Yuya Nishihara <yuya@tcha.org> [Sun, 01 Nov 2015 17:42:03 +0900] rev 26965
import-checker: reset context to verify convention in function scope
I got the following error by rewriting hgweb/webcommands.py to use
absolute_import. It is false-positive because the import line appears in
"help" function:
hgweb/webcommands.py:1297: higher-level import should come first: mercurial
This patch makes the import checker aware of the function scope and apply
rules recursively.
Yuya Nishihara <yuya@tcha.org> [Sun, 01 Nov 2015 00:37:22 +0900] rev 26964
import-checker: allow import of child modules from package root
I got the following error by rewriting hgweb/__init__.py to use
absolute_import, which is obviously wrong:
Import cycle: mercurial.hgweb.__init__ -> mercurial.hgweb.__init__
"from foo import bar" should not make a cycle if "foo" is a package and
if "bar" is a module or a package. On the other hand, it should be detected
as a cycle if "bar" is a non-module name. Both cases are doc-tested already,
so this patch does not add new doctest.
Matt Mackall <mpm@selenic.com> [Mon, 02 Nov 2015 17:17:33 -0600] rev 26963
encoding: extend test cases for utf8b
This adds a round-trip helper and a few tests of streams that could
cause synchronization problems in the encoder.
Siddharth Agarwal <sid0@fb.com> [Fri, 13 Nov 2015 22:43:09 -0800] rev 26962
merge: make 'cd' and 'dc' actions store the same arguments as 'm'
We're going to treat these conflicts similarly to merge conflicts, and this
change to the way we store things in memory makes future code a lot simpler.
Siddharth Agarwal <sid0@fb.com> [Fri, 13 Nov 2015 14:24:22 -0800] rev 26961
merge: stop emptying out 'cd' and 'dc' actions
(1) These aren't currently read from anywhere, so emptying this out is
pointless.
(2) These *will* be read from later in upcoming patches, and not emptying
them out will be important then.
liscju <piotr.listkiewicz@gmail.com> [Sun, 15 Nov 2015 22:18:48 +0100] rev 26960
rebase: add returning value from pullrebase function
So far pullrebase function has always returned None value, no matter
what orig function returned. This behaviour made impossible for
pull to change returned value from mercurial process (it has always
ended with 0 value by default). This patch makes pullrebase returning
with returned value from orig.
Siddharth Agarwal <sid0@fb.com> [Fri, 13 Nov 2015 15:56:02 -0800] rev 26959
resolve: restore .orig only after merge is fully complete (issue4952)
Previously, we'd restore the .orig file after the premerge is complete but
before the merge was complete. This would lead to the .orig file potentially
containing merge conflict markers in it, as a leftover from the last merge
attempt.
Anton Shestakov <av6@dwimlabs.net> [Fri, 06 Nov 2015 22:32:00 +0800] rev 26958
monoblue: add a space to separate navigation links in file log footer
Before: '(0)-60tip', after: '(0) -60 tip'.
Martin von Zweigbergk <martinvonz@google.com> [Thu, 12 Nov 2015 13:14:03 -0800] rev 26957
merge: move messages about possible conflicts a litte earlier
I actually wanted to reduce the amount of code around the call to
applyupdates(), so I tried moving these warnings a little earlier, and
I think it makes the output make a little more sense (see changes to
test cases).
Yuya Nishihara <yuya@tcha.org> [Sun, 01 Nov 2015 15:46:06 +0900] rev 26956
import-checker: include lineno in warning message
This makes it easy to look for imports in function scope.