| 13 Sep 2023 |
 hexa | for some reason I can reproduce the too many open files issue on firefox-unwrapped.tests on staging-next right now | 09:03:48 |
 vcunat | I expect that's still the same issue. | 10:06:09 |
| hexa | yeah, I'm not saying anything on staging-next is different 🙂 | 10:07:07 |
 emily | can you check if the you can reproduce it on master/unstable now as well -- on the same machine? | 10:09:04 |
| hexa | I could, but not anymore 🙂 | 10:10:33 |
| hexa | I had been building other stuff alongside it | 10:10:46 |
| hexa | I don't have an idea how that would impact qemu's fds however | 10:11:09 |
| emily | out of curiosity: what filesystem is that machine running? | 10:14:09 |
| hexa | I think the machine where it was scheduled on runs on ZFS | 10:14:44 |
| hexa | actually all my remote builders do 🤡 | 10:15:02 |
| emily | alright, ty | 10:17:27 |
 ajs124 | something broke the thunderbird build on release-23.05:
0:13.72(B checking the wasm C++ compiler can find wasi headers... yes(B(B
0:13.81(B checking the wasm C++ linker can find wasi libraries... yes(B(B
0:13.81(B checking for dump_syms... /nix/store/vvzjvkppdwpdysy32xc4sdmycgf9vyl0-dump_syms-2.2.1/bin/dump_syms(B(B
0:13.96(B checking for getcontext... yes(B(B
0:13.96(B checking for nss >= 3.90... no(B(B
0:13.96(B ERROR: Requested 'nss >= 3.90' but version of NSS is 3.79.4(B(B
*** Fix above errors and then restart with "./mach build"
| 13:34:57 |
| ajs124 | I think maybe f81059c16b16d9826ef7fd0ecc23bf3d0888f859?
does that affect thunderbird? | 13:35:32 |
| ajs124 | ok, I just read some of that code as a refresher (I mostly forgot about buildMozillaMach) and now I'm quite sure that commit broke thunderbird | 13:38:33 |
| vcunat | Yes, confusion about nss versions. | 13:45:21 |
| hexa | sorry, I forgot to test thunderbird | 13:47:47 |
| vcunat | shouldn't firefox-esr have the same issue? | 13:48:41 |
| ajs124 | maybe. I'm testing that right now. | 13:48:57 |
| ajs124 | yup. firefox-esr-115 is also broken on release-23.05. should I open a revert or does one of you want to? | 13:51:00 |
| hexa | yeah, reverting should be sufficient | 13:51:16 |
| vcunat | I wonder how we should choose the versions. Say, do we keep the nixpkgs default nss on the old unmaintained branch in 23.05? | 13:51:19 |
| hexa | I think we keep the old LTS the release was branched with | 13:51:42 |
| hexa | so we don't touch anything else | 13:51:57 |
| ajs124 | IMO we can backport, but that will have to go trough staging iirc. so a revert should be done either way. | 13:52:23 |
| hexa | nss, nss_esr, nss_latest would be the alternative | 13:52:56 |
| vcunat | If we simply revert that commit, the ESR animals on 23.05 will get nss_latest, which will work for now but it's newer than we do on nixpkgs master. | 13:53:36 |
| vcunat | * If we simply revert that commit, the ESR animals on 23.05 will get nss_latest, which will work for now (I think) but it's newer than we do on nixpkgs master. | 13:53:51 |
| hexa | I think that is an okay quickfix | 13:53:58 |
| hexa | but we're currently lacking the correct version which is in between nss and nss_latest | 13:54:12 |
| vcunat | I confirm that firefox-esr fails the same way. | 13:54:15 |
