| 18 Aug 2025 |
 raitobezarius | In reply to @winter:catgirl.cloud
fyi https://git.lix.systems/lix-project/lix/commit/858de5f47a1bfd33835ec97794ece339a88490f1 should probably be backported to a release, getting build errors on unstable bc lowdown bump it's backported to all releases but i surmise that we didn't do any further release | 14:33:30 |
 emily | I did, but it looks like I resolved the comment as all-looks-resolvable-in-principle. I did not get around to sending the most basic fixes off to Nixpkgs for review though. I was pipelining testing it with the LLVM 21 bump which has felt more time-critical; if I'd known there was a hurry I could have prioritized but with the author-submits model I was just planning to come back to it once I felt confident and you said you were looking into one of the patches in the stack so I was also waiting on that | 14:48:51 |
| emily | I'm concerned about changing the format/prefix length twice | 14:49:01 |
| raitobezarius | there's no hurry | 14:49:02 |
| raitobezarius | it's HEAD | 14:49:05 |
| raitobezarius | i want to keep stack size small because once people starts sending their own large stacks, frustration starts | 14:49:26 |
| raitobezarius | yeah i'm still looking into one patch of the stack | 14:49:42 |
| raitobezarius | it's not submitted | 14:49:44 |
| raitobezarius | i submitted everything up to that specific point* | 14:49:50 |
| raitobezarius | * i submitted everything up to that specific point | 14:49:53 |
| raitobezarius | In reply to @emilazy:matrix.org
I'm concerned about changing the format/prefix length twice the only change that went in afaik is the one that doesn't include derivation names in build directories | 14:50:42 |
| raitobezarius | we can revert that one if you feel like this is not ready yet for prime time | 14:50:52 |
| raitobezarius | and only few people would have seen it | 14:50:57 |
| raitobezarius | i'm not sure if your concern pertains actually CLs that are still unmerged | 14:51:37 |
| raitobezarius | or you are concerned about | 14:51:43 |
| raitobezarius | 
Download clipboard.png | 14:51:45 |
| raitobezarius | being merged as well | 14:51:47 |
| raitobezarius | which seems pretty safe to me | 14:51:57 |
| raitobezarius | either way, just please let me know what you prefer | 14:52:08 |
| raitobezarius | we can revert the entire series if you prefer it that way as well | 14:52:15 |
| raitobezarius | it's indeed "author-submits-the-thing" but with large stacks which receives a lot of attention and due to fast forward only, i take care of playing the merge queue until we have one | 14:52:52 |
| raitobezarius | (if we had a merge queue, your changes would have been sent too probably because they would have been ready to submit) | 14:53:28 |
 Winter | ah, okay. maybe consider doing one, or adding the patch to the module? i imagine more ppl on unstable will start complaining 😉 | 15:05:36 |
| raitobezarius | yep, we should do that, thanks | 15:05:58 |
| emily | this stack includes the changes that make build directories always nondeterministic on Darwin, to be clear | 15:31:39 |
| emily | it looks like the diff I had locally for that part of the stack was only cosmetic. I have been running the entire stack for a long while now and it has not caused me problems, though I have not tested the prefix you merged very extensively and it still produces rather long build directories. I was hoping to resolve the remaining questions soon and waiting on your review for the nested directory stuff | 15:32:50 |
| emily | but it should be harmless given I didn't have significant unpushed changes and the code is good | 15:33:02 |
| raitobezarius | got it, still, I have to ask: is there any action you want to take regarding the merge? | 15:33:15 |
| emily | still some more communication would have been nice if it was urgently wanted, I was asleep for the two hours between ping and merge | 15:33:24 |
| raitobezarius | i.e. a revert or X or Y where X or Y can be 'let's do nothing for now' :P | 15:33:26 |
