scmutil: speed up new-style range extension This improves perfrevset performance by 10% for 2222::33333 on a kernel repo.

merge with crew-stable

subrepo/svn: make rev number retrieval compatible with svn 1.5 (issue2968) ae2664ee0223 introduced "svn info TARGET@REV" to determine if a certain path exists in the specified revision, but in svn 1.5 the error message "Not a valid URL" yields exit code 0 so the error is not caught. Use "svn list TARGET@REV" instead which works with svn 1.5 and is even faster in some situations.

convert/bzr: make tests work with bzr 2.0 again bzr ci --commit-time was introduced in 2.1 bzr add without -q prints "adding" in 2.1, "added" in 2.0

tests: cleanup of svn url handling The subversion tests used different tricks to create properly encoded URLs, partly due to partial support for different ways of running the tests on windows. Now we only need/support one way of running the tests on windows. Windows URLs should look like 'file:///c:/foo%20bar' and on Unix platforms like 'file:///tmp/baz'. 'pwd' in the test framework will on Windows emit paths like 'c:/foo bar'. Explicit handling of backslashes in paths is thus no longer needed and is removed. Paths on windows do however need an extra '/' compared to other platforms. This change makes test-subrepo-svn.t pass on windows with msys. Other tests might need more work.

test-subrepo-svn.t: partially adapt for Windows hg forget 'notafile*' is changed to use a name that is valid on Windows so we still get the same error ... but the error message is disabled because it varies with the Windows version.

convert: ignore svn:executable for subversion targets without exec bit support Calling propset/propdel with subversion 1.6 on FAT gave abort: svn exited with status 256 and made test-convert-hg-svn.t and test-convert-svn-sink.t fail. 1.7 worked. This is a rework of fdee5d614fcc but ignores the executable bit when it isn't supported instead of using an approximation.

tests: add missing 'backup bundle' path glob in test-rebase-collapse.t Introduced by c8eda7bbdcab.

tests: add missing 'backup bundle' path glob in test-rebase-scenario-global.t Introduced by 50f434510da6.

merge with crew

merge with stable

rebase: improve error message on improper phases The previous error message had two issues: The first issue was that it wasn't, in fact, an error but a warning, even though it described a fatal error condition preventing the successful completion of the command. The second was that it didn't mention the immutable changesets, leaving the user guessing at the true cause of the error. The main downside to this change is that we now get an 'abort: can't abort...' message which technically contradicts itself. In this case, I blame that on the two uses we have for the word; if it weren't for backwards compatibility, we could make util.Abort print out 'error: <whatever>'.

subrepo: support Git being named "git.cmd" on Windows (issue3173) Popen does not consider "foo.cmd" equivalent to "foo" on Windows. Unfortunately, the default MSYS Git installation installs only "git.cmd" into the path by default. This patch probes for both possible names on Windows.

subrepo: warn user if Git is not version 1.6.0 or higher

test-http and test-https: partially adapt for Windows

test-subrepo-git.t: adapt for Windows hg forget 'notafile*' is changed to use a name that is valid on Windows so we still get the same error ... but the error message is disabled because it varies with the Windows version.

test-bookmarks-pushpull: partially adapt for Windows The kill call at the end is redundant, as we already have 199: $ hg serve -p $HGPORT -d --pid-file=../hg.pid -E errors.log 200: $ cat ../hg.pid >> $DAEMON_PIDS So there is nothing left that would not already be killed by the $DAEMON_PIDS mechanism.

test-static-http: partially adapt for Windows

test-bad-pull: partially adjust for Windows On a English Windows 7, the testcase fails with $ hg clone http://localhost:$HGPORT/ copy abort: error: No connection could be made because the target machine actively refused it Since the error message on non-English Windows installs are most likely different, we have to glob the entire error message away for Windows.

tests: consistently use printenv.py the same MSYS/Windows-compatible way This will remove some reasons some tests are disabled on Windows.

tests: prepare get-with-headers.py for MSYS get-with-headers.py took the http GET parameter as a command line parameter that had to start with '/'. MSYS on windows will mangle such paths. Instead of applying a workaround everywhere (such as an extra '/') we let get-with-headers.py add the mandatory '/'. That is consistent with the url path handling in the Mercurial url class. A few tests sent 'GET ?cmd=...' which is invalid. They will now send 'GET /?cmd=...'. This will not enable any tests for being run on windows - only remove one reason they were disabled.

tests: hghave outer-repo should be true even if a bad repo is found Only the abort message 'no repository found' is a reliable indication that there is no outer repo.

