- 17 Dec, 2020 2 commits
-
-
Georges Racinet authored
It is based on the previously introduced `changeset_extracts()` and has options for fields of comparison and order. The hope is that it will make failures appearing on tests easier to understand, especially to (relative) newcomers.
-
Georges Racinet authored
This will be, I hope, more readable than the pervasive "log with `:` separated template / split" pattern we've been using. Especially, I expect failures to be more readily understandable. Using namedtuples for the extracts so that it's admissible in dicts and sets. The drawback is that we have to return the class as well, for litteral comparison. Each call gives a new class but instances of two such namedtuple classes are comparable.
-
- 06 Oct, 2020 1 commit
-
-
Georges Racinet authored
This seems to be necessary for CI tests inside the Docker container, because /opt/gitlab/omnibus/bin is not on the PATH of the `git` user.
-
- 22 May, 2020 1 commit
-
-
Georges Racinet authored
This will be necessary for tests about content that's not encoded with the current default encoding. Because of this, the output might not be utf-8 either. For now, just avoiding an error is good enough.
-
- 18 May, 2020 1 commit
-
-
Georges Racinet authored
That also will be useful for all tests involving the CI. In most of them we won't care about what the CI configuration says, we'll only need it to exist and be valid. --HG-- branch : heptapod-0-12
-
- 13 Mar, 2020 1 commit
-
-
Georges Racinet authored
Pushes that contain phase changes only always return code 1, whereas genuine errors return code 255 (-1i8). Up to now, we've been using `hg_unchecked()` for this case, it's time to simplify.
-
- 19 Feb, 2020 2 commits
-
-
Georges Racinet authored
Some tests require the extensions to be active. This happens for instance with `test_prune_inside_topic`, which won't receive the obsmarkers if evolve is not activated in the first place. Up to now this has been fulfilled by some surrounding configuration, but that's about to change.
-
Georges Racinet authored
this is necessary to have the `hg rebase` command. Up to now, tests using rebase have worked because of configuration attached to the user launching the tests, but that's going to change soon.
-
- 15 Feb, 2020 1 commit
-
-
Georges Racinet authored
a shortcut to simplify getting a full node id, aka revision hash
-
- 04 Jan, 2020 1 commit
-
-
Georges Racinet authored
Don't know why I haven't been bitten by this one before testing on source install, but it doesn't matter, it's best to force a neutral username on our test commits anyway
-
- 17 Nov, 2019 1 commit
-
-
Georges Racinet authored
I'm used to retry with `hg_unchecked` and a pdb trace, but it's better to have the information right away.
-
- 17 Sep, 2019 1 commit
-
-
Georges Racinet authored
apart from the usual str/bytes dance in subprocesses and I/O, (we can assume that hg and docker exec outputs are essentially ASCII, can be handled by the default UTF-8 codec or that we'd write something dedicated otherwise), the notable thing is that we need inner fixture generator to become a context manager, because something in teardowns catches StopIteration too early and raises RuntimeError. This is cleaner anyway.
-
- 10 Aug, 2019 1 commit
-
-
Georges Racinet authored
-
- 19 Jul, 2019 1 commit
-
-
Georges Racinet authored
This is a bit easier than direct invocation, although it's also possible to import the SHORT_TEMPLATE directly.
-
- 03 Jul, 2019 1 commit
-
-
Georges Racinet authored
This will make tests much easier to write and more readable, while still permitting to get errors on the exact failed command.
-