20 May 2021 |
nateberkopec | quick patch release going out tomorrow AM | 16:20:02 |
| Sébastien Auriault joined the room. | 16:49:39 |
21 May 2021 |
nateberkopec | Oof, backporting https://github.com/puma/puma/commit/ffa5d56b845971557dccaf50628705535473650b is gonna be more complicated than I thought. Anyone want to take a stab at it? | 17:05:15 |
4 Jun 2021 |
dentarg | There was no GitHub release for 5.3.2 so I created one, hope that was okay nateberkopec (can be a bit confusing for people when they look at github.com/puma/puma at the blurp on the right says "Latest release" and a version that isn't the latest is shown) | 06:31:01 |
7 Jun 2021 |
| mai-lapyst joined the room. | 12:27:37 |
11 Jun 2021 |
nateberkopec | Gut feelings on the Puma 5 experimental perf features: https://github.com/puma/puma/issues/2258 | 20:49:47 |
nateberkopec | refork: great experiment, more data/expansion of the feature needed | 20:49:55 |
nateberkopec | nakayoshi_fork: remove. not enough benefit, breaking weird c-extensions in prod. | 20:50:07 |
nateberkopec | wait_for_less_busy: maybe make default in 6.0, seems all benefit no downside. | 20:50:24 |
22 Jun 2021 |
Y | nateberkopec: about this issue, https://github.com/puma/puma/pull/2610#discussion_r633572121
Currently localhost_authority does not generate jks file which is used in minissl_context builder. I believe there are implementations that tries to generate keystores. example - https://github.com/rylarson/keystores. What do you think about it? Shall we add feature to generate keystores inside puma just for localhost authority?.
| 19:04:30 |
27 Jun 2021 |
nateberkopec | I like localhost much better as a dependency because @ioquatix is a very prolific and reliable maintainer | 17:36:11 |
nateberkopec | So, gut reaction is: Y keep it using localhost, if someone wants JRuby support they can submit it upstream to localhost and we'll support it then. | 17:36:45 |
nateberkopec | Gonna review this PR again but once this lands I think we're good for a minor version release, Y you should think of a codename! | 17:37:39 |
nateberkopec | I think the tests still need some work but the /lib code is good now | 17:53:12 |
nateberkopec | my concerns re tests are here: https://github.com/puma/puma/pull/2610#discussion_r633569770 | 17:53:32 |
28 Jun 2021 |
Y | I have new tests written. I will submit PR for it. | 14:09:15 |
Y | Redacted or Malformed Event | 17:50:41 |
Y | nateberkopec I have added a new test case to validate whether the certificate is signed by localhost authority or not. Can you take a look at it?
This is my changes https://github.com/ye-lin-aung/puma/blob/master/test/test_puma_localhost_authority.rb#L78
| 17:51:43 |
29 Jun 2021 |
nateberkopec | Anyone have strong feelings on just going "all the way" with a code formatter like standard ? | 13:47:57 |
nateberkopec | https://github.com/testdouble/standard | 13:48:09 |
nateberkopec | I'm sort of sick of seeing small style nitpicks that aren't even really worth commenting about on a PR but that annoy me anyway 😆 | 13:48:29 |
dentarg | This is the way | 13:49:35 |
nateberkopec | I think I'd probably do it as the first commit after a minor release or something to minimize git impact | 13:55:07 |
dentarg | Exclude the test suite or not? | 14:19:55 |
nateberkopec | probably include it | 14:20:14 |
nateberkopec | but I would have to see the diff | 14:20:24 |
dentarg | Don't update them at the same time, so tests can catch any bugs coming from style changes | 14:21:21 |
msp-greg | Maybe run it another branch or something? I have no idea to what level everyone wants. Personally, I'm bothered by some style issues, but if it runs, it's okay. Also, I'm used to ruby/ruby, and that would have a hard time with a style linter... | 14:23:45 |
dentarg | Locally it would be good to run the lint after the specs, make it part of the default rake task, to make people aware before pushing changes
In CI it should probably be it's own job to make it clear why the build failed | 15:16:46 |
30 Jun 2021 |
| @roguelazer:matrix.roguelazer.net joined the room. | 23:23:50 |