tests: add missing no-outer-repo requirements Outer repos (if any) will now never be touched by the tests. But it is better to run without any repos around the tmp directory.

tests: convert some 'hghave no-outer-repo' to #if In some places we instead use 'hg init' to create a well-known outer repo without any special extensions or other config settings.

strip: incrementally update the branchheads cache after a strip This function augments strip to incrementally update the branchheads cache rather than recompute it from scratch. This speeds up the performance of strip and rebase on repos with long history. The performance optimization only happens if the revisions stripped are all on the same branch and the parents of the stripped revisions are also on that same branch. This adds a few test cases, particularly one that reproduces the extra heads that mpm observed.

localrepo: convert _updatebranchcache from nodespace to revspace _updatebranchcache used to use revlog.reachable. After the switch to revlog.ancestors, we can now clean it up a bit and switch the algorithm from nodes to revs.

discovery: convert legacy part of checkheads to revs from nodes After a recent switch from revlog.reachable to revlog.ancestors, clean up this old call site, switching it from nodes to revs.

transplant: convert applied() algorithm from nodes to revs After a recent switch from revlog.reachable to revlog.ancestors, clean up the code a bit to use revs rather than nodes.

revlog: remove reachable and switch call sites to ancestors This change does a trivial conversion of callsites to ancestors. Followon diffs will switch the callsites over to revs.

revlog: add incancestors, a version of ancestors that includes revs listed ancestors() returns the ancestors of revs provided. This func is like that except it also includes the revs themselves in the total set of revs generated.

scmutil: change canonpath to use util.samefile (issue2167) Fixes (on Windows in cmd.exe): $ hg -R v:\x\a status V:\x\a\bar abort: V:\x\a\bar not under root where v:\x\a is a valid repository with a checked-out file "bar" (Note the difference in casing: "v:\" versus "V:\")

win32.py: let samefile and samedevice work on directories too

rebase: do not add second parent to rebased changeset (drop detach option) (BC) Rebase now behaves as if --detach was always passed. Non-merges are rebased as non-merges, regardless of their parent being an ancestor of the destination. Merges will usually be rebased as merges unless both of their parents are ancestors of the destination, or one of their parents is pruned when rebased. This only alters the behavior of rebase when using the --source/--rev options. --detach option is deprecated. All test changes were carefully validated.

zsh completion: catch up with 2.2 commands and options * add completion for phase and graft core commands * add completion for rebase extension * add new options for already defined commands: * commit --amend * import --bypass * manifest --all * merge --tool * revert -C (short for --no-backup) * qpush --exact [mq] * email --body [patchbomb]

revset: add a utility for obtaining the source of a given rev graft, transplant and rebase all embed a different type of source marker in extra, and each with a different name. The current implementation of each is such that there will never be more than one of these markers on a node. Note that the rebase marker can only be resolved if the source is still present, which excludes the typical rebase usage (without --keep) from consideration (unless the resulting bundle in strip-backup is overlayed). There probably isn't any reason to use rebase --keep as a substitute for transplant or graft at this point, but maybe there was at one point and there are even a few rebases in the hg repo, so it may be of historical interest.

revset: add a predicate for finding converted changesets This selects changesets added because of repo conversions. For example hg log -r "converted()" # all csets created by a convertion hg log -r "converted(rev)" # the cset converted from rev in the src repo The converted(rev) form is analogous to remote(id), where the remote repo is the source of the conversion. This can be useful for cross referencing an old repository into the current one. The source revision may be the short changeset hash or the full hash from the source repository. The local identifier isn't useful. An interesting ramification of this is if a short revision is specified, it may cause more than one changeset to be selected. (e.g. converted(6) matches changesets with a convert_revision field of 6e..e and 67..0) The convert.hg.saverev option must have been specified when converting the hg source repository for this to work. The other sources automatically embed the converted marker.

merge with stable

merge with i18n

i18n-de: More corrections for translation of default, and other translations

i18n-de: translate default as Vorgabe

i18n-de: Update strings, translated many fuzzy strings

i18n-pt_BR: synchronized with 0fdd8193c8b5

merge with i18n

i18n-de: More corrections for translation of default, and other translations

tests: remove execute permission from two unified test files

graft: allow -r to specify revisions

consistency: use REV instead of REVISION

export: add optional -r to synopsis It is not already covered by [OPTION]..., because in this case REV... would need to be marked optional.

