18 Jul 2024 | |||
lizmat | patrickb: I have the power to merge, but I'm seeing CI failing.. | 06:14 | |
worse: I can't even see anymore why they are failing, as the logs have expired | 06:15 | ||
perhaps we need a PR to get CI whipped into shape first ? | |||
patrickb | lizmat: Thanks for the offer! Just to be clear, I have the power to merge as well. What I need is a review. (Because we said, that the policy should be to not merge without a review.) I'll see what I can do about the CI. | 06:51 | |
Geth | cro-http: patrickbkr++ created pull request #196: Bump certs |
12:27 | |
patrickb | lizmat: ^ This should get the tests passing again. Once that's merged, I can rebase my PRs onto it. | 12:51 | |
Open question: Should we for now pause the "require review" rule for non-feature (i.e. bugfix) changes that don't change the API? (Such changes can't add irreversible "damage". They can easily be reverted.) I'm unsure this is a good idea | 13:43 | ||
or a bad idea. | |||
[Coke] | In general, I'm in favor of require review for something in the community bucket. | 13:51 | |
Which sounds like I'm volunteering to help review PRs in the bucket. (I guess I am?) | 13:52 | ||
My issue with committing to review is that, IIRC, I couldn't install all of Cro on my mac the last time I tried. | 13:53 | ||
patrickb | I think there is often less incentive reviewing PRs than there is creating them, because a PR I make scratches my itch, while a PR someone else creates scratches their itch. | 14:07 | |
And then reviewing a PR often feels like taking responsibility for a change while creating a PR feels less so. I guess that's just psychology we have to live with. | 14:11 | ||
Question is, what can we change to improve the dynamics? | 14:12 | ||
Can you provide more info on the Cro on Mac failure? We should do something about that. | 14:17 | ||
[Coke] | I don't remember the actual issue at this point, I'll try a fresh install and see. | 15:41 | |
the only reason to run with --/test is for speed, yes? I should still expect all tests to pass? | |||
doing a zef install --serial --verbose cro - (did on without serial, and it got to cro itself but appeared to hang) | 16:02 | ||
Do we need NETWORKING_TESTING set for anything in cro's testing? | 16:03 | ||
cro seems to hang on tests here: [cro] ok 11 - Got log line mentioning status code | 16:19 | ||
I'll check out that version and test in-repo | |||
That's midway through t/tools-runner.rakutest | 16:21 | ||
patrickb | It managed to get to "cro"? Given that "cro" is just the command line tool to help a bit with minting and running projects, the interesting bits seem to all have been installed. | 17:29 | |
[Coke] | ah, I think this might be github.com/croservices/cro/issues/85 | 17:51 | |
lizmat | Mojave is sooo 2019 :-) | 17:52 | |
[Coke] | I hate OS names so much. I'm on Sonoma now | 17:53 | |
lizmat | actually... so 2015 | ||
the thing is, the IRC logs server runs on a 2015 iMac with Mojave | |||
and it's based on cro :-) | |||
patrickb | So it is non-reproducably failing on Mojave onwards? | 17:57 | |
It's a failure in the "cro" distribution, it doesn't affect any Cro apps themself as "cro" is a dev tool. | 17:58 | ||
Looking at the test file, it's possible the failures are fixed by the PRs of mine. | 18:01 | ||
lizmat | sadly, that box is the only one I have running Mojave still, and it's doing production, so I'd rather not run any testing on it | 18:02 | |
patrickb | Let me revise that. I don't think the error is fixed by the PRs of mine. It's using Cro::HTTP::Client, but not reusing connections. | 18:16 | |
[Coke] | Should we get issues/commits/prs announced in here? | 22:05 | |
[Coke] opens github.com/croservices/cro-http/issues/197 | |||
oops, it's actually in cro::uri | 22:13 | ||
19 Jul 2024 | |||
Geth | cro-core: patrickbkr++ created pull request #40: Fix parsing URIs with empty path segments |
11:00 | |
cro-core: patrickbkr++ created pull request #41: Fix CI |
13:34 | ||
22 Jul 2024 | |||
patrickb | Can I get an eye on github.com/croservices/cro-http/pull/196 | 14:32 | |
coleman | looks good | 15:59 | |
23 Jul 2024 | |||
Geth | cro-http/master: 4 commits pushed by (Patrick Böker)++ | 07:42 | |
patrickb | CI now works again. I have rebased cro-http #191, #192 and #193 onto HEAD, all now pass the CI and can be reviewed. | 07:57 | |
Also cro-core #41 (Fixing CI in cro-core) passes it's test after 2 retries. That can also be reviewed. | 08:02 | ||
[Coke] | patrickb++ | 12:21 |