Pierre-Yves David <pierre-yves.david@fb.com> [Fri, 17 Oct 2014 14:59:10 +0200] rev 23179
getbundle: send highest changegroup format supported by both side
When using bundle2, we find the common subset of supported changegroup-packers
and we pick the max of them. This allow to use generaldelta aware changegroup
through bundle2.
Sune Foldager <cryo@cyanite.org> [Fri, 17 Oct 2014 14:41:21 +0200] rev 23178
changegroup: allow use of different cg#packer in getchangegroupraw
This will allow the use of general delta aware changegroup formats.
Sune Foldager <cryo@cyanite.org> [Fri, 17 Oct 2014 14:41:02 +0200] rev 23177
changegroup: introduce "raw" versions of some commands
The commands getchangegroup, getlocalchangegroup and getsubset now each
have a version ending in -raw. The raw versions return the chunk generator
from the changegroup packer directly, without wrapping it in a chunkbuffer
and unpacker. This avoids extra chunkbuffers in the bundle2 code path.
Also, the raw versions can be extended to support alternative packers
in the future, to be used from bundle2.
Matt Mackall <mpm@selenic.com> [Tue, 04 Nov 2014 16:43:33 -0600] rev 23176
merge with stable
Pascal Quantin <pascal.quantin@gmail.com> [Tue, 04 Nov 2014 21:54:27 +0100] rev 23175
win32: remove Mercurial.ini file from Inno Setup installer (issue4435)
Pascal Quantin <pascal.quantin@gmail.com> [Tue, 04 Nov 2014 21:35:49 +0100] rev 23174
win32: fix win32 installers generation
cd0c51c07e5f introduced a typo preventing a proper generation of the
installers. Also remove ConcatenateFiles() function as it is no more
required.
Pierre-Yves David <pierre-yves.david@fb.com> [Tue, 04 Nov 2014 13:38:33 +0000] rev 23173
patchbomb: add label and color to the confirm output
The current output is mostly a wall of text. This makes it hard to
actually check something for people with lazy eyes. We use labels and
colors to make it more joyful (and get the patch summaries to stand
out). The colors have been arbitrarily choosen. They can be changed
later if someone has a more scientific choice.
Yuya Nishihara <yuya@tcha.org> [Tue, 04 Nov 2014 23:41:46 +0900] rev 23172
tests: write hgrc of more than two lines by using shell heredoc
Here document should be readable than repeating echo commands.
Pierre-Yves David <pierre-yves.david@fb.com> [Tue, 04 Nov 2014 10:40:06 +0000] rev 23171
perf: use a formatter for output
We use a `formatter` object in the perf extensions. This allow the use of
formatted output like json. To avoid adding logic to create a formatter and pass
it around to the timer function in every command, we add a `gettimer` function
in charge of returning a `timer` function as simple as before but embedding an
appropriate formatter.
This new `gettimer` function also return the formatter as it needs to be
explicitly closed at the end of the command.
example output:
$ hg --config ui.formatjson=True perfvolatilesets visible obsolete
[
{
"comb": 0.02,
"count": 126,
"sys": 0.0,
"title": "obsolete",
"user": 0.02,
"wall": 0.0199398994446
},
{
"comb": 0.02,
"count": 117,
"sys": 0.0,
"title": "visible",
"user": 0.02,
"wall": 0.0250301361084
}
]
Pierre-Yves David <pierre-yves.david@fb.com> [Wed, 24 Sep 2014 21:33:12 -0700] rev 23170
bundle2: support a "version" argument in `changegroup` part
When included, this mandatory parameter (mandatory == cannot be ignored) lets the
part handler select the right cgunpacker class.
Pierre-Yves David <pierre-yves.david@fb.com> [Wed, 24 Sep 2014 21:28:54 -0700] rev 23169
bundle2caps: advertise the available versions for changegroup packer
This will let the bundle2 client and server detect what packer they should be using.
This detection part is not done. I expect it to be done with the addition of the
second packer (with generaldelta support).
Pierre-Yves David <pierre-yves.david@fb.com> [Wed, 24 Sep 2014 21:24:06 -0700] rev 23168
changegroup: add a "packermap" dictionary to track different packer versions
We only have "01" right now, but we should get general delta in soon.
Bundle2 is expected to make use of this to advertise and select the right packer
to use on both sides.
Matt Harbison <matt_harbison@yahoo.com> [Mon, 03 Nov 2014 12:08:03 -0500] rev 23167
templater: don't overwrite the keyword mapping in runsymbol() (issue4362)
This keyword remapping was introduced in e06e9fd2d99f as part of converting
generator based iterators into list based iterators, mentioning "undesired
behavior in template" when a generator is exhausted, but doesn't say what and
introduces no tests.
The problem with the remapping was that it corrupted the output for keywords
like 'extras', 'file_copies' and 'file_copies_switch' in templates such as:
$ hg log -r 142b5d5ec9cc --template "{file_copies % ' File: {file_copy}\n'}"
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
File: mercurial/changelog.py (mercurial/hg.py)
What was happening was that in the first call to runtemplate() inside runmap(),
'lm' mapped the keyword (e.g. file_copies) to the appropriate showxxx() method.
On each subsequent call to runtemplate() in that loop however, the keyword was
mapped to a list of the first item's pieces, e.g.:
'file_copy': ['mercurial/changelog.py', ' (', 'mercurial/hg.py', ')']
Therefore, the dict for the second and any subsequent items were not processed
through the corresponding showxxx() method, and the first item's data was
reused.
The 'extras' keyword regressed in de7e6c489412, and 'file_copies' regressed in
0b241d7a8c62 for other reasons. The common thread of things fixed by this seems
to be when a list of dicts are passed to the templatekw._hybrid class.
Pierre-Yves David <pierre-yves.david@fb.com> [Thu, 16 Oct 2014 23:15:35 -0700] rev 23166
revset-matching: call 'getset' on a 'fullreposet'
Calling 'baseset(repo.changelog)' builds a list for all revisions in
the repo. And we already have the lazy and efficient 'fullreposet'
class for this purpose.
