- Dec 23, 2019
-
-
Georges Racinet authored
-
- Mar 05, 2020
-
-
Manuel Jacob authored
* run `find . -name "*.py" -print0 -o -path ./tests/hghave | xargs -I {} -0 2to3 --fix=print --write {}` * run `find . -name "*.py" -print0 -o -path ./tests/hghave | xargs -I {} -0 2to3 --doctests_only --fix=print --write {}` * remove extra newline at end of files introduced by last command * add `from __future__ import print_function` to Python files
-
- Nov 22, 2019
-
-
timeless developer authored
-
- Nov 24, 2019
-
-
timeless developer authored
-
- Nov 12, 2019
-
-
muxator authored
Command executed: find . -name "*.py" -print0 | xargs -I {} -0 2to3 --fix=import --write {}
-
muxator authored
Command executed: find . -name "*.py" -print0 | xargs -I {} -0 2to3 --fix=numliterals --write {}
-
muxator authored
Command executed: find . -name "*.py" -print0 | xargs -I {} -0 2to3 --fix=except --write {}
-
muxator authored
Before this patch: --- /opt/mercurial/hg-git/tests/test-check-pyflakes.t +++ /opt/mercurial/hg-git/tests/test-check-pyflakes.t.err @@ -17,3 +17,8 @@ > -X tests/ \ > 2>/dev/null \ > | xargs pyflakes 2>/dev/null + hggit/__init__.py:59: local variable 'e' is assigned to but never used
-
- Mar 05, 2020
-
-
Manuel Jacob authored
After copying the files, a6ddd3f835af was reapplied.
-
- Jun 03, 2019
-
-
Manuel Jacob authored
This functionality was there before, but didn’t work with dulwich >= 0.19.0, which we now depend on, and was therefore removed.
-
Manuel Jacob authored
Dulwich 0.19.0 switched to urllib3 and therefore doesn’t accept the `opener` argument anymore. Therefore the code that builds and tries to pass the opener can be removed.
-
- Mar 01, 2020
-
-
Georges Racinet authored
In dbbae122f5e4 of Mercurial, the `default` kwarg has been removed. Now `manifest.flags()` always returns an empty string as a default, and that's exactly what we were asking anyway.
-
- Jan 11, 2020
-
-
Georges Racinet authored
This is the same as what's been done earlier for heptapod/hg-git: `/usr/bin/ssh` is no longer installed with `git-core` because the base image skips install of `Recommends` dependencies.
-
- Dec 08, 2019
-
-
Georges Racinet authored
This `_maybe_filtered` attribute has been introduced in Mercurial revision 8a37203ab1d5, and is henceforth expected by some callers. Adding it inconditionnally should not harm on previous versions: it would simply not get used.
-
Georges Racinet authored
as of Mercurial c5548b0b6847, status is no more a tuple. Its attributes must be accessed explicitely. These attributes have been available since 3.2 (introduced in cb4449921a1d), so that we don't need to introduce a conditional such as `isinstance(tuple)`. Technically, these are slots, so that performance-wise, such prefetching in locals should not matter, but they make for the smallest changeset.
-
- Dec 22, 2019
-
-
Georges Racinet authored
Currently, the tests fail for this version of Mercurial, hence we'll add the job once they are fixed. Ideally, this changeset should be rebased on top of a revision that fixes hg-git for Mercurial default, hence validating the fix.
-
Georges Racinet authored
All commands that make use of Python and are meant to retrieve or prebuild code should not themselves use potentially incompatible stuff from /ci/repos/mercurial. For instance, updating Mercurial to 'stable' should not risk using a random version of the mercurial Python package. It must instead use the one provided system-wide, that is consistent with the `hg` command.
