| [19:10:35] | <symbioquine[m]> | I think we can just run the Code Sniffer first and save the exit status in a variable, then check that variable at the end to fail the build if they failed. |
| [19:11:40] | <symbioquine[m]> | e.g. something like `CODE_SNIFFER_SUCCESS=$?` |
| [19:11:49] | <mstenta[m]> | i suppose we could also just run codesniffer after the tests? |
| [19:12:02] | <mstenta[m]> | i actually intentionally moved it before the tests at some point |
| [19:12:19] | <symbioquine[m]> | Depends whether it's important for a single run to yield both pieces of information... |
| [19:12:32] | <mstenta[m]> | because i wanted to fail early... (i had passing tests and then got annoyed that codesniffer was the only thing wrong) |
| [19:12:42] | <symbioquine[m]> | hahaha :) yeah |
| [19:12:59] | <mstenta[m]> | i mean... that does still feel like it's better, now that i think about it |
| [19:13:13] | <symbioquine[m]> | Does Code Sniffer really even need the full dev image? |
| [19:13:13] | <mstenta[m]> | i want to know fast if i that's going to fail |
| [19:13:21] | <mstenta[m]> | probably not |
| [19:13:38] | <mstenta[m]> | it's just convenient... but yea we could do something simpler |
| [19:15:02] | <symbioquine[m]> | Maybe something like https://hub.docker.com/r/cytopia/phpcs |
| [19:15:44] | <mstenta[m]> | ah interesting |
| [19:15:58] | <mstenta[m]> | yea i was even just thinking install php in the action environment itself - no docker |
| [19:16:03] | <mstenta[m]> | but maybe that |
| [19:16:11] | <symbioquine[m]> | It could be both in a separate job as a sort of canary, but move the phpcs at the end of the unit tests on the dev image as the source of truth. |
| [19:16:13] | <mstenta[m]> | 's more involved |
| [19:17:30] | <mstenta[m]> | well... |
| [19:18:05] | <mstenta[m]> | the main reason i started splitting things out to separate jobs was so that we could eventually use this action to push images to docker hub |
| [19:19:07] | <mstenta[m]> | but i dunno... maybe that shouldn't be in the same workflow as this |
| [19:19:28] | <mstenta[m]> | i was hoping it could be... so that images wouldn't get pushed to docker hub if the tests failed |
| [19:20:16] | <mstenta[m]> | hmm |
| [19:22:53] | <mstenta[m]> | currently... docker hub is building the images automatically, regardless of tests/sniffs |
| [19:23:26] | <mstenta[m]> | so building images in a separate job without running tests would be status quo |
| [19:25:23] | <mstenta[m]> | assuming we did that (hypothetically), then the remaining issue is codesniffer preventing tests from running |
| [19:26:09] | <mstenta[m]> | but maybe that's not a bad thing? 🤷 |
| [19:26:32] | <mstenta[m]> | it would be nice if they could run in parallel... but if that requires the performance hit, is it really worth it? |
| [19:27:39] | <mstenta[m]> | and if we don't run them in parallel, i feel like it's better to run the codesniffer step before the tests... so it fails sooner and you can fix that before waiting the ~10 minutes for tests to run before you find out you had an extra newline somewhere haha |
| [19:42:53] | <mstenta[m]> | <mstenta[m]> "(pushed a commit to test anyway)" <- paul121: yea looks like gzipping didn't make any difference |
| [19:46:44] | <mstenta[m]> | i'm testing one more commit, which only uploads/downloads the `farmos/farmos:2.x-dev` image via artifacts, since that's the only one we use... might shave off a bit of time |
| [19:47:11] | <mstenta[m]> | (originally both `farmos/farmos:2.x` and `farmos/farmos:2.x-dev` were included) |
| [19:47:41] | <mstenta[m]> | although... now that i think of it, the reason i did that was so that both could be pushed to docker hub ultimately... :-/ |
| [10:13:16] | * farmBOT has joined #farmos |
| [11:23:02] | * natewrench123456 has joined #farmos |
| [12:41:34] | * natewrench123456 has quit (Quit: Leaving) |
| [13:26:33] | * natewrench123456 has joined #farmos |
| [14:11:50] | * natewrench123456 has quit (Quit: Leaving) |