- Feb 25, 2025
-
-
Pierre-Yves David authored
This will exercise the fncache update.
-
- Jan 28, 2025
-
-
Pierre-Yves David authored
-
- Jan 23, 2025
-
-
Raphaël Gomès authored
We should actually figure out if we're benchmarking a very old version and try to use the old way of setting up... but for now this'll do, and it actually seems to work (until setuptools deprecates something else) for older changesets as well;
-
- Jan 18, 2025
-
-
Raphaël Gomès authored
-
Raphaël Gomès authored
This is has been unusable for a while now and we will most likely be using `egui` within the `pouple-dashboard` thing to build a web-ui.
-
- Jan 15, 2025
-
-
Raphaël Gomès authored
-
Raphaël Gomès authored
-
Raphaël Gomès authored
-
Raphaël Gomès authored
-
- Dec 20, 2024
-
-
Raphaël Gomès authored
There is still a lot of work to do, but this should be a better "first contact" point.
-
- Dec 31, 2024
-
-
Raphaël Gomès authored
-
- Dec 20, 2024
-
-
Raphaël Gomès authored
I am going to experiment with a GUI for poulpe that can run first natively, but later in the web (provided we have a server component to talk to later). This is basically me installing the template and removing all the stuff I'm sure we are not going to need. No actually specific things are happening.
-
- Jan 08, 2025
-
-
Raphaël Gomès authored
This was picked up by running the doctest
-
Raphaël Gomès authored
This is given to us by hyperfine directly, and is important information.
-
- Dec 23, 2024
-
-
Pierre-Yves David authored
I wrongly published the "pre-fix" version, (some baby interruption might have been involved in the lack of pushing).
-
- Dec 19, 2024
-
-
Raphaël Gomès authored
This is still loading from disk, but will at least use the postgres storage that we're interested in.
-
Raphaël Gomès authored
Explanations inline, this will be useful for SWH postgres testing
-
Raphaël Gomès authored
This corresponds to the `--conclude` option of hyperfine and will be used in an upcoming test where a background process is spawned.
-
Raphaël Gomès authored
This will make it easy for benchmarks to have small temporary artifacts. In the future, we might want to configure where this temp dir is created, because it might be important for performance reasons.
-
- Dec 17, 2024
-
-
Raphaël Gomès authored
-
- Dec 13, 2024
-
-
Raphaël Gomès authored
-
- Dec 09, 2024
-
-
Raphaël Gomès authored
This is very basic, but is enough for our first step of creating an entire separate suite.
-
- Dec 06, 2024
-
-
Raphaël Gomès authored
-
- Dec 18, 2024
-
-
Raphaël Gomès authored
There is no way for poulpe itself to determine all the parameters that actually *uniquely* define a bin-env, because it has no knowledge of how to gather its dependencies, for example. We have to assume that bin-envs in general wouldn't survive being moved/renamed, because any hardcoding of the path that would happen in its setup would get broken. Thus, we setup this contract with the suite setup that the unique ID is indeed unique to all relevant parameters for this particular bin-env.
-
- Dec 13, 2024
-
-
Raphaël Gomès authored
The war wages on ;)
-
- Dec 19, 2024
-
-
Raphaël Gomès authored
This is still loading from disk, but will at least use the postgres storage that we're interested in.
-
Raphaël Gomès authored
Explanations inline, this will be useful for SWH postgres testing
-
Raphaël Gomès authored
This corresponds to the `--conclude` option of hyperfine and will be used in an upcoming test where a background process is spawned.
-
Raphaël Gomès authored
This will make it easy for benchmarks to have small temporary artifacts. In the future, we might want to configure where this temp dir is created, because it might be important for performance reasons.
-
- Dec 17, 2024
-
-
Raphaël Gomès authored
-
- Dec 13, 2024
-
-
Raphaël Gomès authored
-
- Dec 09, 2024
-
-
Raphaël Gomès authored
This is very basic, but is enough for our first step of creating an entire separate suite.
-
- Dec 06, 2024
-
-
Raphaël Gomès authored
-
- Dec 18, 2024
-
-
Raphaël Gomès authored
There is no way for poulpe itself to determine all the parameters that actually *uniquely* define a bin-env, because it has no knowledge of how to gather its dependencies, for example. We have to assume that bin-envs in general wouldn't survive being moved/renamed, because any hardcoding of the path that would happen in its setup would get broken. Thus, we setup this contract with the suite setup that the unique ID is indeed unique to all relevant parameters for this particular bin-env.
-
- Dec 19, 2024
-
-
Raphaël Gomès authored
-
- Dec 17, 2024
-
-
Raphaël Gomès authored
-
Raphaël Gomès authored
-
Raphaël Gomès authored
This couldn't reasonably be done in two separate steps because of how intertwined the logic was.
-
Raphaël Gomès authored
-
Raphaël Gomès authored
-