Jordi Gutiérrez Hermoso <jordigh@octave.org> [Thu, 14 Feb 2019 10:41:47 -0500] rev 41700
crecord: remove obsolete version check
An internal function shouldn't be checking compatibility with
Mercurial versions.
Jordi Gutiérrez Hermoso <jordigh@octave.org> [Wed, 13 Feb 2019 16:02:44 -0500] rev 41699
histedit: remove "chistedit" mention from interface
"chisted" is internal jargon. The end user should not need to be aware
that it's different from histedit.
Navaneeth Suresh <navaneeths1998@gmail.com> [Wed, 13 Feb 2019 18:17:42 +0530] rev 41698
revset: improve documentation on expectsize()
This is a follow-up patch to D5813. It improves the documentation of
`expectsize(set, size)`.
Differential Revision: https://phab.mercurial-scm.org/D5953
Martin von Zweigbergk <martinvonz@google.com> [Wed, 13 Feb 2019 12:09:36 -0800] rev 41697
scmutil: fix a comment that doesn't match the code
Differential Revision: https://phab.mercurial-scm.org/D5956
Gregory Szorc <gregory.szorc@gmail.com> [Thu, 14 Feb 2019 13:16:36 -0800] rev 41696
remotefilelog: remove strkwargs()
The previous commit activated previously unused code paths on
Python 3 and revealed that we were calling strkwargs() on a dict
that already had str keys. The only caller of _forceprefetch()
is _adjustlinknode() a few lines up and the static commonlogkwargs
dict features str keys.
Differential Revision: https://phab.mercurial-scm.org/D5958
Gregory Szorc <gregory.szorc@gmail.com> [Wed, 13 Feb 2019 16:31:20 -0800] rev 41695
remotefilelog: use raw strings when looking for variable names
Keys in self.__dict__ and vars() are always str. So we need to use
raw strings to ensure lookups work on Python 3.
# skip-blame just r'' prefixes
Differential Revision: https://phab.mercurial-scm.org/D5957
Georges Racinet <georges.racinet@octobus.net> [Thu, 10 Jan 2019 18:25:18 +0100] rev 41694
rust-cpython: binding for headrevs()
This uses the core `dagops::retain_heads` to give a Rust implementation
to `mercurial.dagop.headrevs`.
Testing happens for now from `test-rust-ancestors.py`
(for quick and minimal change), but it'd made more sense to put the binary
index data elsewhere and to create a new test python module