strip: add optional -r to synopsis

qimport: use [FILE]... because if -r is used no file is needed

mq: abort if no files or revisions are specified for qimport

tests: convert some hghave unix-permissions to #if

tests: run test-walk.t on windows A couple of detours and changed paths are used to work around msys path conversion.

tests: better testing of 'glob:glob' in test-walk.t

tests: move tests in test-walk.t using ':' in filenames to conditional section eol-in-paths will also protect against testing ':' in filenames on FAT.

tests/printenv.py: eliminate trailing spaces on output

tests: fix test-i18n.t after translation update

merge with i18n

Corrected hyphenation, translated subrepo help and other strings

i18n-de: translate default as Vorgabe

i18n-de: Update strings, translated many fuzzy strings

i18n-ru: synchronized with 2255950e1f76

tests: remove hghave system-sh from test-pending.t The alternative would have been to add a unix-permissions requirement.

tests: use hghave eol-in-paths to test for support for ':' in filenames and " " It is an approximation but seems to work ok ... and testing that 'foo:bar' creates a real file and not just a resource fork is not completely trivial.

tests: use a different evil name in test-hgweb-raw.t The test used a filename with ':' which prevented the test from running on Windows and FAT. It now uses a filename with space and '%' and will thus still exercise proper url escaping.

tests: add some missing #if's / hghave requirements Seen on Linux with FAT.

tests: introduce hghave hardlinks Some tests can't be run on FAT filesystems because it doesn't support hardlinks.

tests/hghave: test that the inotify unix socket actually can be created Inotify do not work on FAT filesystems.

tests/hghave: test that a fifo actually can be created on the filesystem Some tests were failing on FAT filesystems.

tests/hghave: consistently use dir='.', prefix=tempprefix for tempfiles Some feature tests were done in the wrong directory and could thus give the wrong result.

test-merge-tools: adapt for Windows

tests/hghave: extract hghave.py hghave is degraded to a bare script, moving the functions to hghave.py so they can be shared later on.

test-alias: adapt for Windows The test used 'echo' to test '!' style aliases. On Windows 'echo' is handled by cmd and thus behaves very differently from the 'normal' echo command. The simple workaround used here for using the same alias on all platforms is to use 'printf' instead. Msys 'printf' will also handle sh quoting and escaping in cmd. Environment variable expansion with sh syntax is handled by launching sh.

test-hook: adapt for Windows

tests/printenv.py: replace \ with / in output saves us quite a bunch of (glob)'s

test-clone-pull-corruption: adapt for Windows

test-merge-tools: inline pointless domerge function This also uncovers the exit codes of the 'hg merge' calls

test-bisect: adapt for Windows

test-journal-exists: use #if

merge with stable

acl: user docs for the "!" prefix in user or group names

acl: use of "!" prefix in user or group names The "!" prefix allows you to prevent anyone except a given user or group to push changesets in a given branch or path. This patch enables a use case suggested by a user (Julien Bonnet): There's a branch that only a given user (or group) should be able to push to, and you don't want to restrict access to any other branch that may be created. With the "!" prefix, you simply deny access to "!givenuser" (or "!@givengroup").

parsers: add a C function to pack the dirstate This is about 9 times faster than the Python dirstate packing code. The relatively small speedup is due to the poor locality and memory access patterns caused by traversing dicts and other boxed Python values.

test-convert: disable autocrlf for git

debugwalk: observe ui.slash config option

test-commit: remove extra cd .. This specific cd .. leaves the base directory of the test ($TESTTMP). Removing it avoids that test artifacts (e.g. files) are created outside of the base directory.

merge with stable

notify: various fixes to docstring Typo corrections, grammar adjustments, clarity improvements.

help: sort results from keyword search

tests: reintroduce globs in test-convert-authormap.t They are needed on windows and were accidentally removed in eaf6a6d7f015.

merge with stable

merge with crew

tests: removed unneeded section in acl configuration: 'extensions' Declaring 'acl' in the [extensions] section doesn't change anything.

config: use util.compilere to compile regexps

matcher: use re2 bindings if available There are two sets of Python re2 bindings available on the internet; this code works with both. Using re2 can greatly improve "hg status" performance when a .hgignore file becomes even modestly complex. Example: "hg status" on a clean tree with 134K files, where "hg debugignore" reports a regexp 4256 bytes in size. no .hgignore: 1.76 sec Python re: 2.79 re2: 1.82 The overhead of regexp matching drops from 1.03 seconds with stock re to 0.06 with re2. (For comparison, a git repo with the same contents and .gitignore file runs "git status -s" in 1.71 seconds, i.e. only slightly faster than hg with re2.)

