The new "diff" field lets you use the matching revset keyword to find revisions
that apply the same change as the selected revisions.
The match must be exact (i.e. same additions, same deletions, same modified
lines and same change context, same file renames and copies).
Two revisions matching their diff must also match their files. Thus, to match
the diff much faster we will always check that the 'files' match first, and only
then check that the 'diff' matches as well.
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.
Factor update code common to all callers of _addpath into _addpath.
By centralizing the update code here, it provides one place to put
updates to new data structures - in a future patch. It also removes
a few lines of duplicate code.
"extinct" and "unstable" predicates use "obsolete" implementation
internally, but own predicate name should be used in error messages of
them instead of "obsolete".
An `obsolete` boolean property is added to changeset context. Function to get
obsolete marker object from a changeset context are added to the obsolete
module.
All necessary data to fire a simple revset query are already known. No call to
ancestors are needed. Such ancestors calculation was already done to compute
outgoing.missing.
spawndetached() was the only user of _STARTF_USESHOWWINDOW and it creates the
process with _CREATE_NO_WINDOW anyway. If the process has no window, then
there is nothing to hide.
Before this change, a console window briefly popped up on "hg serve -d" and
disappeared again, stealing the focus window (which was very annyoing when
running tests).
Specifying _CREATE_NO_WINDOW instead of _DETACHED_PROCESS fixes this (as tested
on Windows 7 x64).
The problem occured when pushing a changeset that at the same time creates a
new named branch head and moves a bookmark. The code invoked methods that only
exist on localrepo instances, so it failed for any other type of remote. The
test suite only tested against local remotes.
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.
When applying a patch renaming/copying 'a' to 'b' on a revision where
'a' does not exist, the patching process would abort immediately,
without processing the remaining hunks and without reporting it. This
patch makes the patching no longer abort and possible hunks applied on
the copied/renamed file be written in reject files.
b67b333b0d8a attempted to force the filecaches in localrepo to reload
everything after a rollback. But simply clearing _filecache isn't enough,
invalidate() needs to be called before/after. localrepo._rollback calls
invalidate() already, so we clear the map right afterwards which ensures
everything will be reread.
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.
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.)
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.
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.
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.
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.
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.
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.
Keyword search in help (introduced in d455a324f54f and ff267c569bea by Augie
Fackler) tries to translate already translated strings, which results in
Unicode errors in gettext when non-ASCII locale is used. Also command
descriptions should be translated before searching there (thanks to FUJIWARA
Katsunori for pointing this out and actual fix), (issue3482).
When developing, we may see non-standard version strings of the form
5d64306f39bb+20120525
which caused tuplever() to raise
ValueError: invalid literal for int() with base 10: '5d64306f39bb'
and shadowing the real traceback.
Caching has no performance effect on the revset aliases which triggered
the recent recursive evaluation bug. I wrote it not to feel bad about
expanding several times the same complicated expression.
'exact' match objects are sometimes created with a non-list 'pattern'
argument:
- using 'set' in queue.refresh():hgext/mq.py
match = scmutil.matchfiles(repo, set(c[0] + c[1] + c[2] + inclsubs))
- using 'dict' in revert():mercurial/cmdutil.py (names = {})
m = scmutil.matchfiles(repo, names)
'exact' match objects return specified 'pattern' to callers of
'match.files()' as it is, so it is a non-list object.
but almost all implementations expect 'match.files()' to return a list
object, so this may causes problems: e.g. exception for "+" with
another list object.
this patch ensures that '_files' of 'exact' match objects is a list
object.
for non 'exact' match objects, parsing specified 'pattern' already
ensures that it it a list one.
The alias expansion code it changed from:
1- Get replacement tree
2- Substitute arguments in the replacement tree
3- Expand the replacement tree again
into:
1- Get the replacement tree
2- Expand the replacement tree
3- Expand the arguments
4- Substitute the expanded arguments in the replacement tree
and fixes cases like:
[revsetalias]
level1($1, $2) = $1 or $2
level2($1, $2) = level1($2, $1)
$ hg log -r "level2(level1(1, 2), 3)"
where the original version incorrectly aborted on infinite expansion
error, because it was confusing the expanded aliases with their
arguments.
The current revset alias expansion code works like:
1- Get the replacement tree
2- Substitute the variables in the replacement tree
3- Expand the replacement tree
It makes it easy to substitute alias arguments because the placeholders
are always replaced before the updated replacement tree is expanded
again. Unfortunately, to fix other alias expansion issues, we need to
reorder the sequence and delay the argument substitution. To solve this,
a new "virtual" construct called _aliasarg() is introduced and injected
when parsing the aliases definitions. Only _aliasarg() will be
substituted in the argument expansion phase instead of all regular
matching string. We also check user inputs do not contain unexpected
_aliasarg() instances to avoid argument injections.
