| 22 May 2021 |
 sterni (he/him) | pennae: by poking the job repeatedly I've got it to happen 4 times on three different machines it seems https://hydra.nixos.org/build/143399542 | 00:59:06 |
 pennae | happens about 50% of the time on the 12 core box we have here | 00:59:53 |
| sterni (he/him) | maybe we should just pass something like min 8 $NIX_BUILD_CORES to the test suite to work around this issue | 01:00:35 |
|  jhgarner set a profile picture. | 01:45:46 |
| jhgarner changed their profile picture. | 01:48:19 |
| jhgarner changed their profile picture. | 01:49:55 |
 cdepillabout | sterni maralorn I merged in the static-linking fix (https://hydra.nixos.org/eval/1671982), as well as merging master into haskell-updates, so now Hydra will have to work through a full rebuild. Let's see what happens in a day or so. | 02:05:40 |
| sterni (he/him) | We may have to repeat that in a few days or so -- staging-next is planned to be merged tomorrow | 02:06:24 |
| sterni (he/him) | excited for the pkgsStatic fix though, we probably should add a job of that to our jobset as well, so we can monitor it and also have some binary cache | 02:07:08 |
| cdepillabout | Oh, also, how would you guys feel about adding something like pkgsStatic.haskellPackages.something to our Hydra jobset? This would prove that statically linking Haskell programs is at least minimally working. | 02:07:16 |
| cdepillabout | Oh, haha, I guess we were thinking the exact same thing. | 02:07:25 |
| sterni (he/him) | I think it is important to make it viable in the first place -- no fun using that feature if you first have to compile GHC | 02:07:47 |
| cdepillabout | Although I think we should try to get someone to volunteer to be a maintainer for it. Possibly @rnhmjoj or @nh2. I imagine fixing it might mean actually fixing the underlying static-linking machinery more than just Haskell-related stuff. | 02:08:32 |
| cdepillabout | In reply to @sternenseemann:systemli.org
We may have to repeat that in a few days or so -- staging-next is planned to be merged tomorrow Ah, then I guess I should merge in master again tomorrow.
BTW, how do you follow the status of when staging-next will be merged in?
| 02:09:41 |
| sterni (he/him) | well I'm subscribed to the PR again now, I think jon commented it today that that is the plan | 02:14:37 |
| sterni (he/him) | I think the best source of information is really searching staging next in the pull request view and reading through the discussion | 02:14:57 |
| sterni (he/him) | In reply to @cdepillabout:matrix.org
Although I think we should try to get someone to volunteer to be a maintainer for it. Possibly @rnhmjoj or @nh2. I imagine fixing it might mean actually fixing the underlying static-linking machinery more than just Haskell-related stuff. if no one does it, I guess I can volunteer, but I don't expect this to frequently break tbh | 02:15:27 |
| sterni (he/him) | I guess it may break when we finally refactor pkgsStatic, but apart from that it is fairly stable I'd hope | 02:16:01 |
| cdepillabout | My understanding is that when nh2 was working on https://github.com/nh2/static-haskell-nix, he would constantly have problems with static-linking breaking for various reasons. Although I'm not very familiar with it, so maybe it has stabalized, or my understanding is just incorrect. | 02:20:19 |
| cdepillabout | In reply to @sternenseemann:systemli.org
I guess it may break when we finally refactor pkgsStatic, but apart from that it is fairly stable I'd hope Oh, is there some sort of refactoring coming to pkgsStatic? | 02:21:02 |
| sterni (he/him) | it is currently based on an overlay being injected by default and John Ericson wants to get rid of that at some point | 02:22:20 |
| sterni (he/him) | emphasis on the at some point, no idea at which point someone will have the time and energy to work on that | 02:22:38 |
| cdepillabout | Ah, I see | 02:22:48 |
| sterni (he/him) | it probably requires some substantial reworking of stdenv and how platforms work | 02:22:54 |
| cdepillabout | In reply to @sternenseemann:systemli.org
if no one does it, I guess I can volunteer, but I don't expect this to frequently break tbh Okay, if maralorn doesn't oppose this, I'll reach out to @rnhmjoj and see if he'd be willing to volunteer. | 02:23:45 |
| sterni (he/him) | sure | 02:23:58 |
 John Ericson | i think the getting rid of the static overlay should be a fairly small thing | 04:12:43 |
 joe (he/him) | Redacted or Malformed Event | 05:48:17 |
| joe (he/him) | cdepillabout: RE https://github.com/NixOS/nixpkgs/pull/123354 the patch to fix things (in hnix (for which I'm a maintainer)) is tiny, I'd rather fix this by adding a patch to hnix, is that OK? | 07:46:15 |
| joe (he/him) | patch (and test): https://github.com/haskell-nix/hnix/pull/936/files | 07:46:21 |
