| 18 Jul 2024 |
 Atemu | Can we somehow make Mike feel the pain? :D | 08:41:19 |
 K900 | Honestly the change is pretty minimal and we can probably convince people of it as is | 08:41:41 |
| K900 | Just need to go talk to someone | 08:41:51 |
| K900 | Speaking of mesa, one day I will finish building these three stdenvs | 08:42:45 |
| K900 | One day | 08:42:54 |
| Atemu | stdenvs? | 08:43:04 |
| Atemu | What's mesa got to do with those? | 08:43:13 |
| K900 | I'm testing a mesa bump for staging | 08:43:18 |
| Atemu | Oh you mean desktop world rebuilds | 08:44:01 |
| Atemu | With our 7 chromiums | 08:44:11 |
| K900 | Not desktop world, mostly just stdenv and a few more things | 08:46:26 |
| K900 | But I'm testing it for x86_64-linux, aarch64-linux, i686-linux and both darwins | 08:46:44 |
| K900 | Oh and musl | 08:46:51 |
 ElvishJerricco | In reply to @elvishjerricco:matrix.org
yea... I'm actually rebuilding the world against a downgraded libinput right now it's times like this I really wish https://github.com/NixOS/nix/pull/8105 would get some attention... | 08:51:36 |
 woobilicious | ohh yeah, bring's back memories of nix trying to recompile my kernel and emacs and running out of 32GB of RAM. | 09:09:43 |
 mjm | I’ve run out of 32G when compiling plasma, has to limit to 8 jobs to handle it | 09:40:43 |
| ElvishJerricco | well, that's over 3 hours of building down the drain. Downgrading libinput didn't solve it | 10:10:17 |
| ElvishJerricco | I guess I'll try another mass rebuild downgrading mesa this time | 10:28:04 |
| ElvishJerricco | if it's not gnome-shell, mutter, mesa, or libinput, I'm at a loss as to what else it could be | 10:28:32 |
| Atemu | Could you do a first-parent bisect? | 10:32:29 |
| Atemu | That'd at least reveal the staging-next cycle in which the issue first appeared | 10:32:58 |
| ElvishJerricco | In reply to @atemu12:matrix.org
Could you do a first-parent bisect? ......... TIL about --first-parent | 10:38:33 |
| ElvishJerricco | that seems like it would have been useful a bunch of times | 10:38:39 |
| ElvishJerricco | I'm assuming the first parent in github merges is always the target branch? | 10:39:01 |
| ElvishJerricco | i.e. you should only ever be bisecting on commits that master actually pointed to at some point? | 10:39:28 |
| Atemu | There's also a tool by Alyssa IIRC that only considers revisions that are cached for bisect | 10:39:43 |
| Atemu | Yes, that's correct | 10:39:47 |
| ElvishJerricco | In reply to @atemu12:matrix.org
There's also a tool by Alyssa IIRC that only considers revisions that are cached for bisect IIRC that's no longer working because it relied on graham's channel history tracker thing that doesn't get updated anymore | 10:40:12 |
| ElvishJerricco | unless it's something new | 10:40:15 |
| Atemu | In a git merge, the branch "you're on" is the first parent and the revisions you're merging become the second, third etc. parents of the merge commit | 10:40:40 |