help: fix extension commands help in keyword search This patch fixes the synopsis shown for extension commands in keyword search results. A previous patch erroneously caused the extension synopsis to be shown instead. Test cases for keyword search are missing, so I added a one.

localrepo: lowercase "unexpected response" message

ui: lowercase "no username" warning

ui: lowercase "not trusting file" warning message

ui: lowercase ConfigError warning message

cmdutil: lowercase finddate status message This message is output by 'hg revert -d DATE' and 'hg update -d DATE'.

bisect: lowercase status message

serve: lowercase "no repo here" message

debuginstall: lowercase status messages

dispatch: lowercase --time message

win32text: lowercase warning message

patchbomb: lowercase status messages

purge: lowercase messages

mq: lowercase warning messages

largefiles: lowercase messages

gpg: lowercase messages

fetch: lowercase abort message

convert: lowercase status and abort messages

churn: lowercase message

test-parse-date: move remaining date parsing tests from test-log

test-log: removed tests already covered by test-parse-date

test-parse-date: remove cruft from 8c6f823efcc9 The test contained what looks like a copy-paste mistake.

merge with stable

archive: make progress only show files that are actually archived Before this, files that are excluded (or not included) were shown when using progress bar or --debug. Reported by Andrew Shadura.

test-hardlinks: enable for Windows

test-command-template: enable for Windows

test-rollback: enable for Windows but without the part that needs serve

test-init: enable for Windows On Windows, we can't create a directory with the name " ", as that is not a valid name for a directory.

test-run-tests: add tests for "#if no-false" and no-true

tests: add missing trailing 'cd ..' Many tests didn't change back from subdirectories at the end of the tests ... and they don't have to. The missing 'cd ..' could always be added when another test case is added to the test file. This change do that tests (99.5%) consistently end up in $TESTDIR where they started, thus making it simpler to extend them or move them around.

tests: cleanup of tests that got lost in their own nested directories Some tests ended up in a directory several directories deeper than $TESTTMP, usually because some 'cd ..' had been forgotten between different test cases. Add 'cd ..' where they are missing so the tests get back where they started.

tests: cleanup of tests that ended up outside $TESTTMP No harm done, but it might confuse later on.

tests: remove 'hghave symlink' from test-import-git.t The test had a long chain of commits depending on execbits in some of first commits. The hashes are checked throughout the file, so there was no elegant way to make it pass both with and without execbits. We now rollback the execbits-or-not commits and make a stable change instead. The hash chain is thus updated once but is now a bit more stable. The test coverage should be unaltered.

tests: remove 'hghave symlink' from test-convert-svn-sink.t The symlink was present all over the test and could not easily be guarded with #if. Symlink testing is now moved to a separate #if section. All the relevant test coverage should unaltered.

tests: convert some 'hghave symlink' to #if This will enable some tests for windows.

tests: drop a couple of unnecessary 'hghave symlink'

tests: use the right directory for running hghave from run-tests.py

tests: make .t tests stop immediately if a cd fails This prevents tests from escaping from TESTTMP as a consequence of a failing directory creation.

test-remove-new: integrate into test-remove.t

test-bundle: enable for Windows

test-rename.t: enable for Windows

tests: consistently use a HGEDITOR pattern that works with msys on windows Removing the need for an executable sh script that can be run from a native windows application enables some tests for running on windows.

tests: glob the remaining 'saved backup bundle to' to prepare for windows

tests: convert some 'hghave execbit' to #if This enables some new tests for running on windows.

tests: change odd uses of 'if hghave' to #if

tests/run-tests: avoid C:/ in arguments MSYS replaces C:/... in arguments with C;... as it interprets the C:/ as a colon separated POSIX path list. The colon is replaced with ; (path separator on Windows) according to http://www.mingw.org/wiki/Posix_path_conversion So we must not replace \ with / for neither $TESTTMP nor $TESTDIR, but we have to keep replacing \ with / for the Popen4 call of function hghave. If we don't do the latter, test-run-tests.t will fail with $ python run-tests.py --local test-run-tests.t --- C:\Users\adi\hgrepos\hg-main\tests\test-run-tests.t +++ C:\Users\adi\hgrepos\hg-main\tests\test-run-tests.t.err @@ -70,6 +70,7 @@ tested #else $ echo skipped + skipped #endif #if false An additional tweak in test-ssh.t is needed that globs away an encoded path, as it can't be translated back to $TESTTMP, because the backslashes in the output have been already encoded as %5C. This patch makes test-ssh.t pass in MSYS on Windows.

