| 8 Mar 2022 |
|  finn joined the room. | 16:11:04 |
| finn left the room. | 16:11:52 |
| 16 Mar 2022 |
 hexa | https://github.com/NixOS/nixpkgs/pull/164102 doesn't start eval ☹️ | 18:05:54 |
 piegames | Could it be the current outage? Webhooks are marked as "Partial Outage" on https://www.githubstatus.com/ | 18:06:43 |
| piegames | * GitHub Actions too | 18:06:55 |
| hexa | ah yes, I'm stupid 😀 | 18:06:58 |
 cole-h | FYI: I'll be redeploying ofborg shortly to pick up https://github.com/NixOS/ofborg/pull/573
Barring unforeseen issues, it should be over ~quickly (but I have a commitment in ~30m and won't be watching that closely at that point)
If you notice anything wonky, please ping / DM / whatever :) | 20:19:31 |
| cole-h | OK, deploy finished, and it seems to still be working... :)
https://github.com/NixOS/nixpkgs/pull/164504#event-6254353751 🎉 | 22:25:02 |
| 17 Mar 2022 |
 nf | is it intentional that failed ofborg builds are marked as "neutral" instead of failed? | 08:14:40 |
| nf | nvm https://github.com/NixOS/ofborg/issues/286 | 08:15:20 |
| piegames | AFAIK, yes | 10:03:38 |
 7c6f434c | Graham has said many times (and correctly), that even a relatively low rate of false-positive-red will devalue red for very clear failures. Build test failures… vary in their expected impact on decisions | 10:08:55 |
| nf | fair enough | 11:52:21 |
| 19 Mar 2022 |
| hexa | https://github.com/NixOS/nixpkgs/pull/164601 unwilling to eval apparently | 14:08:37 |
| hexa |
fatal: couldn't find remote ref pull/164601/head
o_O
| 14:09:02 |
| cole-h | Oh boy... | 17:45:39 |
| cole-h | Ah, found the real issue:
This SHA and context has reached the maximum number of statuses. | 17:50:38 |
| hexa | what does that even mean | 17:52:12 |
| cole-h | It likely means, when GitHub was having their bad day, that the API was doing weird things and caused ofborg to continuously attempt to add statuses or something | 17:53:14 |
| cole-h | https://api.github.com/repos/NixOS/nixpkgs/statuses/4e07eedf4dc9dbba87d3591271d158ceb9341f03?per_page=100 | 17:56:15 |
| cole-h | : ~ ; curl -v "https://api.github.com/repos/NixOS/nixpkgs/statuses/4e07eedf4dc9dbba87d3591271d158ceb9341f03?per_page=1" 2>&1 | rg -i 'link:'
< link: <https://api.github.com/repositories/4542716/statuses/4e07eedf4dc9dbba87d3591271d158ceb9341f03?per_page=1&page=2>; rel="next", <https://api.github.com/repositories/4542716/statuses/4e07eedf4dc9dbba87d3591271d158ceb9341f03?per_page=1&page=1167>; rel="last"
| 17:57:23 |
| cole-h | So that commit has ~1167 statuses on it lol | 17:57:36 |
| cole-h | This same thing happened in the past, and I sent in a support ticket and got the above hint to look for how it's exceeding the maximum number of statuses (1000) | 17:59:31 |
| cole-h | hexa: Fetching from the pull head ref should work now :) | 18:00:57 |
| hexa | ❤️ | 18:03:45 |
| 23 Mar 2022 |
 Sandro | In reply to @7c6f434c:nitro.chat
Graham has said many times (and correctly), that even a relatively low rate of false-positive-red will devalue red for very clear failures. Build test failures… vary in their expected impact on decisions and how often to we muss grey checks by accident and break things? | 10:56:12 |
| Sandro | Either I stressed ofborg to much with staging PRs or GitHub is having issues again. Checks are taking very long... | 10:57:47 |
| 7c6f434c | In reply to @sandro:supersandro.de
and how often to we muss grey checks by accident and break things? It's GitHub, you cannot have things work actually as they should with GitHub | 11:06:29 |
| piegames | In reply to @7c6f434c:nitro.chat
It's GitHub, you cannot have things work actually as they should with GitHub cries in CODEOWNERS… | 11:07:20 |
| Sandro | In reply to @7c6f434c:nitro.chat
It's GitHub, you cannot have things work actually as they should with GitHub Its not as terrible as everyone is saying. Launchpad or gerrit have their flaws, too. | 11:26:26 |