This gives us the usual benefits of the fullreposet but it is less visible
because the matching process itself is very expensive:
revset) matching(100)
before) wall 6.413281 comb 6.420000 user 5.910000 sys 0.510000 (best of 3)
after) wall 6.173608 comb 6.170000 user 5.750000 sys 0.420000 (best of 3)
However for some complex list, this provide a massive speedup
revset) matching(parents(100))
before) wall 23.890740 comb 23.890000 user 23.450000 sys 0.440000 (best of 3)
after) wall 6.382280 comb 6.390000 user 5.930000 sys 0.460000 (best of 3)
Pierre-Yves David <pierre-yves.david@fb.com> [Thu, 16 Oct 2014 23:15:06 -0700] rev 23165
revset-parentspec: call 'getset' on a 'fullreposet'
Calling 'baseset(repo.changelog)' builds a list for all revisions in
the repo. And we already have the lazy and efficient 'fullreposet'
class for this purpose.
This gives us the usual benefits of the fullreposet:
revset) 100^1
before) wall 0.002694 comb 0.000000 user 0.000000 sys 0.000000 (best of 897)
after) wall 0.000997 comb 0.000000 user 0.000000 sys 0.000000 (best of 2324)
revset) parents(100)^1
before) wall 0.003832 comb 0.000000 user 0.000000 sys 0.000000 (best of 587)
after) wall 0.001034 comb 0.000000 user 0.000000 sys 0.000000 (best of 2309)
revset) (100^1)^1
before) wall 0.005616 comb 0.000000 user 0.000000 sys 0.000000 (best of 405)
after) wall 0.001030 comb 0.000000 user 0.000000 sys 0.000000 (best of 2258)
Pierre-Yves David <pierre-yves.david@fb.com> [Thu, 16 Oct 2014 23:14:17 -0700] rev 23164
revset-children: call 'getset' on a 'fullreposet'
Calling 'baseset(repo.changelog)' builds a list for all revisions in the
repo. And we already have the lazy and efficient 'fullreposet' class
for this purpose.
This gives us the usual benefits of the fullreposet:
revset) children(tip~100)
before) wall 0.007469 comb 0.010000 user 0.010000 sys 0.000000 (best of 338)
after) wall 0.003356 comb 0.000000 user 0.000000 sys 0.000000 (best of 755)
Pierre-Yves David <pierre-yves.david@fb.com> [Thu, 16 Oct 2014 23:11:25 -0700] rev 23163
revset-ancestorspec: call 'getset' on a 'fullreposet'
Calling 'baseset(repo.changelog)' builds a list for all revisions in
the repo. And we already have the lazy and efficient 'fullreposet'
class for this purpose.
This gives us the usual benefits of the fullreposet:
revset) 100~5
before) wall 0.002712 comb 0.000000 user 0.000000 sys 0.000000 (best of 918)
after) wall 0.000996 comb 0.000000 user 0.000000 sys 0.000000 (best of 2493)
revset) parents(100)~5
before) wall 0.003812 comb 0.010000 user 0.010000 sys 0.000000 (best of 667)
after) wall 0.001038 comb 0.000000 user 0.000000 sys 0.000000 (best of 2361)
revset) (100~5)~5
before) wall 0.005614 comb 0.000000 user 0.000000 sys 0.000000 (best of 446)
after) wall 0.001035 comb 0.000000 user 0.000000 sys 0.000000 (best of 2424)
Pierre-Yves David <pierre-yves.david@fb.com> [Thu, 16 Oct 2014 23:10:44 -0700] rev 23162
revset-rangeset: call 'getset' on a 'fullreposet'
Calling 'baseset(repo.changelog)' builds a list for all revisions in
the repo. And we already have the lazy and efficient 'fullreposet'
class for this purpose.
This gives us the usual benefit ofs the fullreposet:
revset) 10:100
before) wall 0.002774 comb 0.000000 user 0.000000 sys 0.000000 (best of 797)
after) wall 0.001977 comb 0.000000 user 0.000000 sys 0.000000 (best of 1244)
revset) parents(10):parents(100)
before) wall 0.005054 comb 0.000000 user 0.000000 sys 0.000000 (best of 481)
after) wall 0.002060 comb 0.000000 user 0.000000 sys 0.000000 (best of 1056)
Martin von Zweigbergk <martinvonz@google.com> [Sun, 19 Oct 2014 22:19:22 -0700] rev 23161
test-revert: make sure all 'tracked' files are really tracked
When a file is missing in the 'parent' version and is tracked but
missing in the working directory, which happens by the 'missing' or
'removed' types, and the 'clean' type in the working directory, the
file does not exist in the working directory (unlike it would had the
'deleted' type been used). Thus, the *_missing_missing_tracked are not
actually tracked and they end up testing the same state as
*_missing_missing_untracked. To make them tracked, add a temporary
file, just like we do for the delete case. For simplicity's sake,
let's make sure the gen-revert-cases.py script always puts a file in
the working directory, whether or not it's going to be deleted.
Martin von Zweigbergk <martinvonz@google.com> [Sat, 18 Oct 2014 18:12:54 -0700] rev 23160
test-revert: sort by output filename again
Future patches will change how the output of 'gen-revert-cases.py
filelist' is generated, so now we want the order to depend on just the
filename again.
Martin von Zweigbergk <martinvonz@google.com> [Mon, 20 Oct 2014 22:54:18 -0700] rev 23159
test-revert: name files by state, not by state transition
This is the main patch in a series. See motivation in earlier patch.
In this patch, we actually change the names of the generated
files. For example, the file that is currently called missing_clean
becomes missing_missing_missing-tracked and it's clearer that it
should be tracked. It turns out that since the state was not
previously clear, it ended up testing an untracked state, which was
the same as for missing_clean. We'll fix this in a later patch.
Let's also change the content from (base,parent,wc) to
(content1,content2,content3) to make them all the same length so they
line up when displayed.
Martin von Zweigbergk <martinvonz@google.com> [Fri, 17 Oct 2014 06:27:43 -0700] rev 23158
test-revert: temporarily sort by input states instead of output filename
The next patch will change the names of the files produced by the
script in test-revert. In order to reduce the size and increase the
clarity of the next patch, make the order produced by the internal
'gen-revert-cases.py filelist' command independent of the filenames.
Martin von Zweigbergk <martinvonz@google.com> [Sat, 18 Oct 2014 22:23:19 -0700] rev 23157
test-revert: put content, not keys, into 'combination'
By putting the file content rather than keys in the 'combination'
list, we restrict the knowledge of 'ctxcontent' and 'wccontent' to the
loop generating the combinations. That will make it easier to replace
the generation code.