test-commit: remove extra cd .. This specific cd .. leaves the base directory of the test ($TESTTMP). Removing it avoids that test artifacts (e.g. files) are created outside of the base directory.

help: test keyword search in translated text. Tests that help keyword search is able to search the translated help text when using a non-English language.

test-mq-subrepo-svn: enable for Windows

test-convert-git: remove the read-only attribute of repository files Git object files are stored read-only in the filesystem. Trying to remove a read-only file on windows will fail with access denied, so we have to make them writeable before they can be removed.

test-convert: disable autocrlf for git Git might have autocrlf=true as a global or default setting, especially on windows. That is not expected in the tests and can cause + warning: LF will be replaced by CRLF in d/b. + The file will have its original line endings in your working directory. Explicitly setting it false will make the test pass in some setups - but still not out of the box.

tests/hghave: implement #if true / #if false For unconditionally testing / skipping a section. Useful for testing the test infrastructure in test-run-tests.t and for debugging/developing tests.

tests/run-tests: use $TMP on Windows (issue3490) This is just a short-term workaround for that issue. More work needs to be done on scmutil.canonpath & friends. $TMP on Windows is specified to be defined, and it has correct casing, so we can use that as the default dir for tempfile.mkdtemp on Windows.

test-import: enable for Windows

test-import: remove unneeded chmod call As pointed out by Mads, setting the x bit on dummypatch.py is unneeded, because it is called with 'python ../dummypatch.py'.

test-log: enable for Windows

test-revert: enable for Windows

merge heads

help: fix 'hg help -k' matching an extension without docs getattr is not needed, __doc__ always exists and defaults to None

revlog: zlib.error sent to the user (issue3424) Give the user the zlib error message instead of a backtrace when decompression fails.

statichttprepo: don't send Range header when requesting entire file When requesting files using statichttprepo.httprangereader, a request for the entire file is sent with a Range: bytes=0- header. This causes problems with web servers such as Cherokee that return an HTTP 416 when an empty file is requested in this way, which in turn cause some repository clone attempts to fail. This patch omits the Range header when the entire file is being requested, which fixes the problem.

merge with stable

merge with i18n

merge with i18n

i18n-ru: synchronized with 4cf41c333e74

update: fix help regarding update to ancestor As pointed out on the user mailing list, hg will update just as happily to an ancestor if there are uncommitted changes as to a descendant.

bugzilla: stop bugs always being marked as fixed in xmlrpc (issue3484) Bugs should only be marked fixed if the comment parser gives them the fixed state. xmlrpc+email got this right, xmlrpc screwed it up.

test-add: enable for Windows

test-addremove: remove bits about con.xml Windows reserved filename warnings are already enough tested in test-add.t. This enables the test for Windows.

util: simplify queue management in chunkbuffer This also fixes a small wire protocol performance regression.

hghave: wrap command in 'sh -c "..."' for has_pyflakes() Without this, the has_pyflakes() check always fails in MSYS on Windows.

merge with stable

help: fix keyword search result formatting This patch fixes the broken formatting of keyword search results. Some blank lines were missing from the RST markup, which caused markup to be printed.

help: drop -a from heads syntax summary (issue3483)

revlog: add optional stoprev arg to revlog.ancestors() This will be used as a step in removing reachable() in a future diff. Doing it now because bryano is in the process of rewriting ancestors in C. This depends on bryano's patch to replace *revs with revs in the declaration of revlog.ancestors.

revlog: descendants(*revs) becomes descendants(revs) (API) Once again making the API more rational, as with ancestors.

revlog: ancestors(*revs) becomes ancestors(revs) (API) Accepting a variable number of arguments as the old API did is deeply ugly, particularly as it means the API can't be extended with new arguments. Partly as a result, we have at least three different implementations of the same ancestors algorithm (!?). Most callers were forced to call ancestors(*somelist), adding to both inefficiency and ugliness.

config: make sortdict keys() and iterkeys() methods respect the item order The config.sortdict class is a simple "sorted dictionary" container class, based on python's regular dict container. The main difference compared to regular dicts is that sortdicts remember the order in which items have been added to it. Without this patch the items() method returns the sortdict elements in the right order. However, getting the list of keys by using the keys() or iterkeys() methods, and consequencly, looping through the container elements in a for loop does not respect that order. This patch fixes this problem.

