- Mar 06, 2019
-
-
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.
-
Boris Feld authored
-
Boris Feld authored
-
Boris Feld authored
-
- Feb 27, 2019
-
-
Boris Feld authored
-
- Mar 05, 2019
-
-
Raphaël Gomès authored
Both benchmarks/ and scmperf_lib/ make use of PyYAML to load yaml files. This library has used an unsafe (i.e. can load yaml bombs, etc.) loader by default since its inception and has been urging people to move to `safe_load` instead. We could have still used `load` by switching to SafeLoader, but the refactor shrunk the code so much that it would not really be an improvement.
-
Raphaël Gomès authored
Strip variants can differ between repos and should not be hardcoded. This change uses the data from the .benchrepo file for each repo to pass to each corresponding test. We are new storing the entire repo prefix instead of only storing the hash, since it is just as unique, removing the need to recompute the prefix from the hash later in the code.
-
- Mar 04, 2019
-
-
Raphaël Gomès authored
This list used to be hardcoded, this change gathers the unique strip variants into a list from every .benchrepo file for use in the ASV test matrix.
-
- Mar 05, 2019
-
-
Raphaël Gomès authored
Every repository may have its own partial sets. This change is reflected in the way this information is passed to `make-partial-repos`: Remove the positional argument for the partial-sets file and fetch the partial sets config from the reference repository .benchrepo file.
-
Raphaël Gomès authored
-
Raphaël Gomès authored
-
- Mar 04, 2019
-
-
Raphaël Gomès authored
This file is useful for configuring the creation of reference repos, as all repos will not need the same number (or critetia) of partials.
-
- Mar 05, 2019
-
-
Boris Feld authored
-
Boris Feld authored
As we know use kwargs in setup too, we need to change how we compute the repo_name.
-
Boris Feld authored
-
Boris Feld authored
Removes unused argument
-
Boris Feld authored
-
Boris Feld authored
-
Boris Feld authored
-
Boris Feld authored
-
- Feb 26, 2019
-
-
Boris Feld authored
This was it will be easier to identify which variants is defined by the repo and which ones are defined by the test itself.
-
- Mar 05, 2019
-
-
Boris Feld authored
Update `update-benchrepo` script to rename partial archives if the benchrepo hash has been updated and the partial archives are present on disk.
-
- Mar 04, 2019
-
-
Boris Feld authored
We will update update-benchrepo to process partials in the next changeset, prepare the needed test files for it.
-
Boris Feld authored
This code was duplicated in several helper scripts, but it in our library.
-
- Feb 25, 2019
-
-
Boris Feld authored
Repositories created using `make-reference` have the same id, prefix and on-disk file name. Update test-update-benchrepo.t to respect the same standard.
-
Boris Feld authored
The update-benchrepo script was initially designed to rebuild the `.benchrepo` file from scratch. The downside of doing that is that it loses all manual changes and important information recorded when the reference was made (like the hg version). The big test change is because the test output was initially made with the previous code that recomputed the whole benchrepo. As the base benchrepo file is quite empty, the fact that we recompute only the format-info generate a big diff in the test file.
-
- Mar 04, 2019
-
-
Boris Feld authored
-
Boris Feld authored
-
- Feb 13, 2019
-
-
Raphael Gomes authored
Some of the performance methods in Mercurial are not defined at all points in the history. There might be times when they were not yet defined or broken, etc.. This decorator adds a way of telling asv that this method should be skipped for the hash being tested.
-
- Mar 05, 2019
-
-
Pierre-Yves David authored
-
- Jan 25, 2019
-
-
Pierre-Yves David authored
Index related performance is one of our hot topic right now, so we better track how it goes. There was a previous entry for perfindex but it lacked parameters.
-
Pierre-Yves David authored
For large repository this computing heads is in the 10ms order of magnitude, this is worth tracking.
-
- Mar 05, 2019
-
-
Pierre-Yves David authored
Change in output from e8c6165e7175 were not applied.
-
- Mar 04, 2019
-
-
Raphaël Gomès authored
We currently are limited in the way that tests can be run against the test data, which has to reside in the same folder. For reasons of flexibility and/or performance, one might need to change the filesystem path to the repositories being tested. This change introduces a new config variable "repodir" which must contain a valid path. While this path is ran through `os.path.abspath`, it is recommended to use an absolute path, since the working directory in this context is not garanteed to be stable.
-
- Feb 13, 2019
-
-
Raphaël Gomès authored
-
- Feb 21, 2019
-
-
Boris Feld authored
-
Boris Feld authored
-
Boris Feld authored
make-reference creates reference tar with a name containing the repo hash, update the script to follow the same naming scheme.
-
Boris Feld authored
Make-reference now creates tar archives with files that includes the short hash and the benchrepo also contains the short hash and the repository prefix. Update benchrepo to follow the same structure and benchrepo content.
-
- Feb 12, 2019
-
-
Raphael Gomes authored
-