Martin von Zweigbergk <martinvonz@google.com> [Fri, 17 Oct 2014 09:02:30 -0700] rev 23156
test-revert: replace 'removed' in working copy with 'untracked-deleted'
The 'wccontent' variable has eight different states, four of them
tracked, and the other four untracked (at least when the file existed
in the parent revision). Among these eight states, 'removed' sticks
out by lacking the 'untracked-' prefix despite resulting in an
untracked state. To make the symmetry clearer, and to prepare for
future patches, rename 'removed' to 'untracked-deleted', which is
exactly what it is.
Note that, unlike 'remove', 'deleted' is configured in
gen-revert-cases.py to have content in the working directory and that
that content is instead expected to be removed in the test script.
However, no changes are needed to the test script, since it already
contains 'hg forget *untracked*' and 'rm *deleted*', which together
have the same effect as 'hg remove'.
See additional motivation in earlier patch.
Martin von Zweigbergk <martinvonz@google.com> [Thu, 16 Oct 2014 23:59:08 -0700] rev 23155
test-revert: removing a missing file has no effect
The tests for removed_deleted and removed_removed test the same state
as removed_clean and removed_untracked-clean, respectively. Drop the
duplicate tests.
See additional motivation in earlier patch.
Martin von Zweigbergk <martinvonz@google.com> [Fri, 17 Oct 2014 00:39:26 -0700] rev 23154
test-revert: reverting an addition is the same as removing
The tests for added_revert and added_untracked-revert test the same
state as added_deleted and added_removed, respectively. Drop the
duplicate tests.
See additional motivation in earlier patch.
Martin von Zweigbergk <martinvonz@google.com> [Thu, 16 Oct 2014 23:36:40 -0700] rev 23153
test-revert: reverting no change means it's clean
This is the first step in a series that aims to put the state, not the
state transitions, in the filenames of the files generated by the
gen-revert-cases.py script. The possible state of a file in a revision
and in the working copy is only whether it exists and what its content
is (the tests don't care check flags). In the dirstate, the only state
is whether it's tracked or not. With the new naming, the file that is
currently called modified_untracked-clean now becomes
content1_content2_content2-untracked, for example.
By putting these states in the filename, it becomes easier to see that
we're not missing or duplicating any state, and to check that the
state is what we think it is. For example, the file that is currently
called missing_clean becomes missing_missing_missing-tracked and it's
clearer that it should be tracked.
Putting the content in the filename will also make the tests of file
content (e.g. "cat ../content-parent.txt") very obvious.
When we put the state in the filename, the filenames clearly need to
be unique. However, it turns out that some states are currently tested
multiple times. The 'revert' transition in the script means to take
the content from the grandparent. If the parent is the same as the
grandparent, there is no change compared to the parent, which is
exactly what 'clean' means. Avoid testing the same state twice.
Matt Mackall <mpm@selenic.com> [Mon, 03 Nov 2014 16:56:32 -0600] rev 23152
merge with stable
Matt Harbison <matt_harbison@yahoo.com> [Sun, 02 Nov 2014 15:27:15 -0500] rev 23151
extdiff: drop the command alias without options example in the help text
In the dropped example, the extension would look for 'vdiff.diffargs' in the
configuration, and not finding it, would run kdiff3 without the configured
options. That's not obvious to a new user who sees a kdiff3 configuration in
the prepackaged mergetools.rc file, and sees that kdiff3 still runs. While it
is conceivable that the user wants a kdiff3 command that runs without the
preconfigured options, it is more likely what they want is this, which uses the
canned options:
[alias]
vdiff = kdiff3
[extdiff]
kdiff3 =
We could mention alias here, but that seems like it belongs elswhere.
Matt Harbison <matt_harbison@yahoo.com> [Fri, 31 Oct 2014 21:34:55 -0400] rev 23150
extdiff: allow a preconfigured merge-tool to be invoked
There are three ways to configure an extdiff tool:
1) cmd.tool = (/path/to/exe optional)
2) tool = (path/to/exe optional)
3) tool = sometool someargs
Previously, if no executable is specified in the first two forms, the named tool
must be in $PATH, or the invocation fails. Since the [merge-tools] section
already has the path to the diff executable, and/or the registry keys to find
the executable on Windows, reuse that configuration for forms 1 and 2 instead of
failing. We already fallback to [diff-tools] and then [merge-tools] for program
arguments if they aren't specified in the [extdiff] section.
Since this additional lookup only occurs if an executable is not on the $PATH
for the named tool, this is backwards compatible. For now, we assume the user
knows what he is doing if a path is provided.
This change allows a configuration file like this (assuming beyondcompare3 is
configured in merge-tools), instead of hardcoding system specific a path:
[extdiff]
beyondcompare3 =
Matt Mackall <mpm@selenic.com> [Mon, 03 Nov 2014 16:30:21 -0600] rev 23149
extdiff: sort files when snapshotting
This fixes output stability and is generally
filesystem-performance-friendly.
Matt Harbison <matt_harbison@yahoo.com> [Sun, 02 Nov 2014 14:58:50 -0500] rev 23148
filemerge: split the logic for finding an external tool to its own function
This will be used by extdiff in an subsequent patch.
Martin von Zweigbergk <martinvonz@google.com> [Sun, 02 Nov 2014 13:18:08 -0800] rev 23147
largefiles: simplify check for lack of path arguments
Instead of checking for a partial merge by checking that the matches
has no files and no patterns, check that it's not an
always-matcher. Except for being shorter, it also catches the rare
case of an exact-matcher with no files.
Martin von Zweigbergk <martinvonz@google.com> [Fri, 31 Oct 2014 14:11:47 -0700] rev 23146
largefiles: shortcircuit status code also for non-matching patterns
We currently shortcircuit the checking for large file standins if only
patterns of type 'path' are given on the command line. That makes e.g.
"hg st 'glob:foo/**'" unnecessarily slow when the only large files are
in a sibling directory.
Relax the check to be that it is not an always-matcher and that no
large files match the patterns given on the command line.
Note that before this change, only the latter of the following two
would show the status of files in .hglf (since the -I makes
match.anypats() true). After this change, they both display the
status. This behavior doesn't seem correct, but it would be a separate
change to explicitly filter out .hglf even in the shortcircuit case.
hg st .hglf/$file
hg st .hglf/$file -I .
Martin von Zweigbergk <martinvonz@google.com> [Mon, 27 Oct 2014 21:10:24 -0700] rev 23145
largefiles: remove confusing 'or None' from predicate
The match function that is overriden returns a boolean value, so
adding 'or None' is both unnecessary and confusing.
