| opendevreview | Radomir Dopieralski proposed openstack/horizon master: Revert "Create and use local upper-constraints to avoid Selenium failing" https://review.opendev.org/c/openstack/horizon/+/978266 | 08:31 |
|---|---|---|
| opendevreview | Radomir Dopieralski proposed openstack/horizon master: Revert "Create and use local upper-constraints to avoid Selenium failing" https://review.opendev.org/c/openstack/horizon/+/978266 | 08:32 |
| opendevreview | Jan Jasek proposed openstack/horizon master: zuul: Remove PhantomJS deps and update parent job https://review.opendev.org/c/openstack/horizon/+/978264 | 10:29 |
| opendevreview | Jan Jasek proposed openstack/horizon master: Remove old Horizon integration tests https://review.opendev.org/c/openstack/horizon/+/978552 | 15:32 |
| stephenfin | rdopiera: tmazur: Before I invest ages in manually fixing a whole load of pep8 failures after lots of sed'ing: how open are you folks to adopting ruff? 🤔 | 16:11 |
| stephenfin | (context is https://review.opendev.org/c/openstack/horizon/+/978278 but also my earlier selenium 4.x change) | 16:12 |
| stephenfin | *context is mainly | 16:12 |
| rdopiera | stephenfin: I used ruff on horizon code before, it should just work | 16:48 |
| stephenfin | I mean `ruff format` (though also `ruff check`). We use it in all SDK projects and most oslo projects for some time now, and I believe keystone and a few others are using it also. | 16:49 |
| rdopiera | stephenfin: this is my ruff.toml if you want it: https://gist.github.com/deshipu/394526181650b1df663ceaa1dc206dd7 | 16:50 |
| rdopiera | I don't want to go through the global requirements process to include it in test requirements | 16:51 |
| stephenfin | ruff is a linter so it should be blacklisted from g-r | 16:51 |
| rdopiera | ruff is a lot of things | 16:52 |
| stephenfin | though we run it via pre-commit, which bypasses the g-r check, e.g. https://github.com/openstack/openstacksdk/blob/cf6f9a00af3bd84de1e77b07a367511df1f2001d/tox.ini#L67 | 16:52 |
| rdopiera | ideally we would have it in our pep8 job instead of hacking/flake | 16:52 |
| rdopiera | personally I'm not a fan of autoformatters, I think they remove important signal from the code | 16:53 |
| stephenfin | aye, that's what I would suggest. I have a local series to add it (`ruff check` first, then `ruff format` later) but I didn't want to propose it unless you were open to it | 16:53 |
| stephenfin | rdopiera: do tell? | 16:53 |
| rdopiera | if you see badly formatted code, it gives you an idea about where the bugs might be | 16:54 |
| stephenfin | hmm, fair, but in a world of LLMs is that still valid? | 16:54 |
| rdopiera | even more so | 16:54 |
| stephenfin | thinking of https://nadh.in/blog/code-is-cheap/ | 16:54 |
| rdopiera | llms generated code is a liability | 16:55 |
| rdopiera | I mean, all code is liability, but code that no human looked at is doubly so | 16:55 |
| stephenfin | *"vibe coded" LLM code is a liability | 16:55 |
| rdopiera | automation is a blunt instrument | 16:55 |
| * stephenfin personally approaches LLM-written code as code written by someone else and reviews it like another person's code | 16:56 | |
| rdopiera | humans make certain kinds of errors commonly, llms make errors no human would even think about | 16:57 |
| stephenfin | so what I love about auto-formatters is that they allow you to large, sed/awk/<whatever>-driven updates and not have to go fix a few hundred pep8 violations after | 16:58 |
| rdopiera | anyways, I would love to see a cross-project session at the PTG about new tools that we could use | 16:58 |
| stephenfin | *allow you to make | 16:58 |
| rdopiera | I would rather not go against the grain of the official openstack recommendations | 16:58 |
| rdopiera | we are already in an icky situation with pytest | 16:58 |
| stephenfin | So I've seen. Makes sense | 16:58 |
| opendevreview | Merged openstack/horizon master: zuul: Remove PhantomJS deps and update parent job https://review.opendev.org/c/openstack/horizon/+/978264 | 17:00 |
| opendevreview | Stephen Finucane proposed openstack/horizon master: Migrate setup configuration to pyproject.toml https://review.opendev.org/c/openstack/horizon/+/978792 | 17:01 |
| opendevreview | Stephen Finucane proposed openstack/horizon master: Migrate coverage configuration to pyproject.toml https://review.opendev.org/c/openstack/horizon/+/978793 | 17:01 |
| opendevreview | Stephen Finucane proposed openstack/horizon master: Move horizon.hacking to a private module https://review.opendev.org/c/openstack/horizon/+/978794 | 17:01 |
| opendevreview | Stephen Finucane proposed openstack/horizon master: Migrate to pre-commit for linter checks https://review.opendev.org/c/openstack/horizon/+/978795 | 17:01 |
| rdopiera | I'm off for today | 17:02 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!