Martin von Zweigbergk <martinvonz@google.com> [Fri, 27 Dec 2019 14:08:02 -0800] rev 44039
overlwayworkingctx: remove doubly bad reference to wrapped ctx for copies
`_wrappedctx` lives on overlwayworkingctx, not on the repo object, so
we should access it as `._wrappedctx`, not `._repo._wrappedctx`. More
importantly, the overlayworkingctx is relative to its base, not
including it, so the copies returned should not include copies made in
the base.
Differential Revision: https://phab.mercurial-scm.org/D7801
Martin von Zweigbergk <martinvonz@google.com> [Fri, 27 Dec 2019 12:41:56 -0800] rev 44038
movedirstate: get copies from dirstate before setting parents
Setting dirstate parents can modify the copies recorded in the
dirstate when there are two dirstate parents. I don't think we ever
call movedirstate() when there is more than one parent, but it seems
clearer to get the copies from the dirstate first anyway.
Differential Revision: https://phab.mercurial-scm.org/D7800
Kyle Lippincott <spectral@google.com> [Thu, 12 Dec 2019 14:31:11 -0800] rev 44037
fix: convert clang-format-ignorelist to use wildcards
It's very brittle to specify the exact filenames for these vendored/thirdparty
libraries; when there's an upgrade, it's likely not going to be updated.
Differential Revision: https://phab.mercurial-scm.org/D7619
Connor Sheehan <sheehan@mozilla.com> [Wed, 08 Jan 2020 11:59:32 -0500] rev 44036
packaging: update Ubuntu docker build dependencies to Python 3
Changeset
7574ccd87200f02e updated the Debian docker builds to Python 3. In doing
so, the Ubuntu docker-based builds broke, as the Dockerfile does not
include the new dependencies as specified in `contrib/packaging/debian/control`.
This commit changes the dependencies in the Ubuntu Dockerfile template
to their Python 3 equivalents, fixing `make docker-ubuntu-bionic`.
Differential Revision: https://phab.mercurial-scm.org/D7810
Pierre-Yves David <pierre-yves.david@octobus.net> [Tue, 07 Jan 2020 12:09:36 +0100] rev 44035
mmap: add a size argument to mmapread
With this argument, we can control the size of the mmap created. (previously it
was always the whole file.
Differential Revision: https://phab.mercurial-scm.org/D7808
Pierre-Yves David <pierre-yves.david@octobus.net> [Tue, 07 Jan 2020 12:26:37 +0100] rev 44034
revlogutils: move the NodeMap class in a dedicated nodemap module
We will introduce more nodemap related code, so it make sense to move the
existing code in that module first.
Differential Revision: https://phab.mercurial-scm.org/D7809
Kyle Lippincott <spectral@google.com> [Thu, 02 Jan 2020 17:27:37 -0800] rev 44033
progress: flush stderr after clearing
On python3, ui.stderr is buffered, it seems, so we need to flush it to actually
get the progress bar off the screen. This is important since ui.write() will
call into progbar.clear() if it thinks there's a progress bar on the screen,
with the intent that the next thing it outputs is at the beginning of the line
(instead of at the end of the progress bar line). Without the flush, we buffer
up the clearing of the screen, and we get some really weird/corrupt output.
Differential Revision: https://phab.mercurial-scm.org/D7784
Matt Harbison <matt_harbison@yahoo.com> [Sun, 29 Dec 2019 21:10:04 -0500] rev 44032
rcutil: drop the `defaultrcpath()` method (API)
The resource based code can service py2, py3 and an oxidized executable, so
there's no reason to leave this around. It was flagged as an API change when it
was introduced, so flagging it again.
Differential Revision: https://phab.mercurial-scm.org/D7777
Matt Harbison <matt_harbison@yahoo.com> [Sun, 29 Dec 2019 21:06:34 -0500] rev 44031
ui: add the ability to apply `defaultrc` configs from resources
We will want the ability to cat out these resources, but the same would apply to
templates, so I'm going to wait for the dust to settle on that. Reading the
default config directly from the filesystem is still in place for now.
Differential Revision: https://phab.mercurial-scm.org/D7776
Matt Harbison <matt_harbison@yahoo.com> [Sun, 29 Dec 2019 20:51:44 -0500] rev 44030
ui: refactor `readconfig()` into a form that can consume resources
The old form can't completely go away, because files outside of packages still
need to be read. The name passed in here is a tuple of `package name, resource`
as needed by the resource API.
I like the idea of stating the config file is embedded in the executable by
listing is as `exe!package.resource`. This would be consistent with how
`debuginstall` points to the executable for the python executable, lib, and
installed modules. While in practice the filesystem path is available from the
backing ResourceReader when the resource is opened, it is a relative path on py2
and absolute on py3. Further, while this would show in the `hg config` output
for each option if set as such here, it doesn't show in the `reading from...`
line when `--debug` is used. The file isn't actually open where that prints, so
there's no way I see to get that info there. So I opted for the simple prefix
to distinguish resources from files.
Differential Revision: https://phab.mercurial-scm.org/D7775