Martin von Zweigbergk <martinvonz@google.com> [Thu, 30 Oct 2014 22:32:39 -0700] rev 23144
largefiles: drop unnecessary setting of matcher._always
In two very similar segments of code, an existing matcher is modified
by changing its _files attribute through a map and a filter
operation. Neither operation can cause an empty list to become
non-empty, so a matcher that always matches can not stop always
matching. Drop the setting of the attribute, so we don't unnecessarily
prevent the fast paths to be taken where these matchers end up being
used.
Mads Kiilerich <madski@unity3d.com> [Sun, 19 Oct 2014 03:22:23 +0200] rev 23143
config: move mergetools configuration from contrib to default configuration
The merge tool configuration is an essential part of a good initial user
experience. 'make osx' installers and direct 'make' installation did not have
merge tool configuration. Now they have.
Note: The installer fixes for windows have been done blindly and might require
additional changes.
Mads Kiilerich <madski@unity3d.com> [Thu, 04 Sep 2014 21:36:35 +0200] rev 23142
config: introduce "built-in" default configuration settings in default.d
This helps providing a more consistent user experience on all platforms and
with all packaging.
The exact location of default.d depends on how Mercurial is installed and
whether it is 'frozen'. The exact location should never be relevant to users
and is intentionally not explained in details in the documentation. It will
however always be next to the help and templates files.
Note that setting HGRCPATH also disables these defaults. I don't know if that
should be considered a bug or a feature.
Mads Kiilerich <madski@unity3d.com> [Sat, 18 Oct 2014 21:48:38 +0200] rev 23141
contrib: buildrpm checking of md5 checksums of downloaded Python and Docutils
Mads Kiilerich <madski@unity3d.com> [Sun, 02 Nov 2014 16:39:02 +0100] rev 23140
contrib: update build defaults to latest Python and docutils versions
Use Python 2.7.8 and Docutils 0.12 for Makefile.python and buildrpm (CentOS 5
rpms).
Mads Kiilerich <madski@unity3d.com> [Thu, 17 Apr 2014 22:47:38 +0200] rev 23139
spelling: fixes from proofreading of spell checker issues
Michael Fyles <mf@vorston.net> [Thu, 26 Jul 2012 11:38:13 +0100] rev 23138
extdiff: quote user-supplied options passed to shell
$ hg extdiff -p cmd -o "name <user@example.com>"
resulted in a shell redirection error (due to the less-than sign),
rather than passing the single option to cmd. This was due to options
not being quoted for passing to the shell, via util.system(). Apply
util.shellquote() to each of the user-specified options (-o) to the
comparison program before they are concatenated and passed to
util.system(). The requested external diff command (-p) and the
files/directories being compared are already quoted correctly.
The discussion at the time of changeset be98c5ce4022 correctly noted
that this course of action breaks whitespace-separated options specified
for external diff commands in the configuration. The lower part of the
patch corrects this by lexing options read from the configuration file
into separate options rather than reading them all into the first
option.
Update test to cover these conditions.
Related changesets (reverse-chronological):
- be98c5ce4022 (fix reverted to make configuration file options work)
- 453097750fbf (issue fixed but without fix for configuration file)
Yuya Nishihara <yuya@tcha.org> [Sun, 02 Nov 2014 13:20:07 +0900] rev 23137
serve: correct meta variable of --daemon-pipefds option
It was changed to lock path at e22695b4472f.
Yuya Nishihara <yuya@tcha.org> [Sun, 02 Nov 2014 13:08:46 +0900] rev 23136
tests: have dumbhttp.py use cmdutil.service() to wait for child to listen()
Because the original dumbhttp.py exited without waiting for listen(), several
tests could fail with "abort: error: Connection refused" if subsequent hg
command is fast enough.
Martin von Zweigbergk <martinvonz@google.com> [Sun, 19 Oct 2014 22:09:03 -0700] rev 23135
test-revert: remove obsolete comment about known misbehavior
It seems like the last known misbehvior that the comment was referring
to was dealt with in 2ff28e07d7d6 (revert: properly back up added
files with local modification, 2014-08-31).
Matt Mackall <mpm@selenic.com> [Mon, 03 Nov 2014 11:06:51 -0600] rev 23134
test-revert.t: fix wc check-code false positive
Matt Mackall <mpm@selenic.com> [Mon, 03 Nov 2014 14:01:52 -0600] rev 23133
merge with stable
Matt Mackall <mpm@selenic.com> [Sat, 01 Nov 2014 22:48:49 -0500] rev 23132
Added signature for changeset ced632394371
Matt Mackall <mpm@selenic.com> [Sat, 01 Nov 2014 22:48:12 -0500] rev 23131
Added tag 3.2 for changeset ced632394371
Pierre-Yves David <pierre-yves.david@fb.com> [Sat, 01 Nov 2014 23:52:53 +0000] rev 23130
setdiscovery: limit the size of all sample (issue4411)
Further digging on this issue show that the limit on the sample size used in
discovery never works for heads. Here is a quote from the code itself:
desiredlen = size - len(always)
if desiredlen <= 0:
# This could be bad if there are very many heads, all unknown to the
# server. We're counting on long request support here.
The long request support never landed and evolution make the "very many heads,
all unknown to the server" case quite common.
We implement a simple and stupid hard limit of sample size for all query. This
should prevent HTTP 414 error with the current state of the code.
Pierre-Yves David <pierre-yves.david@fb.com> [Sat, 01 Nov 2014 23:17:50 +0000] rev 23129
hook: protect commit hooks against stripping of temporary commit (issue4422)
History rewriting commands like histedit tend to use temporary
commits. They may schedule hook execution on these temporary commits
for after the lock has been released. But temporary commits are likely
to have been stripped before the lock is released (and the hook run).
Hook executed for missing revisions leads to various crashes.
We disable hooks execution for revision missing in the repo. This
provides a dirty but simple fix to user issues.
Pierre-Yves David <pierre-yves.david@fb.com> [Sat, 01 Nov 2014 22:59:37 +0000] rev 23128
mq: do not call [0] on revset
The __getitem__ method have been removed. The "first" method is to be used
instead. Test have been extended to test this code path.
Pierre-Yves David <pierre-yves.david@fb.com> [Sat, 01 Nov 2014 22:58:30 +0000] rev 23127
addset: fix `first` and `last` on sorted addset (issue4426)
The lazy sorting were not enforced on addset. This was made visible through MQ.
