!PbcQeaWcgMyjVfeGQN:nixos.org

Nix Mozilla 🦊🐦🐒

175 Members
A room about a number of weird animals (also known as Mozilla products): Firefox, Thunderbird, Spidermonkey, NSS, cacert. Also a little bit of fun times, small amounts of extreme, when building weird animals. But for bugs please file GitHub issues. | Release Schedule: https://whattrainisitnow.com | Crash-Stats: https://crash-stats.mozilla.org/search/?distribution_id=%3Dnixos&product=Firefox&product=Thunderbird51 Servers

Load older messages


SenderMessageTime
19 Sep 2023
@nbp:mozilla.orgnbppatchelf, is the thing Nixpkgs uses to make binary lookup their dependencies with a custom RPATH. relrhack is something Mozilla recently added.16:16:44
@vcunat:matrix.orgvcunat I suspect it will be more like which side is easier to fix than which side has larger blame. 16:16:52
@nbp:mozilla.orgnbpvalgrind for example does not care about patchelf-ed binaries …16:18:02
@k900:0upti.meK900Ideally patchelf should produce spec compliant ELF files16:19:06
@k900:0upti.meK900And so should relrhack16:19:10
@vcunat:matrix.orgvcunatSure, ideally always, but you know how it often is in real life when you call things "hack" 😉16:20:15
@k900:0upti.meK900Well yeah16:20:24
@k900:0upti.meK900So we need to figure out what's going wrong probably16:20:38
@nbp:mozilla.orgnbp"ideally" (repeated 10 times, and faster, sounds like "dali")16:21:18
@nbp:mozilla.orgnbpOr to find the easiest path to fix the issue the fastest.16:21:50
@nbp:mozilla.orgnbpI do not know, but maybe increasing the size of the header could be a temporary hack in relrhack, to satisfy patchelf until patchelf can be fixed.16:22:26
@nbp:mozilla.orgnbp * I do not know, but maybe increasing the size of the header could be a temporary hack in relrhack, to satisfy patchelf until patchelf can be fixed. (just a guess, I haven't look the issue deeply yet)16:22:43
@nbp:mozilla.orgnbpis there a patchelf channel where people would know?16:23:34
@vcunat:matrix.orgvcunatThe only patchelf-related room I can find is this wide one? https://matrix.to/#/#systems-programming:nixos.org16:25:20
@vcunat:matrix.orgvcunat(i.e. I don't know a good channel)16:26:00
@vcunat:matrix.orgvcunat * (i.e. I don't know a good channel/room)16:26:28
@vcunat:matrix.orgvcunat * (i.e. I don't know a good channel/room/place)16:26:45
@vcunat:matrix.orgvcunatA github ticket could be used for this, though that's a less "chatty" place.16:27:39
@vcunat:matrix.orgvcunatI wonder if FF compiled by ourselves will also suffer from this, once it reaches these changes.16:28:25
@vcunat:matrix.orgvcunatThat's certainly a more important package than the binary one.16:28:48
@hexa:lossy.networkhexamaybe just invite mic92 here?16:29:11
@vcunat:matrix.orgvcunatCould be. I have no idea how many people are active around patchelf really.16:29:44
@vcunat:matrix.orgvcunatEither way, if a place doesn't exist, inviting them all here might be the easiest way.16:30:30
@hexa:lossy.networkhexa invited @joerg:thalheim.ioMic92.16:31:31
@joerg:thalheim.ioMic92 joined the room.16:31:54
@vcunat:matrix.orgvcunat👋 the mic is yours, I guess16:37:41
@joerg:thalheim.ioMic92Patchelf is fun. You fix 1 bug and create 3 new ones.16:38:25
@joerg:thalheim.ioMic92And than you get emails from people that complain about c++14 because it doesn't work on their 15 year old OS.16:39:22
@vcunat:matrix.orgvcunatNow I see that nixpkgs patchelf defaults to 0.15 whereas upstream is at 0.18.16:41:40
@vcunat:matrix.orgvcunatBut if it goes like 1 for 3, maybe it's better to default to the oldest version.16:42:04

Show newer messages


Back to Room ListRoom Version: 9