "the system cannot execute the specified program".

merge with crew

merge with crew

Merge with stable

eol: setup the repo.ui in reposetup() This fixes problems some have reported using the eol extension together with TortoiseHg. It ensures the ui associated with the repository has the preupdate.eol hook configured.

tests: unify test-hardlinks-safety

tests: unify test-extra-filelog-entry

tests: unify test-extdiff

tests: add test for bookmark push/pull

bookmarks: set VALUE in push/pull

tests: unify test-username-newline

tests: unify test-url-rev

tests: unify test-update-renames

tests: unify test-rename-merge1

merge with crew

clone: only use stream when we understand the revlog format This patch fixes issues with stream cloning in the presense of parentdelta, lwcopy and similar additions that change the interpretation of the revlog format, or the format itself. Currently, the stream capability is sent like this: stream=<version of changelog> But the client doesn't actually check the version number; also, it only checks the changelog and it doesn't capture the interpretation-changes and flag-changes in parentdelta and lwcopy. This patch removes the 'stream' capability whenever we use a non-basic revlog format, to prevent old clients from receiving incorrect data. In those cases, a new capability called 'streamreqs' is added instead. Instead of a revlog version, it comes with a list of revlog-format relevant requirements, which are a subset of the repository requirements, excluding things that are not relevant for stream. New clients use this to determine whether or not they can stream. Old clients only look for the 'stream' capability, as always. New servers will still send this when serving old repositories.

localrepo: factor out requirement application and write

merge with stable

merge with i18n

merge with i18n

l10n: starting to unify the old translations New status: 1525 messages complete with 135 fuzzies and 621 untranslated. Transmitted-via: Transifex (www.transifex.net).

l10n: translations: env. variables / templating system (DE) New status: 1517 messages complete with 134 fuzzies and 630 untranslated. Transmitted-via: Transifex (www.transifex.net).

l10n: added some new translations New status: 1443 messages complete with 130 fuzzies and 708 untranslated. Transmitted-via: Transifex (www.transifex.net).

i18n-pt_BR: synchronized with 0f6164806283

tests: unify test-default-push

tests: unify test-clone-update-order

tests: unify test-debugrename

tests: unify test-confused-revert

tests: unify test-command-template

mq: rename strip --nobackup option to --no-backup (issue2377) Same option name as for hg revert --no-backup. Deprecate strip --nobackup.

mq: consistently use opts.get() to prevent potential KeyError

tests: remove exec bit from .t files

combine tests

merge with stable

color: add win32 support for non-black background This complicates the code a bit, since attributes need to be masked together correctly before they are applied. Perhaps the code should be redesigned at some point, but this works well for now.

i18n: unmark untranslatable string

test-subrepo-recursion: test missing subrepo

incoming: recurse into subrepositories with --subrepos/-S flag As with push and outgoing, the optional source path is ignored for the subrepositories. Fixing this is Issue1852.

incoming: move code from commands to cmdutil This makes it easier to reuse it in a recursive fashion for subrepos.

outgoing: recurse into subrepositories with --subrepos/-S flag As with push, the optional destination path is ignored for the subrepositories, i.e., they are always compared with their default push path. Fixing this is Issue1852.

outgoing: move code from commands to cmdutil This makes it easier to reuse it in a recursive fashion for subrepos.

add: recurse into subrepositories with --subrepos/-S flag

add: move main part to cmdutil to make it easier to reuse

narrowmatcher: propagate bad method The full path is propagated to the original match object since this is often used directly for printing an error to the user.

narrowmatcher: fix broken rel method

patch: break import cycle with cmdutil The patch module imported cmdutil but used it only in updatedir.

patchbomb: show progress when sending emails or writing mbox

patchbomb: consistent code style in getaddrs()

win32: 64-bit Inno Setup installer

win32: update Inno Setup build instructions with up-to-date software

merge with stable

tests: unify test-merge-remove

tests: unify test-merge-prompt

tests: unify test-merge-local

tests: unify test-merge-internal-tools-pattern

tests: unify test-merge-force

tests: unify test-merge-default

tests: unify test-merge-commit

tests: unify test-manifest*

convert/darcs: handle non-ASCII metadata in darcs changelog (issue2354) Given a commit author or message with non-ASCII characters in a darcs repo, convert would raise a UnicodeEncodeError when adding changesets to the hg changelog. This happened because etree returns back unicode objects for any text it can't encode into ASCII. convert was passing these objects to changelog.add(), which would then attempt encoding.fromlocal() on them. This patch ensures converter_source.recode() is called on each piece of commit data returned by etree. (Also note that darcs is currently encoding agnostic and will print out whatever is in a patch's metadata byte-for-byte, even in the XML changelog.)

