11:18 <emmett1_> hellow guys

11:18 <emmett1_> any of you got 'Segmentation fault' for firefox?

11:19 <emmett1_> also happen with my firefox-esr port

11:22 <emmett1_> https://we.tl/t-rrhdW9AkOo

11:22 <emmett1_> here file for 'strace firefox &> stracelog' if its helpful

11:28 <SiFuh> emmett1_: Not clicking that. Too many questions about accepting cookies and crap

11:29 <emmett1_> owh sh*t

11:29 <emmett1_> so how should i share that?

11:29 <SiFuh> Yeah, it wants subscription and accept cookies and review what cookies you allow and not.

11:29 <SiFuh> I use dpaste

11:29 <SiFuh> Other use another.

11:29 <emmett1_> fi;e too large to get into pastebin

11:30 <emmett1_> or hastebin

11:30 <SiFuh> File or a log?

11:30 <emmett1_> its a log file

11:30 <emmett1_> because pastebin wont accept too long log

11:30 <emmett1_> i upload file to wetransfer

11:30 <SiFuh> http://ix.io

11:31 <SiFuh> A lot of the guys use this

11:31 <emmett1_> i never use that, can you enlighten me a little how to use it

11:32 <SiFuh> Scroll down

11:32 <SiFuh> EXAMPLES <--

11:33 <emmett1_> http://ix.io/4ytK

11:33 <emmett1_> ok i hope its there

11:33 <SiFuh> Hell yeah

11:39 <emmett1_> been make a little research yesterday

11:39 <emmett1_> some say it happen on intel cpu, caused by latest mesa

11:40 <SiFuh> As in runs out of memory?

11:40 <emmett1_> some say it caused by rust's LTO

11:40 <emmett1_> runs out of memory?

11:41 <SiFuh> I was asking

11:41 <SiFuh> About the Intel and Mesa thing if it was a memory issue

11:42 <emmett1_> owh i dont know

11:42 <emmett1_> because others works fine, build fine and etc

11:42 <farkuhar> it was a recent update (either mesa or vulkan stuff) that coincided with more segfaults on my machine, but not with firefox. Instead the segfaults are happening in lagrange (the gemini browser).

11:42 <emmett1_> firefox just start crash suddenly few days ago

11:43 <SiFuh> Downgrade mesa?

11:44 <farkuhar> interestingly, a different machine (older video card) is unaffected, and lagrange runs there without segfaults.

11:44 <emmett1_> owh ok, i will try downgrade mesa

11:45 <emmett1_> does webkitgtk has anything to do with firefox?

11:45 <emmett1_> because i found this

11:45 <emmett1_> https://wiki.linuxfromscratch.org/blfs/ticket/18051

11:48 <farkuhar> emmett1_: i think firefox uses its own rendering engine, independent of webkitgtk.

11:48 <emmett1_> owh ok, just point something i found incase it helps

11:48 <emmett1_> will try downgrade mesa first

11:50 <emmett1_> will relogin

11:50 <farkuhar> emmett1_: that ticket seems limited to Intel GPUs, so it's no suprise that my Epiphany browser with an AMD GPU didn't exhibit any such flickering.

11:52 <emmett1_> still segfaults after downgrade mesa

11:56 <emmett1_> http://ix.io/4ytO

11:57 <emmett1_> with my firefox-esr running firefox through term tell me this

11:57 <emmett1_> firefox runs for few second then closed

12:03 <farkuhar> same description applies to my lagrange segfaults. At first it would crash immediately on loading, but after a mesa/vulkan version bump I got it to run for a few seconds before segfaulting.

12:04 <emmett1_> alright

12:04 <emmett1_> for easy bugs sometimes i could fix it after some research

12:05 <emmett1_> but this one i'm lost

12:05 <emmett1_> looks like i have to use other browser for now

12:45 <ukky> emmett1_: Try to log strace output with '--follow-forks' option

12:51 <emmett1_> ukky, ok now log too big to put into ix.io

12:54 <ukky> I know, it must be huge, but without tracing sub-processes, it might be hard to figure out where segfault occurs. You can try to compress the log and see if accepted.

13:59 <emmett1_> how to compress it?

14:01 <emmett1_> ukky, maybe we gotta continue this later, i gotta go

16:08 <ppetrov^> hey r0ni, any issues with setting up CRUX?

20:10 <r0ni> ppetrov^ nah it's pretty straightforward, your posts are honestly almost the only helpful info i've found on the net, good job with them. also I really like the distro for how simple yet powerful it is

20:10 <r0ni> outside of the main sites docs, that is

20:25 <ppetrov^> if you think i can halp with sth, let me know

20:40 <r0ni> i'm mostly working with the arm port, but i have setup a x86 box as well so i can cross-check fixes to ports before submitting them to the relevant upstreams, but a few things are broken on arm as-is without a fix so far

20:42 <r0ni> anyway, i'm bout to walk out for work, i'll be around more in like 13 hrs... gah :(

21:58 <farkuhar> ppetrov^: i've been tinkering with pkg-get in response to r0ni's recent questions. I managed to achieve almost a 4x speedup in the pkg-repgen script, at least on one trial.

22:00 <ppetrov^> that sound sgreat

22:00 <ppetrov^> *great

22:00 <farkuhar> time results, old version (real, user, system): 39.59s, 25.53s, 14.17s

22:01 <farkuhar> time results, latest rewrite: 10.62s, 9.99s, 0.62s

22:05 <ppetrov^> and my underscore issues were not just me, right?

22:06 <farkuhar> Right, the dashes in port names needed better handling, which was another part of my recent patching efforts. Thanks to emmett1 I had reason to test the interaction of --install-root with 'runscripts yes'. This combination was equally problematic in pkg-get as it was in prt-get, so I patched that oversight too.

22:07 <ppetrov^> any chance your fixes will make it to the official pkg-get?

22:11 <farkuhar> that's a question better suited for #crux-devel. But changes in a long-untouched code base take time to be reviewed, and this chat is the first place where anyone would have learned about them.

22:16 <farkuhar> anyway, my updated README now gives credit to ppetrov^ for discovering some of the workarounds needed with this program: https://git.crux.nu:82/farkuhar/pkg-get

22:23 <ppetrov^> thanks!