Matt Mackall <mpm@selenic.com> [Sat, 01 Nov 2014 17:30:57 -0500] rev 23126
clone: properly mark branches closed with --uncompressed (issue4428)
On streaming clone, we were priming the local branch cache with the
remote branchmap, without checking which heads were closed.
This fixes an issue introduced in:
changeset: 17740:e6067bec18da
user: Tomasz Kleczek <tomasz.kleczek@fb.com>
date: Wed Oct 03 13:19:53 2012 -0700
summary: branchcache: fetch source branchcache during clone (issue3378)
that was exposed in 2.9 by:
changeset: 20192:38fad5e76ee8
user: Brodie Rao <brodie@sf.io>
date: Mon Sep 16 01:08:29 2013 -0700
summary: branches: simplify with repo.branchmap().iterbranches()
Mads Kiilerich <madski@unity3d.com> [Sat, 01 Nov 2014 20:00:31 +0100] rev 23125
Makefile: update .PHONY
Based on
sed -n 's/^\([a-z0-9-]*\):\(\s.*\)\?$/\1/gp' Makefile | xargs echo
add check, check-code, update-pot, some packaging targets
Mads Kiilerich <madski@unity3d.com> [Sat, 01 Nov 2014 20:00:00 +0100] rev 23124
buildrpm: fix use of invalid $PLATFORM in mercurial.repo
Matt Mackall <mpm@selenic.com> [Sat, 01 Nov 2014 13:13:04 -0500] rev 23123
debuglocks: add missing usage summary
Mads Kiilerich <madski@unity3d.com> [Sat, 01 Nov 2014 19:02:31 +0100] rev 23122
help: don't crash on help for 'sections' with multiple '.'
Mads Kiilerich <madski@unity3d.com> [Sat, 01 Nov 2014 18:28:54 +0100] rev 23121
help: fix config description of ui.reportoldssl
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Sat, 01 Nov 2014 18:03:17 +0900] rev 23120
i18n-ja: synchronized with f3baaeffe841
Wagner Bruna <wbruna@yahoo.com> [Fri, 31 Oct 2014 22:30:39 -0200] rev 23119
i18n-pt_BR: synchronized with 0b7853f969ac
Wagner Bruna <wbruna@yahoo.com> [Fri, 31 Oct 2014 22:22:41 -0200] rev 23118
merge with i18n
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Sat, 01 Nov 2014 01:03:11 +0900] rev 23117
i18n-ja: synchronized with 5459b30aa498
Durham Goode <durham@fb.com> [Fri, 31 Oct 2014 12:56:25 -0700] rev 23116
clone: fix copying bookmarks in uncompressed clones (issue4430)
ef62c66bee1b broke bookmarks getting copied during uncompressed clones. Since
most of the pull logic has been moved into exchange.py, lets just call
exchange.pull to fix up the repo with the latest bits after the streaming clone
has bootstrapped the repo. This keeps us from having to duplicate the bookmark
logic.
Martin von Zweigbergk <martinvonz@google.com> [Fri, 31 Oct 2014 10:41:36 -0700] rev 23115
revset: don't recreate matcher for every revision
The matcher variable 'm' in checkstatus() is reset to None on each
call, so the caching of the matcher no longer happens as it was
intended. This seems to be a regression in ed7b674824a3 (revset: added
lazyset implementation to checkstatus, 2014-01-03).
Fix by moving the cached matcher into the enclosing function so it's
actually cached across calls. This speeds up
hg log -r 'modifies(mercurial/context.py)' >/dev/null
from 7.5s to 4s.
Also see similar fix in f2aeff8a87b6 (revset: avoid recalculating
filesets, 2014-10-22).
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Sat, 01 Nov 2014 02:43:08 +0900] rev 23114
help: use ":hg:`command`" instead of incorrect ":hg:'command'" notation
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Sat, 01 Nov 2014 02:43:08 +0900] rev 23113
i18n: add i18n comment to error messages of filesets predicates
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Sat, 01 Nov 2014 02:43:08 +0900] rev 23112
i18n: add i18n comment to error messages of template functions
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Sat, 01 Nov 2014 02:43:08 +0900] rev 23111
help: refer ":merge3" instead of "internal:merge3"
According to warning message (introduced by 38e0363dcbe0) in
filemerge.py, the former should be used as official name.
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Sat, 01 Nov 2014 02:43:08 +0900] rev 23110
i18n: make hint message of exception translatable
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Sat, 01 Nov 2014 02:43:08 +0900] rev 23109
help: use "hg files" instead of "hg locate" in "hg help filesets"
The latter command is already deprecated.
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Sat, 01 Nov 2014 02:41:18 +0900] rev 23108
files: refer "hg help filesets" instead of "hg help revsets" in help text
"specifying FILE patterns" should refer the former.
Matt Mackall <mpm@selenic.com> [Thu, 30 Oct 2014 17:52:01 -0500] rev 23107
tests: silence output race in test-run-tests.t
Matt Mackall <mpm@selenic.com> [Thu, 30 Oct 2014 16:57:28 -0500] rev 23106
merge with i18n
Wagner Bruna <wbruna@softwareexpress.com.br> [Mon, 27 Oct 2014 20:38:17 -0200] rev 23105
i18n-pt_BR: synchronized with 8d43c6bb38c0
Wagner Bruna <wbruna@softwareexpress.com.br> [Thu, 23 Oct 2014 18:17:00 -0200] rev 23104
i18n-pt_BR: synchronized with c312ef382033
Pierre-Yves David <pierre-yves.david@fb.com> [Fri, 17 Oct 2014 21:25:48 -0700] rev 23103
transaction: only generate file when we actually close the transaction
Before this change, the file were written for every call to `tr.close()`
exposing data to reader far too early.
Pierre-Yves David <pierre-yves.david@fb.com> [Mon, 29 Sep 2014 00:59:25 -0700] rev 23102
transaction: extract file generation into its own function
We extract the code generating files into its own function. We are
about to move this code around to fix a bug. We'll need it in a
function soon to reuse it for "pending" logic. So we move the code
into a function instead of moving it twice.
Matt Harbison <matt_harbison@yahoo.com> [Tue, 28 Oct 2014 23:05:19 -0400] rev 23101
amend: abort early if no username is configured with evolve enabled (issue4211)
Amend will reuse the original username if a new one is not provided
with -U, but obsolete.createmarkers() only considers ui.username() for
the obsolete marker's metadata. Allowing the metadata field to be
spoofed seems to defeat the point of the field in the first place.
This covers 'evolve amend' and 'ci --amend' with evolve enabled.
