| 20 Sep 2025 |
 Grimmauld (any/all) | and yeah libxslt is failing right after, so we probably can't do that bump yet | 10:25:12 |
 Wolfgang Walther | Related: https://github.com/NixOS/nixpkgs/pull/437567
Might not be the worst pattern in general. | 10:39:06 |
 nim65s | rebuild number seems to be around 2.5k, I'll make individual PR, some on master, others on staging | 10:57:24 |
 K900 | You can just send them all to staging really | 10:59:27 |
| Grimmauld (any/all) | https://github.com/NixOS/nixpkgs/pull/444599
K900 if you ever have cycles to burn, feel free to build this towards libreoffice and tell me what all exploded on the way. I expec tthe explosions to be quite significant, but we should do the libxml upgrade before 25.11 or else we are stuck backporting fixes forever. | 11:08:43 |
| Grimmauld (any/all) | i'd truly be astonished if this worked... | 11:09:52 |
| K900 | Gonna let the Plasma rebuilds run first | 11:12:17 |
| K900 | Should be done soon-ish | 11:12:17 |
| Grimmauld (any/all) | no hurry, libxml 2.15 is something for next cycle | 11:13:31 |
| Grimmauld (any/all) | this libxml update is feeling "fun" and it truly becomes hell if this goes to staging too | 11:13:57 |
 Lun | The pattern is kinda displeasing but it's not more displeasing than the excessive rebuilds it's avoiding so it's impossible to say if it's bad :p | 11:54:31 |
 emily | the Python interface is deprecated and disabled by default now, yes? | 14:01:54 |
| emily | so why don't we just not enable it? | 14:02:14 |
| emily | python-packages.nix can override the flag | 14:02:29 |
| emily | checklist will have the billion PRs to fix packages nobody uses that should be dropped thing. I was thinking about putting up a pinned issue about "hey we expect fallout of major bumps here's the best way to handle it" | 14:04:27 |
 dish [Fox/It/She] | re: glog
ending up just reverting back to original glog for packages that take substantial effort to patch for ng-log. glog is now glog_0_7 in my tree so i'm just using that when things don't build well | 14:16:41 |
| dish [Fox/It/She] | however anything that is trivial to build I'm leaving alone | 14:16:58 |
| dish [Fox/It/She] | * however anything that is trivial to build with nglog is keeping nglog | 14:17:09 |
| emily | not sure we want two packages when we could have one | 14:26:45 |
| emily | even if one is deprecated | 14:26:50 |
| dish [Fox/It/She] | well its either keep glog for now and slowly kill remaining dependents or just never get this PR merged because I don't have the energy to fix every dependent that breaks with nglog
the other option is just abandoning this migration and leaving things as-is | 14:28:56 |
| dish [Fox/It/She] | * well its either keep glog for now and slowly kill remaining dependents or just never get this PR merged because I don't have the energy to fix every dependent that breaks with nglog
the other option is just abandoning this migration and leaving things as-is with glog | 14:29:03 |
| dish [Fox/It/She] | * well its either keep glog for now and slowly kill remaining dependents or just never get this PR merged because I don't have the energy to fix every dependent that breaks with nglog
the other option is just abandoning this migration and leaving things as-is with glog even though it's deprecated | 14:29:08 |
| dish [Fox/It/She] | oh good something fails for unrelated reasons even though I switched back to glog for it | 14:29:22 |
| dish [Fox/It/She] | wonderful /s | 14:29:24 |
| dish [Fox/It/She] | segfaults, how fun | 14:29:44 |
| emily | well, that's what I mean by saying it's not clear if ng-log is much of an improvement if it doesn't solve the problem of "glog dead" 😅 | 14:30:05 |
| dish [Fox/It/She] | yeah | 14:30:15 |
| emily | if the solution is ultimately going to be upstreams migrating | 14:30:26 |
| dish [Fox/It/She] | considering it changes the API so much i definitely wouldn't consider it a drop-in replacement | 14:30:30 |
