Augie Fackler <augie@google.com> [Mon, 24 Jul 2017 20:38:09 -0400] rev 33594
osx: new script for generating OS X package versions
If you're shipping prerelease or rc packages using Munki, you'll
eventually discover that Munki's version comparison logic is not as
good as pip's. In theory we should be able to fix Munki, but it seems
entirely reasonable to produce version strings that sort reasonably
under these conditions. Since the requried logic not brief, add a new
script and some tests of that logic.
A followup change will wire this into the Makefile.
Augie Fackler <augie@google.com> [Mon, 24 Jul 2017 20:34:08 -0400] rev 33593
tests: update mac packaging test expectations
Some more files are included these days.
Augie Fackler <augie@google.com> [Wed, 26 Jul 2017 17:56:44 -0400] rev 33592
tests: add support in run-tests.py for an environment variable to stop color
I don't like the color output, so this gives me a way to easily never
see it while still having pygments on my machine.
Augie Fackler <augie@google.com> [Wed, 26 Jul 2017 13:21:34 -0400] rev 33591
setup: fix mistake that prevented Python 3 from being excluded
My intent was to only allow Python 3 if the right environment variable
is set (for when people want to use `pip install .` on hg
locally). This fixes the bug in my previous change. I verified that
`python3.6 run-tests.py` still passes the tests that passed before,
and that all tests pass on 2.7 (including our virtualenv-using
installation test).
Differential Revision: https://phab.mercurial-scm.org/D185
Jun Wu <quark@fb.com> [Mon, 24 Jul 2017 23:52:56 -0700] rev 33590
rebase: move bookmark to destination for commits becoming empty (issue5627)
When rebasing a changeset X and that changeset becomes empty, we should move
the bookmark on X to rebase destination.
This is a regression caused by the scmutil.cleanupnodes refactoring for
rebase.
The `adjustdest` function calculates the destination of bookmark movement.
It was back-ported from https://phab.mercurial-scm.org/D21. It might be
slightly more powerful than the minimal requirement to solve this issue.
For example, it's impossible for a merge changeset to become empty while any
of its ancestors does not become empty, but the code could handle that case.
Since the code is reasonably short and clean, and helps the upcoming D21
series, I'd like to check-in `adjustdest` now.
Thanks Martin von Zweigbergk for spotting corner cases (-k and descendant
with bookmarks) in this area!
Yuya Nishihara <yuya@tcha.org> [Wed, 26 Jul 2017 23:39:42 +0900] rev 33589
setup: drop use of set literal so parsable on Python 2.6
Otherwise friendly error wouldn't be displayed.
Augie Fackler <augie@google.com> [Fri, 21 Jul 2017 10:46:31 -0400] rev 33588
setup: add extra logic to try and recommend a new pip on bad Python
Modern pip can detect supported Python versions (which we now
declare), and pull down a reasonable release. This trick was suggested
in http://bit.ly/pycon2017-build-bridges, and seems like a good
defensive maneuver so that when we want to move to Python 3 it's
less risky for existing users.
This moves the version-check logic after defining our printf function
so we can print more informative messages.
Augie Fackler <augie@google.com> [Fri, 21 Jul 2017 10:39:52 -0400] rev 33587
setup: explicitly declare supported Python versions
I think we should probably backport this to 4.2 as well, and do one
more release there that explicitly declares 2.6 support. That way
anyone stuck on Python 2.6 will end up getting the right hg if they
use a modern pip to install. Users can still use `python setup.py`
incantations to attempt installing Mercurial on unsupported Pythons,
including 3.5 and 3.6.
A followup change will switch to only doing our own
Python-version-check logic if we're not being installed by a
reasonable pip.