Without this, the transaction aborts but the parent changeset is set to -1. The
corresponding test will be added to evolve separately.
Durham Goode <durham@fb.com> [Tue, 28 Oct 2014 14:06:06 -0700] rev 23100
revset: fix O(2^n) perf regression in addset
hg log -r 1 ... -r 100 was never returning due to a regression in the
way addset computes __nonzero__. It used 'bool(self._r1 or self._r2)'
which required executing self._r1.__nonzero__ twice (once for the or,
once for the bool). hg log with a lot of -r's happens to build a one
sided addset tree of N length, which ends up being 2^N performance.
This patch fixes it by converting to bool before or'ing.
This problem can be repro'd with something as simple as:
hg log `for x in $(seq 1 50) ; do echo "-r $x "; done`
Adding '1 + 2 + ... + 20' to the revsetbenchmark.txt didn't seem to repro the
problem, so I wasn't able to add a revset benchmark for this issue.
Matt Mackall <mpm@selenic.com> [Mon, 27 Oct 2014 23:47:41 -0500] rev 23099
tests: add missing glob for Windows
Matt Mackall <mpm@selenic.com> [Mon, 27 Oct 2014 18:44:05 -0500] rev 23098
test-convert-svn-sink: properly isolate symlink section
This was fixed earlier by moving all the symlink bits to a section to
the end of the file, but then it was broken (by the same person) by
adding more tests at the end.
Siddharth Agarwal <sid0@fb.com> [Fri, 24 Oct 2014 11:39:39 -0700] rev 23097
util.fspath: use a dict rather than a linear scan for lookups
Previously, we'd scan through the entire directory listing looking for a
normalized match. This is O(N) in the number of files in the directory. If we
decide to call util.fspath on each file in it, the overall complexity works out
to O(N^2). This becomes a problem with directories a few thousand files or
larger.
Switch to using a dictionary instead. There is a slightly higher upfront cost
to pay, but for cases like the above this is amortized O(1). Plus there is a
lower constant factor because generator comprehensions are faster than for
loops, so overall it works out to be a very small loss in performance for 1
file, and a huge gain when there's more.
For a large repo with around 200k files in it on a case-insensitive file
system, for a large directory with over 30,000 files in it, the following
command was tested:
ls | shuf -n $COUNT | xargs hg status
This command leads to util.fspath being called on $COUNT files in the
directory.
COUNT before after
1 0.77s 0.78s
100 1.42s 0.80s
1000 6.3s 0.96s
I also tested with COUNT=10000, but before took too long so I gave up.
Matt Mackall <mpm@selenic.com> [Mon, 27 Oct 2014 16:53:01 -0500] rev 23096
test-clone.t: drop message about listing bookmarks with no hardlinks
Matt Mackall <mpm@selenic.com> [Mon, 27 Oct 2014 16:39:57 -0500] rev 23095
tests: don't try to test unix sockets on vfat
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Tue, 28 Oct 2014 00:19:18 +0900] rev 23094
tests: change obsolete timestamp to avoid "gmtime()" problem on Windows
Before this patch, "test-obsolete.t" fails on Windows environment,
because strings corresponded to "tm_wday" (day of the week) field are
incorrect.
On POSIX environment, "gmtime()" returns correct "tm_wday" value even
for negative "time_t" value. On the other hand, it returns incorrect
one on Windows environment. At least, "gmtime()" of the Windows
runtime library bundled with Python 2.7.3 does.
According to 9a7d0f7e0561 introducing original timestamp value '56
120', it shouldn't cause negative "time_t" value.
test-obsolete: remove subminute timezone in test
Obsmarker format "1" does not supports sub minute timezone. So we
change the test to something slightly more sensible.
It replaced "-d '56 12'" by "-d '56 120'".
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Tue, 28 Oct 2014 00:19:18 +0900] rev 23093
tests: use "%HG_ARGS%" in shell alias on Windows instead of "$HG_ARGS"
Before this patch, a part of "test-alias.t" fails unexpectedly on
Windows environment, because "cmd.exe" can't evaluate "$HG_ARGS"
expression in shell alias correctly.
This patch uses "%HG_ARGS%" in shell alias on Windows instead of
"$HG_ARGS" to expand it correctly.
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Tue, 28 Oct 2014 00:19:18 +0900] rev 23092
tests: introduce "checkeditform-n-cat.sh" script to invoke "cat" in it safely
Before this patch, a part of "test-transplant.t" fails unexpectedly on
Windows environment, because semicolon (";") in HGEDITOR isn't
recognized as the command separator by "cmd.exe".
This patch newly introduces "checkeditform-n-cat.sh" script to invoke
"cat" in it safely anywhere.
Jordi Gutiérrez Hermoso <jordigh@octave.org> [Fri, 24 Oct 2014 13:50:00 -0400] rev 23091
doc: change 'revision or range' to 'revision or revset'
The phrase "revision or range" comes from a pre-revset era. Since the
documentation for ranges now is under the revset docs, and as a
helpful hint nudging users towards revsets, I think it's better to say
"revision or revset"
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Tue, 28 Oct 2014 01:14:12 +0900] rev 23090
largefiles: add examination of exec bit in "hg status --rev REV" case
Before this patch, "hg status --rev REV" doesn't list largefiles up
with "M" mark, even if exec bit of them is changed, because
"lfilesrepo.status" doesn't examine exec bit in such case.
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Tue, 28 Oct 2014 01:14:11 +0900] rev 23089
largefiles: ignore removal status of files not managed in the target context
Before this patch, "hg status --rev REV" listed largefiles removed in
the working directory up with "R" mark, even if they aren't managed in
the REV. Normal files aren't listed up in such case.
When "lfilesrepo.status" is invoked for "hg status --rev REV", it
treats files on conditions below as "removed" (to avoid manifest full
scan in "ctx.status" ?):
- marked as "R" in lfdirstate, or
- files managed in the target revision but unknown in the manifest
of the working context (= not including "R" files)
But the former can include files not managed in the target context.
To ignore removal status of files not managed in the target context,
this patch drops files unknown in the target revision from "removed"
list.
Mads Kiilerich <madski@unity3d.com> [Sun, 26 Oct 2014 12:14:12 +0100] rev 23088
changelog: use headrevsfiltered
2b5940f64750 introduced use of the new filtering headrevs C implementation. It
caught TypeError to detect when to fall back to the implementation that was
compatible with old extensions. That method was however not reliable.