merge with stable

parsers: replace magic number 64 with symbolic constant

revset: introduce and use _revsbetween This is similar in spirit to revlog.nodesbetween, but less ambitious, much simpler, and ~2x faster.

revset: implement dagrange directly This is much faster than the older implementation (~8x).

revset: turn dagrange into a function

revset: drop unreachable code

perf: add a benchmark for revrange

test-tags: enable for Windows It turns out that MSYS does have a chmod.exe, but it has no effect. So, the inserted "#if unix-permissions" is somewhat redundant, as the test would pass without it as well: it would simply write the tag cache, despite what the comment says. But I'm actually in favor of inserting the #if, as it makes it clearer what's going on.

tests: roll test-copy2.t into test-copy.t

help: fix keyword search output for extension commands This patch fixes the help keyword search "hg help -k" to show correct results in the section listing extension commands.

help: format all output using RST This change is the last patch needed to implement help text generation based only on formatting a single text object marked up with RST.

help: format command and option list help using RST This patch changes the function which generates help text about commands and options to use RST formatting. Tables describing options have been formatted using RST table markup for some time already, so their appearance does not change. Command lists, however, change appearance. To format non-verbose command lists, RST field list markup was chosen, because it resembles the old format: <http://docutils.sourceforge.net/docs/user/rst/quickref.html#field-lists> In the old (hand-coded) format of non-verbose command lists, the left column is 12 characters wide. Our minirst implementation formats field lists with a left column 14 characters wide, so this patch changes the appearance of help output correspondingly: <http://markmail.org/message/krl4cxopsnii7s6z?q=mercurial+reinert+from:%22Olav+Reinert%22&page=2> The minirst markup most closely resembling the old verbose command lists is definition lists. But using it would cause a blank line to be inserted between each command definition, making the output excessively long, and no more useful than before. To avoid this, I chose to use field lists also for verbose command help, resulting in output like this example: add add the specified files on the next commit annotate, blame show changeset information by line for each file clone make a copy of an existing repository commit, ci commit the specified files or all outstanding changes diff diff repository (or selected files) export dump the header and diffs for one or more changesets forget forget the specified files on the next commit init create a new repository in the given directory log, history show revision history of entire repository or files merge merge working directory with another revision phase set or show the current phase name pull pull changes from the specified source push push changes to the specified destination qdiff diff of the current patch and subsequent modifications qinit init a new queue repository (DEPRECATED) qnew create a new patch qpop pop the current patch off the stack qpush push the next patch onto the stack qrefresh update the current patch remove, rm remove the specified files on the next commit serve start stand-alone webserver status, st show changed files in the working directory summary, sum summarize working directory state update, up, checkout, co update working directory (or switch revisions) This change is a move towards generating all help text as a list of strings marked up with RST.

help: format extension lists using RST This change is a move towards generating all help text as a list of strings marked up with RST.

test-revset: enable for Windows In MSYS, the test fails like this if the hghave exit at the beginning is removed: --- C:\Users\adi\hgrepos\hg-main\tests\test-revset.t +++ C:\Users\adi\hgrepos\hg-main\tests\test-revset.t.err @@ -58,7 +58,7 @@ $ hg co 3 2 files updated, 0 files merged, 0 files removed, 0 files unresolved $ hg branch /a/b/c/ - marked working directory as branch /a/b/c/ + marked working directory as branch a:/b/c/ (branches are permanent and global, did you want a bookmark?) $ hg ci -Aqm"5 bug" @@ -252,7 +252,7 @@ 2 a-b-c- 3 +a+b+c+ 4 -a-b-c- - 5 /a/b/c/ + 5 a:/b/c/ 6 _a_b_c_ 7 .a.b.c. $ log 'children(ancestor(4,5))' due to the posix path conversion done by MSYS globally, as explained here http://www.mingw.org/wiki/Posix_path_conversion The solution is a bit lame, but it is simple and works: don't use strings that look like '/a/b', in order not to trigger the path magic done by MSYS. So, if we can agree not to insist on testing branch names starting with '/', then this relatively simple patch makes the test pass both on Windows with MSYS and Linux.

graft: restore config option on correct ui

tests: roll test-commit-copy.t into test-commit.t

base85: cast Py_ssize_t values to int (issue3481) If it outputs a nonsense value, no harm done, it was nonsense to start with.

tests: roll test-clone-failure.t into test-clone.t

test-clone-failure: enable for Windows using #if