run-tests: skipped test should not produce 'i' mark while retesting When '--retest' option is enabled then skipped test should not produce 'i' mark. This fixes a regression introduced by 92a6b16c9186 and about 200 previous changesets from Gregory Szorc.

test-run-tests.t: test running a failing test

run-tests: filter whitespace at end of error diffs Python 2.6 and 2.7 produce slightly different diff headers.

run-tests: produce '!' mark after running a failing test This patch fixes a regression recently introduced by a refactoring (see 92a6b16c9186 and about 200 previous changesets from Gregory Szorc). '!' mark denotes that the test gets failed while testing.

run-tests: produce error on running a failing test This patch fixes a regression recently introduced by a refactoring (see 92a6b16c9186 and about 200 previous changesets from Gregory Szorc). It produce an error message everytime with a test filename which gets fail while testing except at one condition when '--nodiff' option is enabled.

test-run-tests.t: test running a passing test

run-tests: don't check for the mercurial library used when using --with-hg In the ``--with-hg`` case, we cannot properly infer the location of the mercurial library, so we skip the warning completely.

test: introduce test-run-tests.t This testfile tests the test runner itself. There were multiple regressions in the last few weeks. Having tests would prevent this in the future.

tests: comment the main purpose of ``test-unified-test.t``

parsers.c: fix a couple of memory leaks

tests: rename ``test-run-tests.t`` to ``test-unified-test.t`` This ``test-run-tests.y`` is testing the unified test (.t) syntax interpretation, not really the run-tests.py command tool itself.

wireproto: rephrase the error message for unknown getbundle parameters In case of an unknown parameter passed to ``getbundle``, the server prints a message and ignores the parameter. This message was misleadingly prefixed with "abort: ". The use of "abort: " here is clearly wrong as nothing is actually aborted and the ``getbundle`` call proceeds without the parameter. The message is now prefixed with "warning: " and rephrased to make it clearer that the parameter was simply ignored.

patchbomb: reorder header insertions to clarify code This is done as its own change to make the changes to test-patchbomb.t easier to review.

patchbomb: always use message-id of first patch for series-id This currently has the side effect that the 0 of N message has no series-id. This won't matter for Mercurial's own use, but may be undesirable for other projects depending on their workflow. The way the header is inserted is intentionally a little funny to make the test expectation diff easier to review.

test-patchbomb: normalize glob lines and record their construction This makes it significantly less painful to use --interactive on run-tests, as you can now use the recorded regular expression substitutions to fix up the glob lines and produce zero diffs. This also tightens the expectations of a few of the lines for the MIME boundaries - it just seemed like the thing to do while in here and causing some changes.

patchbomb: includes a unique series ID in email header Now that we have patch index and series size information, having a unique series identifier will helps tool to glue all email back together without any additional logic.

color: add parenthesis to help readability Took me a multiple attempts until my mind eventually stop reading: auto = coloropt = 'auto' And properly reads: auto = colorpopt == 'auto' So we add parenthesis to clarify.

status: document the content of the returned tuple in the docstring The ``status`` function returns a lot of information. We document it.

resolve: fix grammar of no matching files message

merge with stable

bash_completion: add -l|--list support for shelve This was overlooked previously and found via 'hg shelve -p -l <tab>' (to show the diff of the shelved commit).

workingctx: duplicate status list received at _poststatus hook basectx.status may reorder the list after workingctx._poststatus is called, so workingctx must copy it. Otherwise, wctx.deleted() would return "unknown" files, for example.

patch: replace functions in fsbackend to use vfs Several functions in fsbackend are replaced to use vfs functions. vfs operations are applied on filename and path is joined by _join().

vfs: add unlinkpath to vfs This patch adds unlinkpath() function in vfs to replace util.unlinkpath().

shelve: add option combination tests for refactoring in succeeding patch

qfold: allow to specify '--message/'--logfile' and '--edit' at the same time Before this patch, 'hg qfold' disallows to specify '--message/'--logfile' and '--edit' at the same time. 'hg qfold' has disallowed such combination since Mercurial 0.9.2, but this restriction seems not to be reasonable for recent Mercurial, because all other commands creating new changeset allow it. This patch allows 'hg qfold' to specify '--message/'--logfile' and '--edit' at the same time like other commands creating new changeset.

qrefresh: allow to specify '--message/'--logfile' and '--edit' at the same time Before this patch, 'hg qrefresh' disallows to specify '--message/'--logfile' and '--edit' at the same time. 'hg qrefresh' has disallowed such combination since Mercurial 0.9.2, but this restriction seems not to be reasonable for recent Mercurial, because all other commands creating new changeset allow it. This patch allows 'hg qrefresh' to specify '--message/'--logfile' and '--edit' at the same time like other commands creating new changeset.