merge with crew

tests: fix incompatibility with python-2.4 in test-hgweb Thanks to lcantey for catching it.

merge with stable

merge with i18n-default

merge with i18n-stable

i18n-pt_BR: fix a couple of typos

i18n-pt_BR: added messages from crew stable up to 26d86a6cf2af

i18n-ro: merged default

i18n-ro: ongoing translation

i18n: merged in strings from hg cset ce818cf215dc

i18n-ro: ongoing translation

i18n-ro: ongoing translation

i18n-ro: ongoing translation

i18n-ro: ongoing translation

i18n-ro: ongoing translation

i18n-ro: ongoing translation

Continued translation

i18n-ro: ongoing translation

i18n-de: fix awkward translation of "memory"

i18n-da: fix typo

i18n: merge with stable

i18n: merge default heads

i18n-pt_BR: synchronized with 9617803b1acb

i18n: merge with stable

i18n: merge default heads

i18n-pt_BR: synchronized with c6b1be675d3c

i18n: merge with stable

i18n: merge with stable

i18n: merged stable heads

i18n-it: finish translating 'hg summary' output

i18n-it: remove bogus fuzzy translations for revset The translations were all random fuzzy strings, which can only do harm, so remove them. Moreover, given the status of revset (it seems mostly an internal undocumented module, and to understand it you need the source anyway), I prefer not to translate it for now.

i18n-it: update fuzzy messages/translate patchbomb

i18n-it: update mq translation Significant updates were performed - when translating something, I tried to choose a term and use it consistently everywhere possible, so here I translated a lot of messages. Additionally, I use the mq extension a lot, and now I am using the translation, so I will be able to translate it better.

i18n-it: updates after msgmerge on stable branch, 2nd part

i18n-it: general updates fuzzy synopsis texts This patch follows the fix on the translation of [OPTION]... with other needed updates on the same class of messages.

i18n-it: translate [OPTION]..., not [OPTIONS]...: keep the singular Since [OPTION]... is intended to mean more options, it was translated in Italian as plural. However the plural is conveyed by `...', so switch to the singular. This was also useful because for the '-o OPT' extdiff option, I have to translate `OPT' as a singular.

i18n-it: readd other lost accented characters The problem was introduced already in revision 9f9137cd83f5 by the original Italian translator, according to logs. This might be due to a bug in the program he used, "Lokalize 0.3", or maybe not - some accented characters exist also in that commit.

i18n-it: updates after msgmerge on stable branch

i18n-it: perform msgmerge on the tip of the stable branch

i18n-pt_BR: synchronized with 31dde4c3bb83

dirstate: use one pass to filter out files in subrepos

subrepo: handle diff with working copy We cannot unconditionally call bin(node2) since node2 is None when comparing with the working copy.

subrepos: handle diff nodeids in subrepos, not before Subversion nodeids are integer revisions.

mq: add newline after qfold warning

merge with stable

tests: unify test-locate

tests: unify test-journal-exists

tests: unify test-issue1438 and test-issue2137

keyword: explain file-wise expansion in help

convert: move -A option to --authormap

patchbomb: add --confirm option to show series details and ask for confirmation --confirm presents same prompt as --diffstat, but does not write a diffstat to the messages' bodies. A simple test simulating a negative response is included.

patchbomb: let diffstat prompt only once with complete summary This changes the behaviour of --diffstat. Before the user was asked for confirmation of each patch with its description and diffstat, and a final summary. Now there is only one prompt right before sending with a final summary which does not include the patch descriptions, but the message details and the diffstats: Final summary: From: sender To: recipient(s) Cc: (if present) Bcc: (if present) Reply-To: (if present) Subject: [patch 0 of x [flags]] intro (if present) a | 28 ++++++++++++++++++++++++++++ b | 15 +++++++++++++++ Subject: [patch 1 of x [flags]] subject a | 28 ++++++++++++++++++++++++++++ [ ... ] are you sure you want to send (yn)?