-
- Dec 08, 2019
-
-
Georges Racinet authored
Without these two flags, we get the following: ``` devel-warn: changectx.__init__ is getting more limited, see context.changectxdeprecwarn() for details (compatibility will be dropped after Mercurial-4.6, update your code.) at: /tmp/hgtests.ShCiwV/install/lib/python/mercurial/localrepo.py:849 (__contains__) ``` after inspection of the code paths, it turns out that these are self-inflicted Mercurial warnings. This can be seen with test-clone.t, the whole traceback does not involve hg-git at all. So the only practical solution we have is to silence the warnings. I think it's acceptable as long as it's for this single version and in particulary not for any version under active development or maintenance.
-
- Dec 07, 2019
-
-
Georges Racinet authored
Noticed that some tests don't pass while working on the CI configuration. Ideally, this changeset should be rebased on top of a revision that fixes hg-git for Mercurial default, hence validating the fix.
-
- Dec 22, 2019
-
-
Georges Racinet authored
Git behaviour and output has changed, in a way that seems hard to reconcile with glob or regex patterns. we'll probably have to write an entire variant of this test. In the meanwhile, this skipping should restore the main CI runs.
-
Georges Racinet authored
This is to run some tests only if the Git version is smaller than the given one.
-
Georges Racinet authored
Now that Octobus images for CI of Mercurial extensions against a fixed Mercurial version have a prebuilt Mercurial, we can just use it, saving some computing resources and waiting time. Also, this makes sure that the CI will use the test runner provided with hg-git, which will avoid problems on older versions (seen one with a new skip rule). To avoid inconsistencies (hg executable importing the system wide mercurial Python package), we need to also set PYTHONPATH, and conversely unset it to run pip in before_script
-
- Dec 08, 2019
-
-
Georges Racinet authored
Currently, hg-git does not pass the tests for - Mercurial 4.6 - the head of Mercurial default branch In order to start benefit of the CI and enforcing that contributions must pass it, we will therefore add jobs for these two versions in separate topics and Merge Requests. Also worth of note is that 5.2 can have new point releases that will be provided in the base Docker image. Hence we don't exclude it from nightly runs.
-
Georges Racinet authored
This Gitlab CI configuration launches the tests of the target hg-git revision against: - the current head of Mercurial `stable` branch - a fixed Mercurial version, assumed not to change anymore We'll add more versions in a follow-up. ## `image` All operations after cloning of the tested `hg-git` revision happen in a fresh container started on this Docker image. The ones we use are available on Docker Hub, with definitions in https://dev.heptapod.net/octobus/ci-dockerfiles ## `before_script` The `before_script` at toplevel will be executed for each job. In later GitLab versions, we'll also have a `default` toplevel key that will allow to specifiy a common value for `script` (and `before_script`, deprecating its use at toplevel). The installations performed by this `before_script` would be better provided by a dedicated series of base Docker images, saving resources and execution time. If that is done, we are able to remove this `before_script` altogether. ## `except` In a first installment, we'll add a nightly scheduler after landing of this configuration. We don't want it to relaunch tests against Mercurial versions that don't change ## `script` This form has the tests launcher build and use the appropriate Mercurial revision (information about it displayed first).
-
- Nov 20, 2019
-
-
Georges Racinet authored
-
Kevin Bullock authored
-
- Nov 18, 2019
-
-
Kevin Bullock authored
-
Kevin Bullock authored
-
- Sep 07, 2019
-
-
Pierre-Yves David authored
The `mercurial/repository.py` file got moved into `mercurial/interfaces/repository.py`. Hg-Git is now aware of it.
-
- Aug 07, 2019
-
-
Kevin Bullock authored
-
Kevin Bullock authored
-
Kevin Bullock authored
-
Kevin Bullock authored
-
Kevin Bullock authored
-
Kevin Bullock authored
-
Kevin Bullock authored
-
- Aug 01, 2019
-
-
Pierre-Yves David authored
Changeset 27d6956d386b changed the string for the "root" directory from `"."` to `""`. We need to update the hg-git version of that code to adjust. The new code works for both pre and post 5.1 version of Mercurial.
-
Pierre-Yves David authored
The message case got fixed in 5.0. So we need to match both version.
-
- May 19, 2019
-
-
Kevin Bullock authored
-