| 12 May 2023 |
|  @federicodschonborn:matrix.org changed their display name from Federico Schonborn to Federico Damián Schonborn. | 19:34:53 |
|  Kourtni changed their display name from Kourtni Marshall to Kourtni. | 20:05:37 |
| 14 May 2023 |
 delroth | I think there's a way to fix the Linux modules BTF reproducibility issue -- it looks like pahole (the thing that generates the .BTF section from DWARF info) uses multi-threading, and I suspect that's the cause of the unreproducibility (stuff gets processed in a ~random order). But 1. there's a pahole flag we could use (-j 1); 2. we could just patch pahole to not support that at all, or default the number of threads to 1 if e.g. SOURCE_DATE_EPOCH is defined, or similar. | 01:59:14 |
| delroth | yep, ran a --check of linuxPackages.zfs with pahole patched to default to 1 thread, and it succeeded. could be a fluke, but that's promising (my run with no patch failed "as expected") | 02:06:59 |
| delroth | https://github.com/NixOS/nixpkgs/pull/231768 | 02:14:31 |
| delroth | so uh, only Python left for the minimal iso? :) | 04:48:18 |
|  sympt joined the room. | 07:32:45 |
 raboof | I think so! Very curious to see how things will go in practice once the default is switched to python3.11, after the 23.05 branchoff | 09:18:16 |
 raitobezarius | I'd be curious to see what does it take to fix pahole to emit correctly in multi-threaded mode | 15:05:41 |
|  Pascal Bach joined the room. | 18:01:37 |
| 15 May 2023 |
|  casey © left the room. | 17:03:39 |
| 16 May 2023 |
|  dakralter joined the room. | 12:55:59 |
| 17 May 2023 |
| delroth | hmm, still getting unreproducible linux kernel builds on current staging-next | 12:27:44 |
| raboof | ow - still in the BTF sections or something else? | 12:30:36 |
| delroth | checking now, diffoscope OOMs on my test VM so I have to move the files around :p | 12:30:56 |
| delroth | (turns out ballooning doesn't work super well with panic_on_oom... sometimes memory consumption spikes too quickly and it OOMs before growing the RAM size) | 12:32:11 |
| delroth | https://delroth.net/out.html | 13:59:50 |
| delroth | well, that's not BTF, but that also doesn't tell me much | 13:59:58 |
| delroth | (it took 20min of 100% CPU time for diffoscope to generate this diff) | 14:00:14 |
| delroth | 20 bytes, that's likely some SHA-1 build-ids | 14:02:04 |
| delroth | if I had to guess... | 14:02:06 |
| delroth | Redacted or Malformed Event | 14:20:00 |
| delroth | welp, too bad for future history, can't post images and I'm not going to keep this URL alive forever | 14:20:15 |
| delroth | posted the diff in https://github.com/NixOS/nixpkgs/issues/227800#issuecomment-1551487822 and reopened | 14:20:36 |
 @rnhmjoj:maxwell.ydns.eu | In reply to @delroth:delroth.net
really? I thought ability of posting images/videos was one of the reasons for moving over from IRC | 14:23:07 |
| raitobezarius | In reply to @rnhmjoj:maxwell.ydns.eu
really? I thought ability of posting images/videos was one of the reasons for moving over from IRC there's some incident ongoing currently | 14:23:24 |
| delroth | there was a trolling/spam attack on one of the channels recentlyt | 14:23:27 |
| delroth | * there was a trolling/spam attack on one of the channels recently | 14:23:28 |
| delroth | the two ELFs that have mismatching build-ids are... the 64 bit and 32 bit VDSO, respectively | 14:42:12 |
| delroth | love when someone seems to have had the same issue, sent a patch upstream to LKML, got it reviewed, and then it never got merged: https://lkml.org/lkml/2022/6/8/432 | 14:52:42 |