Instead, use the new headrevsfiltered function when passing a filter. It will
reliably fail with AttributeError when an old extension that predates
headrevsfiltered is used.
Mads Kiilerich <madski@unity3d.com> [Sun, 26 Oct 2014 12:14:10 +0100] rev 23087
parsers: introduce headrevsfiltered in C extension
All extensions that have this function do support filtering. The existing
headrevs function may support filtering but we cannot reliably detect whether
it does.
Matt Harbison <matt_harbison@yahoo.com> [Sat, 25 Oct 2014 21:34:49 -0400] rev 23086
httppeer: close the temporary bundle file after two-way streaming it
This fixes several push tests in test-bundle2-exchange.t that were failing on
Windows with messages like the following:
$ hg -R main push http://localhost:$HGPORT2/ -r 32af7686d403 \
--bookmark book_32af
pushing to http://localhost:$HGPORT2/
searching for changes
remote: adding changesets
remote: adding manifests
remote: adding file changes
remote: added 1 changesets with 1 changes to 1 files
remote: 1 new obsolescence markers
updating bookmark book_32af
abort: The process cannot access the file because it is being used by another
process: 'C:\path\to\tmp\bundle.hg'
[255]
Martin von Zweigbergk <martinvonz@google.com> [Fri, 24 Oct 2014 14:24:28 -0700] rev 23085
status: make 'hg status --rev' faster when there are deleted files
In order not to avoid listing files as both added and deleted, for
example, we check for every file in the manifest if it is in the
_list_ of deleted files. This can get quite slow when there are many
deleted files. Change it to a set to make the containment check
faster. On a somewhat contrived example of the Mozilla repo with the
entire testing/ directory deleted (~14k files), this makes
'hg status --rev .^' go from 26s to 2s.
Pierre-Yves David <pierre-yves.david@fb.com> [Mon, 27 Oct 2014 17:52:33 +0100] rev 23084
setdiscovery: limit the size of the initial sample (issue4411)
The set discovery start by sending a "known" command with all local heads. When
the number of local heads is massive (eg: using hidden changesets) such request
becomes too large. This lead to 414 error over http, aborting the whole
process.
We limit the size of the sample used by the first query to fix this.
The test are impacted because they do test massive number of heads. But they do
not test it over real world http setup.
Pierre-Yves David <pierre-yves.david@fb.com> [Mon, 27 Oct 2014 17:40:32 +0100] rev 23083
setdiscovery: extract sample limitation in a `_limitsample` function
We need to reuse this logic for the initial query. We extract it in a function
to unsure sample limiting is applied consistently in all cases.
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 24 Oct 2014 17:24:46 -0500] rev 23082
exchange: don't report failure from identical bookmarks
b901645a8784 regressed the behavior of pushing an unchanged bookmark to
a remote. Before that commit, pushing a unchanged bookmark would result
in "exporting bookmark @" being printed. After that commit, we now see
an incorrect message "bookmark %s does not exist on the local or remote
repository!"
This patch fixes the regression introduced by b901645a8784 by having
the bookmark error reporting code filter identical bookmarks and adds
a test for the behavior.
Gregory Szorc <gregory.szorc@gmail.com> [Fri, 24 Oct 2014 10:40:37 -0700] rev 23081
bookmarks: explicitly track identical bookmarks
bookmarks.compare() previously lumped identical bookmarks in the
"invalid" bucket. This patch adds a "same" bucket.
An 8-tuple for holding this state is pretty gnarly. The return value
should probably be converted into a class to increase readability. But
that is beyond the scope of a patch intended to be a late arrival to
stable.
Martin von Zweigbergk <martinvonz@google.com> [Fri, 24 Oct 2014 15:52:20 -0500] rev 23080
context.status: remove incorrect swapping of added/removed in workingctx
The comment in workingctx.status() says that "calling 'super' subtly
reveresed the contexts", but that is simply not true, so we should not
be swapping added and removed fields.
Martin von Zweigbergk <martinvonz@google.com> [Thu, 23 Oct 2014 13:17:37 -0700] rev 23079
keyword: use parent of working copy as base for status
Instead of calling repo[None].status(), use the more common form that
uses the parent of the working copy as the base:
repo['.'].status(). Note that the former defaults to comparing to
revision '.', while the latter defaults to revision None, so the
contexts being compared are the same.
It might seem like this would result in a reverse diff, but it turns
out that workingctx.status() incorrectly reverses the result. That bug
will be fixed in a later commit.
Yuya Nishihara <yuya@tcha.org> [Thu, 23 Oct 2014 21:53:37 +0900] rev 23078
revset: avoid O(n) lookup of invalid revision in rev()
ba89f7b542c9 was not the final version of that patch. It was really slow
because `l not in repo.changelog` iterates revisions up to `l`. Instead,
rev() should utilize spanset.__contains__().
revset #0: rev(210000)
0) wall 0.000039 comb 0.000000 user 0.000000 sys 0.000000 (best of 67978)
1) wall 0.002721 comb 0.000000 user 0.000000 sys 0.000000 (best of 1055)
2) wall 0.000059 comb 0.000000 user 0.000000 sys 0.000000 (best of 45599)
(0: 3.2-rc, 1: ba89f7b542c9, 2: this patch)
Note that the benchmark result described in ba89f7b542c9 is wrong because
it is the one of the initial version.
Matt Mackall <mpm@selenic.com> [Thu, 23 Oct 2014 13:44:34 -0500] rev 23077
run-tests: output diffs with only forward slashes
This removes some gratuitous variance when testing on Windows with
test-run-tests.t.
Wagner Bruna <wbruna@softwareexpress.com.br> [Thu, 23 Oct 2014 12:35:10 -0200] rev 23076
i18n: add hint to digest mismatch message
Wagner Bruna <wbruna@softwareexpress.com.br> [Thu, 23 Oct 2014 12:28:00 -0200] rev 23075
help/config: fix typo
Wagner Bruna <wbruna@softwareexpress.com.br> [Thu, 23 Oct 2014 12:27:57 -0200] rev 23074
files: fix example list syntax
Mads Kiilerich <madski@unity3d.com> [Thu, 23 Oct 2014 02:42:57 +0200] rev 23073
parsers: use 'k' format for Py_BuildValue instead of 'n' because Python 2.4
'n' was introduced in Mercurial in 2b5940f64750 and broke Python 2.4 support in
mysterious ways that only showed failure in test-glog.t. Py_BuildValue failed
because of the unknown format and a TypeError was thrown ... but it never
showed up on the Python side and it happily continued processing with wrong
data.
