| 1 Jun 2024 |
 raitobezarius | Qyriad: jade_ we removed the testing code for multiple versions of Nix mentioned in https://git.lix.systems/lix-project/lix/issues/33 | 18:05:37 |
| raitobezarius | are our plans still to restore them somewhat? | 18:05:43 |
| raitobezarius | before 2.90 | 18:05:46 |
| raitobezarius | or should we go and release 2.90 given that we have empirical confirmation that they work | 18:05:56 |
| raitobezarius | and move this to 2.91 | 18:05:58 |
| raitobezarius | (a release blocker of 2.91, that is) | 18:06:04 |
 K900 | I keep saying that but Gerrit is terrible except for ONE THING it does really well and the fact that people keep using it says a lot about how important the one thing actually is | 18:06:04 |
| K900 | (and how bad literally everyone else is at it, for some reason) | 18:06:28 |
|  ⚠️ eldritch horrors operating in this area ⚠️ joined the room. | 18:08:54 |
| ⚠️ eldritch horrors operating in this area ⚠️ changed their display name from pennae to ⚠️ eldritch horrors operating in this area ⚠️. | 18:09:06 |
|  nyanbinary joined the room. | 18:21:05 |
| raitobezarius | jade_: is there anything i can do to help you on the release engineering scripts you are working on? or take some efforts off you? | 18:25:29 |
| raitobezarius | * jade_: is there anything i can do to help you on the release engineering scripts you are working on? or take some load off you? | 18:25:38 |
 jade_ | In reply to @raitobezarius:matrix.org
jade_: is there anything i can do to help you on the release engineering scripts you are working on? or take some load off you? i am currently busy with guests here, but hm. i think one thing that would be good is if the installer build stuff were somehow nicely integrated into the overall releng scripts. but idk, hard to really point at easily separable tasks. | 18:33:54 |
|  @draupnir:lix.systemschanged room power levels. | 18:38:10 |
| jade_ | I would also like to not need an arm mac to build the installer | 18:39:38 |
| jade_ | at least for testing | 18:39:48 |
| raitobezarius set the room topic to "(Technical) development of Lix, the package manager, a Nix implementation. Please be mindful of ongoing technical conversations in this channel.". | 18:51:27 |
 Qyriad | In reply to@kloenk:kloenk.eu
only getting a You do not have the can-perform-mutations buildbot role. I just want to see some log what actually failed Currently only existing contributors can see the buildbot logs — just becauase buildbot as it is doesn't have any way to allow only certain people to do things like forcibly cancel or restart arbitrary builds, which is something we intend to fix | 21:19:19 |
 kloenk | ah fair. can you tell me what it says? localy the check (only on x86 linux, everything else does not have a test) does succeed. so unsure what to do, or if it's just flakey | 21:21:54 |
| Qyriad | ugh, a remote builder failure. restarted the build | 21:23:52 |
| Qyriad | didn't realize that was your CL; we'll fix your permissions | 21:24:21 |
| ⚠️ eldritch horrors operating in this area ⚠️ | the remote builder should get fixed again, sorry <,< | 21:25:29 |
| Qyriad | there you go; you might have to log out and log back in, but you should have buildbot perms and issue triage perms in forgejo | 21:25:31 |
|  dadada joined the room. | 21:26:25 |
| kloenk | ah thanks works. good to hear that it's nothing with my change :) | 21:48:59 |
| Qyriad | as heads up, your change will probably be deferred until after we finish 2.90 | 21:49:36 |
| kloenk | No worries. I waited 4 years in nixcpp :p | 22:06:15 |
|  Charles joined the room. | 22:31:41 |
 ma27 | Question: is there any way in a DerivationGoal to find out whether you're in a daemon or in a nix-build as a root? I'm a little out of ideas since the store is always a LocalStore here and I didn't manage to write dynamic_cast<SimpleLogger> since that's not implemented in any of the headers.
Context: a while ago I submitted https://gerrit.lix.systems/c/lix/+/913 to get SSH errors in the stderr of a nix-build and I think I found a nicer solution than this (using logger->log basically in DerivationGoal::tryBuildHook).
THe idea is to still use writeStderr (to have errors actually appear in nix-daemon stderr) and to use logger to make it appear in the client's stderr. Then we can probably get rid of the log-fd=4 thingy, at least for ssh-ng (for legacy ssh tests are breaking badly and I'm not entirely sure why yet).
Any ideas? Or is my idea even reasonable? | 22:46:03 |