backout: accept '--edit' like other commands creating new changeset After this patch, users can invoke editor for the commit message by '--edit' option regardless of '--message'/'--logfile'.

gpg: accept '--edit' like other commands creating new changeset After this patch, users can invoke editor for the commit message by '--edit' option regardless of '--message'.

memfilectx: add a size method This method is needed to have memfilectx behave like the other file contexts. We can't just inherit this method because each file context has different behavior: filectx reads from the filelog, and workingfilectx reads from the disk. Therefore, we define memfilectx to return the size of the data in memory.

resolve: keep wlock while resolving This will make resolve use correct locking and thus make it more safe. Resolve is usually a long running command spending a lot of time waiting for user input on hard problems. It is thus a real world scenario to start multiple resolves at once or run other commands (such as up -C and merge) while resolve is running. Proper locking prevents that.

keyword: suppress keyword expansion while 'hg fetch' for internal merge Before this patch, 'hg fetch' may cause unexpected conflict, if 'hg fetch'-ed changes are located near lines in which keywords are embedded, because keywords are substituted with other strings in the working directory. This patch suppresses keyword expansion while 'hg fetch' for internal merge by adding 'fetch' to 'restricted' command list like 'merge'. This patch uses 'hg import' to safely create the new head to be merged at succeeding 'hg fetch', because: - branch of revision #10 is different from one of #11 in 'Test' repository, so just 'hg fetch -r 11' doesn't cause merging between them this means the new head should be created manually. - 'hg import' is easier and safer than 'cat <<EOF' and 'hg commit' to replay same changes including special characters like '$' safeness of 'hg import' with keyword extension is already examined in 'test-keyword.t'.

keyword: suppress keyword expansion while 'hg histedit' for internal merge Before this patch, 'hg histedit' may cause unexpected conflict, if 'hg histedit'-ed changes are located near lines in which keywords are embedded, because keywords are substituted with other strings in the working directory. This patch suppresses keyword expansion while 'hg histedit' for internal merge by adding 'histedit' to 'restricted' command list like 'merge'. Test in this patch just swaps order of revision #13 and #14: this is enough to cause internal merge.

keyword: suppress keyword expansion while 'hg backout' for internal merge Before this patch, 'hg backout' may cause unexpected conflict, if 'hg backout'-ed changes are located near lines in which keywords are embedded, because keywords are substituted with other strings in the working directory. This patch suppresses keyword expansion while 'hg backout' for internal merge by adding 'backout' to 'restricted' command list like 'merge'.

keyword: suppress keyword expansion while 'hg graft' for internal merge Before this patch, 'hg graft' may cause unexpected conflict, if 'hg graft'-ed changes are located near lines in which keywords are embedded, because keywords are substituted with other strings in the working directory. This patch suppresses keyword expansion while 'hg graft' for internal merge by adding 'graft' to 'restricted' command list like 'merge'.

keyword: suppress keyword expansion while 'hg rebase' for internal merge Before this patch, 'hg rebase' may cause unexpected conflict, if 'hg rebase'-ed changes are located near lines in which keywords are embedded, because keywords are substituted with other strings in the working directory. This patch suppresses keyword expansion while 'hg rebase' for internal merge by adding 'rebase' to 'restricted' command list like 'merge'. This patch specifies '--keep' to 'hg rebase', because revision #10 is useful also for tests in succeeding patches.

keyword: suppress keyword expansion while 'hg unshelve' for internal merge Before this patch, 'hg unshelve' may cause unexpected conflict, if 'hg unshelve'-ed changes are located near lines in which keywords are embedded, because keywords are substituted with other strings in the working directory. This patch suppresses keyword expansion while 'hg unshelve' for internal merge by adding 'unshelve' to 'restricted' command list like 'merge'.

merge with stable

Added signature for changeset 269c80ee5b3c

Added tag 3.0.1 for changeset 269c80ee5b3c

merge with i18n

localrepo: prevent leak of transaction object (issue4258) The onclose() closure added in cd443c7589cc held a regular reference to the transaction object. This was causing the transaction to not gc and a leak to occur. The closure now holds a reference to the weakref instance and the leak goes away.