Quoting https://docs.python.org/2/c-api/arg.html :
n (integer) [Py_ssize_t]
Convert a Python integer or long integer to a C Py_ssize_t.
New in version 2.5.
k (integer) [unsigned long]
Convert a Python integer or long integer to a C unsigned long without
overflow checking.
This will use unsigned long instead of Py_ssize_t. That is not a good solution,
but good is not an option when we have to support Python 2.4.
Durham Goode <durham@fb.com> [Mon, 20 Oct 2014 18:50:09 -0700] rev 23072
rebase: improve base revset performance
The old revset had pretty terrible performance on large repositories (12+
seconds). This new revset achieves the same result in only 0.7s. As we improve
the underlying revset APIs we can probably get this revset down to 'only(base,
dest)::', but at the moment that version still takes 2s.
Ryan McElroy <rmcelroy@fb.com> [Thu, 16 Oct 2014 06:35:06 -0700] rev 23071
amend: fix amending rename commit with diverged topologies (issue4405)
This addresses the bug described in issue4405: when obsolescence markers are
enabled, amending a commit with a file move can lead to the copy information
being lost.
However, the bug is more general and can be reproduced without obsmarkers as
well, as demonstracted by Pierre-Yves and put into the updated test.
Specifically, graph topology divergences between the filelogs and the changelog
can cause copy information to be lost during amends.
Augie Fackler <raf@durin42.com> [Tue, 21 Oct 2014 17:09:37 -0400] rev 23070
hgweb: disable SSLv3 serving (BC)
Because of recent attacks[0] on SSLv3, let's just drop support entirely.
0: http://googleonlinesecurity.blogspot.com/2014/10/this-poodle-bites-exploiting-ssl-30.html
Augie Fackler <raf@durin42.com> [Tue, 21 Oct 2014 17:01:23 -0400] rev 23069
sslutil: only support TLS (BC)
In light of the POODLE[0] attack on SSLv3, let's just drop the ability to
use anything older than TLSv1 entirely.
This only fixes the client side. Another commit will fix the server
side. There are still a few SSLv[23] constants hiding in httpclient,
but I'll fix those separately upstream and import them when we're not
in a code freeze.
0: http://googleonlinesecurity.blogspot.com/2014/10/this-poodle-bites-exploiting-ssl-30.html
Mads Kiilerich <madski@unity3d.com> [Wed, 22 Oct 2014 16:10:23 +0200] rev 23068
eol: fix crash when handling removed files
ci --amend would in some cases fail after 650b5b6e75ed failed to refactor the
eol extension too.
Pierre-Yves David <pierre-yves.david@fb.com> [Wed, 15 Oct 2014 03:52:20 -0700] rev 23067
bundle2: transmit exception during part generation
If an exception is raised during a bundle2 part payload generation it is now
recorded in the bundle. If such exception occurs, we capture it, transmit an
abort exception through the bundle, cleanly close the current part payload and
raise it again. This allow to generate valid bundle even in case of exception so
that the consumer does not wait forever for a dead producer. This also allow to
raise the exception during unbundling at the exact point it happened during
bundling make debugging easier.
Pierre-Yves David <pierre-yves.david@fb.com> [Tue, 14 Oct 2014 10:47:47 -0700] rev 23066
bundle2: add a interrupt mechanism
It is now possible to emit a single part in the middle of a payload production.
This part will be processed with limitation (only access to a `ui` object). The
goal is to let the server raise exception and output while a part is being
processed. The source motivation is to transmit exception that occurs while
generating a part.
This change is was the motivation to bump the bundle2 format from HG2X to HG2Y.
Somehow, the format bump made it into 3.2 without it. So this change go on
stable. It is low risk as bundle2 is still disabled by default.
Durham Goode <durham@fb.com> [Tue, 21 Oct 2014 12:38:28 -0700] rev 23065
transactions: change backupfiles format to use newlines
Previously the journal.backupfiles file was delimited by \0. Now we delimit it
using \n (same as the journal file). This allows us to change the number of
values in each line more easily, rather than relying on the count of \0's.
Durham Goode <durham@fb.com> [Tue, 21 Oct 2014 11:37:29 -0700] rev 23064
transactions: add version number to journal.backupfiles
The transaction format will be changing a bit over the next releases, so let's
go ahead and add a version number to make backwards compatibility easier. This
whole file format was broken prior to 3.2 (see previous patch), so changing it
now is pretty low risk.
Durham Goode <durham@fb.com> [Mon, 20 Oct 2014 16:53:56 -0700] rev 23063
transactions: fix hg recover with fncache backups
The transaction backupfiles logic was broken for 'hg recover'. The file format
is XXX\0XXX\0YYY\0YYY\0 but the parser did a couple things wrong. 1) It went one
step beyond the final \0 and tried to read past the end of the array. 2)
array[i:i+1] returns a single item, instead of two items as intended.
Added a test to catch it, which turns out to be the first actual 'hg recover'
test.
Yuya Nishihara <yuya@tcha.org> [Sun, 19 Oct 2014 16:48:33 +0900] rev 23062
revset: have rev() drop out-of-range or filtered rev explicitly (issue4396)
The recent optimization of "and" operation relies on the assumption that
the rhs set does not contain invalid revisions. So rev() has to remove
invalid revisions.
This is still faster than using `.filter(lambda r: r == l)`.
revset #0: rev(25)
0) wall 0.026341 comb 0.020000 user 0.020000 sys 0.000000 (best of 113)
1) wall 0.000038 comb 0.000000 user 0.000000 sys 0.000000 (best of 66567)
2) wall 0.000062 comb 0.000000 user 0.000000 sys 0.000000 (best of 43699)
(0: bbf4f3dfd700^, 1: 3.2-rc, 2: this patch)
Matt Mackall <mpm@selenic.com> [Wed, 22 Oct 2014 15:47:27 -0500] rev 23061
revset: avoid recalculating filesets
This fixes a regression in 8dabcc889e33 that moved matcher building
into a callback, thus causing it be rebuilt for each revision matched
against.
FUJIWARA Katsunori <foozy@lares.dti.ne.jp> [Mon, 20 Oct 2014 22:08:08 +0900] rev 23060
tests: open file in binary mode to use POSIX end-of-line style anywhere
Before this patch, "test-context.py" fails on Windows environment,
because "diff" output contains unexpected "\r" character.
Opening the target file in text mode causes this automatic end-of-line
conversion.