- Apr 27, 2019
-
-
Pierre-Yves David authored
The strip is an implementation detail. The partial_key seems a clearer name. (note: I am not sure about the "partial" part, but I could not find any better).
-
Pierre-Yves David authored
The new name have been around for 1 years, let us stop hacking the old name back. This will requires people to reclone, however will come in a larger wagon that will requires people to reclone anyway.
-
Pierre-Yves David authored
In some situation, the tear down might try to access an attribute not initialized yet. This is annoying since this usually hide the initial error. To avoid that, we set the default attribute value on the class to make sure the tear down find an appropriate value.
-
- Mar 30, 2019
-
-
Pierre-Yves David authored
If no revision is checkout, the "update" variant of the test will fail since there are no file to update. Before this change, the error was quite obscure: no file name was return, as a result the filepath was just the repo directory and `open(…)` call failed. Now we explicitly detect the error and warn about it. I am not sure what error should be raised, So I am going for a classical ValueError.
-
Pierre-Yves David authored
-
- Apr 24, 2019
-
-
Pierre-Yves David authored
Some of the repository script now rely on asv being available.
-
- Mar 30, 2019
-
-
Pierre-Yves David authored
Otherwise the subscript won't use the virtual env content.
-
Pierre-Yves David authored
-
Pierre-Yves David authored
Same logic as for the previous changeset. Since multiple variant share the same idea, we need so distinct them or horrible things will happens.
-
Pierre-Yves David authored
-
- Apr 24, 2019
-
-
Pierre-Yves David authored
-
Pierre-Yves David authored
-
- Apr 12, 2019
-
-
Raphaël Gomès authored
-
- Apr 06, 2019
-
-
Raphaël Gomès authored
This change was forgotten in 33ba3b5df38e during the draft phase.
-
- Apr 05, 2019
-
-
Raphaël Gomès authored
Most failure cases have to be covered basically end-to-end, so we might as well test the entire script. Moreover, failures would be easy to pinpoint since we are testing a single condition every time.
-
Raphaël Gomès authored
This unit test is here to make sure the base cases are covered and most of the actual logic (that it, not opening files and fetching data) is correct.
-
Raphaël Gomès authored
Pytest is the best python test runner in 2019, so it makes sense for us to use it for testing our own scripts. Since its test discovery relies on python modules (even in Python 3), this change adds the required __init__.py files in the folder structure. A minimal pytest.ini config file was added with usual parameters.
-
Raphaël Gomès authored
Some older benchmarks were not recognized by ASV because of a version discrepancy. ASV generates a version hash from the code that is ran in each benchmark, which means that any change to the benchmark code renders old results unusable by ASV. While the reason for doing this is sensible (any change in the benchmark could invalidate it), this safety measure does not concern us, and we need to bypass it. Since ASV does not yet log a warning (pending another PR) when it ignores results, nor does it allow us to accept those "invalid" results, this change migrates the old results to a form ASV will accept. Most of this changeset is unrelated to the version compatibility issue and is mostly catching up the file structure with the new one: a few new parameters were inserted, the main "repo" parameter was split into itself and "sparse-revlog", etc.
-
Boris Feld authored
-
Boris Feld authored
We have the times in the scheduler logs
-
Boris Feld authored
-
Boris Feld authored
-
Boris Feld authored
Mercurial core commit d7e751ec679e introduced a new format variant named `compression-level` which updates all repo hashes.
-
- Apr 02, 2019
-
-
Raphaël Gomès authored
Some repositories don't have all the variants that are tried. Instead of getting a `KeyError`, we now just skip the impossible combinations.
-
- Mar 23, 2019
-
-
Pierre-Yves David authored
(The non sparse version is 20GB bigger than the sparse one (12 → 32GB). So I think we'll skip it.
-
- Mar 29, 2019
-
-
Pierre-Yves David authored
This mark the end of the infamous hidden directory.
-
- Mar 28, 2019
-
-
Boris Feld authored
-
- Mar 27, 2019
-
-
Boris Feld authored
We need to use the repo name with the hash in order to find the right partial repository.
-
Boris Feld authored
This way we can detect bad file names
-
Boris Feld authored
Also do the check for partial archives with the right name as it can happens too.
-
Boris Feld authored
The test is currently broken The directory in the partial archive doesn't contains the hash, hence extracting two partial for the same repositories with different hashes will overwrite each-other.
-
- Mar 29, 2019
-
-
Pierre-Yves David authored
In the development branch, various file have moved and cache is being warmed. I used Mercurial's core changeset 00c1ee0f746a to generate this output
-
Pierre-Yves David authored
We just need to check that the directory is renamed, we do not actually care about the directory content.
-
Pierre-Yves David authored
-
- Mar 28, 2019
-
-
Raphaël Gomès authored
`hg web` was broken for a certain range of changesets. Make use of the new `not_compatible_with` decorator to reflect that.
-
Raphaël Gomès authored
This decorator will be used to mark versions that do not work most of the time. So instead passing `compatible_with("not(revset)")` and a negative filter function, we can just specify the revset wherein the bench should not run and optionally further filter on test kwargs.
-
- Mar 11, 2019
-
-
Boris Feld authored
-
Boris Feld authored
By default ASV kill the process group only at the very ends, pass a new parameter to `asv run` so it spawn a separate process per test and kill the process group per test in case of timeout.
-
- Mar 06, 2019
-
-
Pierre-Yves David authored
This changeset have a basic description that cover only ½ of its content. In addition, it add a tracked files in repository breaking currently valid setup using a symbolic link to store the refences repositories outside the repository. I'm backing it out to fix this regression. It can we re-submitted for review with a proper explanation about why it is necessary. Changeset 9dff0103c6a6 looks fairly similar, and is also lacking proper justification. However it does not introduce regression (or at least none known yet).
-
Raphaël Gomès authored
From now on, takobasu will be the source of truth for running `asv publish`, so all other machines that run `asv run` will need to rsync their results over to takobasu, which will later deploy the changes on a timer.
-