i18n-ja: synchronized with 9e12579369f1

i18n-zh_TW: set Language header This avoids a warning from msgfmt: $ msgfmt -v -o mercurial/locale/zh_TW/LC_MESSAGES/hg.mo i18n/zh_TW.po -c i18n/zh_TW.po:7: warning: header field 'Language' still has the initial default value

i18n-de: translation improvement for gpg extension

bundlerepo: backout dbf292f65b09 According to foozy: This patch should be backed out, because "bundlename" and "bundle" in this case are not relative paths to the root of repositories. The former is specified via "hg incoming --bundle BUNDLENAME" (relative path to cwd, or absolute one), and the latter is generated in "changegroup.writebundle" by "tempfile.mkstemp" for internal temporary usage (absolute path). To be exact, the latter hunk in this patch can be applied, because "os.join" for two absolute paths can generate correct result. But the former hunk can't, because it may unexpected result, if specified path is relative to cwd and cwd != root.

merge: drop the quotes around commit description We already have a ":" after the user name to denote the start of the description. The current usage of quotes around the description is problematic as the truncation to 80 chars is likely to drop the closing quote. This may confuse syntax coloration in some editors.

memfilectx: remove path since it is now inherited

memfilectx: remove __str__ since it is now inherited

memfilectx: remove __nonzero__ since it is now inherited

memfilectx: call super.__init__ instead of duplicating code This patch changes the calling signature of memfilectx's __init__ to fall in line with the other file contexts. Calling code and tests have been updated accordingly.

memfilectx: inherit from committablefilectx This patch marks the beginning of having memfilectx become a full-fledged file contex so that we can handle diffing and merging.

memctx: remove parents since it is now inherited

memctx: remove flags since it is now inherited commitablectx has a much more robust implementation of flags() so we will use that instead of just blindly calling the flags function for the given path.

memctx: remove extra since it is now inherited

memctx: remove branch since it is now inherited

memctx: remove clean since it is now inherited

memctx: remove ignored since it is now inherited

memctx: remove unknown since it is now inherited

memctx: remove deleted since it is now inherited

memctx: remove removed since it is now inherited

memctx: remove added since it is now inherited

memctx: remove modified since it is now inherited

memctx: remove files since it is now inherited This is a slight change in definition from memctx returning only modified() but its parent's definition is more consistent with other contexts' behavior so we can call this change a slight bugfix and step in the right direction.

memctx: remove description since it is now inherited

memctx: remove date since it is now inherited

memctx: remove user since it is now inherited

memctx: remove p2 since it is now inherited

memctx: remove p1 since it is now inherited

memctx: remove __getitem__ since it is now inherited

memctx: remove __nonzero__ since it is now inherited

memctx: remove __int__ since it is now inherited

memctx: remove __str__ since it is now inherited

memctx: call super.__init__ instead of duplicating code

memctx: inherit from committablectx This patch marks the start of having memctx inherit from committablectx, thereby making it a full-fledged context that will eventually grow the ability to perform diffing and also merging.

committablectx: fix _manifest doc string

context: fix wrong indentation from renaming method

push: use bundle2 to push phases when available We now use a bundle2 container to push all phase updates at the same time. This is a significant step forward, even if further refactoring is needed to unify phase push with the changeset push.

pushkey: add an ``encode`` function This function is just a shorthand for ``encoding.fromlocal``. It will help hiding the encoding business from other code exchanging pushkey data over the wire.

bundle2: add ``pushkey`` support After ``listkeys`` we can now include ``pushkey`` request in a bundle2. The part uses a very simple scheme closest as possible to the current wireproto command for ``pushkey``. We may eventually decide for a more sophisticated part format before the protocol becomes final.

pushkey: add a ``decode`` function This function is just a shorthand for ``decoding.fromlocal``. It will help hiding the encoding business from other code exchanging pushkey data over the wire.

pull: when remote supports it, pull phase data alongside changesets We use bundle2 to retrieve the remote phase data at the same time as changesets. This reduces the amount of requestis and should improve consistency as the server can ensure nothing changed between the retrieval of those parts.

getbundle: support of listkeys argument when bundle2 is used A new ``listkeys`` is supported by getbundle. It is a list of namespaces whose content should be included in the bundle. An appropriate entry has been added to the wireproto map of getbundle arguments and a new bundle2 capability is advertised. There are still no codes that request such parts in core mercurial.

getbundle: raise error if extra arguments are provided for bundle10 New arguments are most likely to be supported by bundle20 (and above) only.

bundle: introduce a listkey handler This makes it possible to handle listkeys-related data stored in a bundle2. There are no producers yet.

pull: split remote phases retrieval from actual application We want to be able to retrieve phases information through bundle2. The logic to apply remote phases will not change.

wireproto: use pushkey.decodekey

pushkey: introduce an ``decodekeys`` function This function provides a standardized way to exchange pushkey content over the wire.

wireproto: use pushkey.encodekey

pushkey: introduce an ``encodekeys`` function This function provides a standardized way to exchange pushkey content over the wire.