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.
_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.
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.
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:\")
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.
If the string provided to the 'tag' predicate starts with 're:', the rest
of the string will be treated as a regular expression and matched against
all tags in the repository.
There is a slight backwards-compatibility problem for people who actually
have tags that start with 're:'. As a workaround, these tags can be matched
using a 'literal:' prefix.
If no tags match the pattern, an error is raised. This matches the behaviour
of the previous exact-match code.
This is a partial backout of 7efea3b5db4c.
7efea3b5db4c switched win32.py to using ctypes with the intention to get rid
of the dependency on the pywin32 package.
But 7efea3b5db4c replaced the usage of the Python standard module _winreg in
lookup_reg as well, which was uneeded (note that lookup_reg was later renamed
into lookupreg).
Basically, we're switching back to the previous _winreg-based implementation,
which uses _winreg.QueryValueEx(). QueryValueEx returns a unicode code string.
See also: issue3467
There have been quite a few places where we pop elements off the
front of a list. This can turn O(n) algorithms into something more
like O(n**2). Python has provided a deque type that can do this
efficiently since at least 2.4.
As an example of the difference a deque can make, it improves
perfancestors performance on a Linux repo from 0.50 seconds to 0.36.
For divergent renames the following message is printed during merge:
note: possible conflict - file was renamed multiple times to:
newfile
file2
When a file is renamed in one branch and deleted in the other, the file still
exists after a merge. With this change a similar message is printed for mv+rm:
note: possible conflict - file was deleted and renamed to:
newfile
Although index_headrevs is much faster than its Python counterpart,
it's still somewhat expensive when history is large. Since headrevs
is called several times when the tag cache is stale or missing (e.g.
after a strip or rebase), there's a win to be gained from caching
the result, which we do here.
The C implementation is more than 100 times faster than the Python
version (which is still available as a fallback).
In a repo with 330,000 revs and a stale .hg/cache/tags file, this
patch improves the performance of "hg tip" from 2.2 to 1.6 seconds.
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.
This is achieved by acting as if the user had given -r<rev> for each head rev
of outgoing changesets on the command line, as well as appropriate
--base <rev>.
The discovery information is computed as normal, and then adjusted as above.
On an Irix 6.5.24 system, TIOCGWINSZ is not available. This means that
any usage of the "hg" tool that looks up the terminal size (e.g. "hg
help") will fail with an AttributeError.
A simple work-around is just to wrap this block in mercurial/posix.py
with a try/except so that it ends up using the default 80 characters
width.
Previously, we were finding the most recent version of a file in a
changeset and comparing it against its first file parent. This was
wrong on three counts:
- it would show a diff in revisions where there was no change to a file
- it would show a diff when only the exec bit changed
- it would potentially compare against a much older changeset, which
could be very expensive if git-style rename detection was enabled
This compares the file in the current context with that context's
parent, which may result in an empty diff when looking at a file not
touched by the current changeset.
Previously, graph data has been encoded for processing done by
JavaScript code run in the browser, employing simple structures
with implicit member positions. This patch modifies the graph
command to also produce data employing a dictionary-based
structure suitable for use with the templating mechanism, thus
permitting other ways of presenting repository graphs using that
mechanism.
In order to test these changes, the raw theme has been modified
to include templates for graph nodes and edges. In a similar
fashion, themes could employ technologies such as SVG that lend
themselves to templating to produce the graph display. This patch
makes use of a much simpler output representation than SVG in
order to maintain clarity.
The underlying C code doesn't support indexing by longs, there are no
legitimate reasons to use a long, and longs should generally be
converted to ints at a higher level by context's constructor.
Eliminates
mpatch.c(73) : warning C4244: 'return' : conversion from '__int64' to 'int',
possible loss of data
mpatch.c(299) : warning C4244: 'function' : conversion from 'Py_ssize_t' to
'int', possible loss of data
mpatch.c(321) : warning C4244: '=' : conversion from 'Py_ssize_t' to 'int',
possible loss of data
mpatch.c(335) : warning C4244: 'function' : conversion from 'Py_ssize_t' to
'int', possible loss of data
mpatch.c(346) : warning C4244: 'function' : conversion from 'Py_ssize_t' to
'int', possible loss of data
when compiling for Windows x64 target using the Microsoft compiler.