patchbomb: rename argument to avoid shadowing patch module

convert: deprecate --authors in preference for --authormap This aligns the authormap option with the other three mapping options. The old --authors option is still supported and 'hg help convert -v' will still show it.

patchbomb: show prompt and selection in non-interactive mode Akin to 01ada7b1861d for the custom patchbomb prompt.

patchbomb: use ui.promptchoice for diffstat to allow localization of choices The extra check for ui.interacive from patchbomb's prompt function is not needed here. Format boolean prompt as in filemerge.py.

tests: unify test-issue*

tests: unify test-globalopts

tests: unify test-git-*

tests: unify test-filebranch

tests: unify test-extension

tests: unify test-execute-bit

tests: unify test-empty-group

merge with stable

convert: better grouping of command line flags in help We normally put related command line flags after one another.

convert: help string cleanups

convert: better quoting in help text

convert: show example splice, author, and branch map entries in help Also document that - empty lines are skipped and comment are supported in author map - whitespace is not allowed in branch map entries since we split on it when parsing the file

hgweb: support very simple caching model (issue1845)

tests: extend get-with-headers to support cache testing

tests: fix test output for test-hgweb-no* tests

tests: rename test for better grouping

check-code: fix ternary expression introduced in 1ef9b0b5efec

bisect: warn about ending on a merge with only one side marked

bisect: allow revsets in addition to single revs (issue2360)

subrepos: add function for iterating over ctx subrepos

subrepos: handle modified but uncommitted .hgsub

localrepo: check nested repos against working directory This checks against what is really in the filesystem, including subrepositories that have been added to .hgsub but not yet committed.

add: clarify that repo[None].add returns rejected filenames

merge with stable

store: Removed extra space in _data list

verify: fix "missing revlog!" errors for revlog format v0 and add test With revlog format v0 the .d files are empty if the only revision stored is an empty file. Since Mercurial can no longer create format v0 repositories, but still use it, add a script which creates a repository with a single empty file. This can be used in other tests if wanted.

zsh completion: bookmarks support

zsh completion: patchbomb support

diff: recurse into subrepositories with --subrepos/-S flag

status: recurse into subrepositories with --subrepos/-S flag

match: add narrowmatcher class This class can be used to adapt an existing match object to a new match object that only cares about paths within a certain subdirectory.

cmdutil: use repo.auditor when constructing match object This gives the repository control over which nested repository paths that should be allowed via the custom path auditor. Since paths into subrepositories are now allowed, dirstate.walk must now filter away more paths than before.

match: accept auditor argument This is used when normalizing filenames and patterns.

localrepo: add auditor attribute which knows about subrepos

tests: remove unneeded -u flags

merge with stable

tests: unify test-empty-file

tests: unify test-empty-dir

tests: unify test-dispatch

tests: remove unneeded -d flags Many tests fixed the commit date of their changesets at '1000000 0' or similar. However testing with "Mon Jan 12 13:46:40 1970 +0000" is not better than testing with "Thu Jan 01 00:00:00 1970 +0000", which is the default run-tests.py installs. Removing the unnecessary flag removes some clutter and will hopefully make it clearer what the tests are really trying to test. Some tests did not even change their output when the dates were changed, in which case the -d flag was truly irrelevant. Dates used in sequence (such as '0 0', '1 0', etc...) were left alone since they may make the test easier to understand.

test-merge5: removed unnecessary parts of test - no need to initialize a new directory - no need to fix the commit time, already done by run-tests - no need to exit 0, use exit code from hg instead - better spacing around comments

test-merge5: ensure updating to tip across branches is tested When doing "hg update" across branches (without any revision specified), hg will abort and ask the user to merge or do hg update -c. This tests for that error.

test-convert-cvs: add a sleep to make test more stable The recent addition of fuzzy tests introduced a new cvs commit which sometimes fails. This adds a sleep to make sure that cvs notices that the file has changed, similar to how it is done in other tests.

merge with stable

tests: unify test-diff-binary-file

tests: unify test-diff-copy-depth

tests: unify test-diff-upgrade

tests: unify test-convert-hg-startrev.

tests: unify test-diffstat