bundle2: add more phase movement in the test To test phase exchange with bundle2 we need some phase movement to happen. Appropriate logging is added to track those changes.

bundle2: use non-publishing repo in test If we are going to test usage of bundle for phases exchange, we need some phase movements to be exchanged.

bundle2: use a smarter template for test We use a more compact template that includes phase information.

getbundle: declare type of parameters In addition to listing the expected options for ``getbundle``, we also list their types and handle the encoding/decoding automatically. This should make it easier for extensions to transmit additional information to getbundle.

bundle2: introduce a ``caps20to10`` function This function factors the creation of appropriate entries to use in ``bundlecaps`` argument of ``getbundle``. This cleans up code calling ``getbundle`` and helps its usage in more part of the code.

bundle2: introduce a bundle2caps function The process of decoding remote bundle2caps blob into a dictionary is cumbersome. We move it into a small helper function. This will clarify code that reads bundle2 capabilities of peers and helps using it in new places.

exchange: reinsert comment in the right place Unrelated new code was inserted between the original comment and the related code block.

docker: check for docker.io first

buildrpm: include release version in .tar.gz name Official releases are fully indentified by the version number, these builds are not. Specs are however traditionally not versioned.

buildrpm: remove prompt for uncommitted changes - it was a bad idea We want a command that is useful in scripts, not an interactive command.

buildrpm: collect code for building local hg and using it in one place

buildrpm: various minor cleanup

fix_bytes: loosen blacklist matching requirements On my Linux machine, paths seen by 2to3 include the build directory. We switch from an exact to substring match to allow 2to3 to work in more environments.

convert: only consider shamap revisions converted if they still exists Rollback or strip could leave a Mercurial repo with a shamap with revisions no longer in the repository. To ensure reliable conversions we now check that the commit actually exists and consider it non-existing if it doesn't exist.

convert: introduce hascommitfrommap sink method Mercurial has stable revision identifiers and rollback and strip. Revisions referenced in the shamap are thus not necessarily still present but we can easily check for it. Subversion do not have stable identifiers and no rollback or strip(?). We must thus assume that all revisions referenced from a shamap still must be present. This method is similar to hascommitforsplicemap but different ...

convert: rename sink hascommit to hascommitforsplicemap The name 'hascommit' sounds like something generic ... but it might also throw exceptions in specific cases and it is thus (apparently) only useful for splicemap.

tests: better tests for two-way convert

mergetools.hgrc: add minimal configuration for editmerge The ``editmerge`` script is shipped in contrib and opens an editor on every conflicting file. It needs minimal configuration to inject the config marker in the file before opening. Otherwise it behaves the same as ``internal:local`` and bad things happen.

merge with stable

convert: drastically speed up git conversions We would formerly exec git cat-file once for every commit, plus once for every tree and file we wnated to read. This switches to using git cat-file's batch mode, which is much, much, much faster. Using this new code, converting the git git repo to hg ran in 106 minutes on my machine. Using the stock mercurial, it required 1239 minutes. I believe this to be typical of the speedups we will see form this patch.

vim: use try catch in vim plugin to avoid conflicts

bundle2: raise BundleValueError error for stream level unsupported params This ensures both consistency and smooth propagation over the wire.

bundle2: support None parttype in BundleValueError This will be used for errors at the stream level.

bundle2: ignore advisory part with unknown parameters Advisory parts are advisory. If a handler exists but does not support the proper parameters, we can safely ignore it. Test has been updated to include this case.

bundle2: enforce all parameters in a part to be handled Once we picked a handler, we check that all mandatory parameter keys are properly supported. If not we raise an exception. We added a test for this case. The code now fails for any part with unknown mandatory parameters. We will ignore such errors for advisory parts in a later changeset.

bundle2: declare supported parameters for all handlers We now update all existing handlers with the supported parameters information.

bundle2: make it possible to declare params handled by a part handler If we are to enforce the mandatory aspect of parameter, we need a way to discover what a handler supports. The best option we end up with is this a simple declaration of known parameters at registration time. We simply plug the list of parameters on the function object because Python lets us do that and there is no benefit for a more complicated way. One of the handlers is updated for example and testing.

bundle2: support transmission of params error over the wire We picked a null character to split each parameter during the transfer. This is fragile if the same character is used in parameter name. However other codes will already behave in a strange way in that case, so we are not introducing any regression. A better format may be picked for the final version of the protocol.

bundle2: introduce a ``params`` attribute to BundleValueError We'll first use it for unsupported mandatory parameters on parts.

bundle2: introduce a parttype attribute to BundleValueError We will use the Exception for more that just unknown part type.