Hi, I did a fresh install this morning and when I was emerging xorg-server, glib was pulled in and failed to compile. I tried with PATH="/sbin:/bin:/usr/sbin:/usr/bin" . It still fails. Running lafilefixer --justfixit didn't help neither. If anyone has any clue... Thanks. Reproducible: Always Steps to Reproduce: 1.fresh install 2.emerge glib Actual Results: compile fails Expected Results: compile runs fine :)
Created attachment 228317 [details] emerge --info
Created attachment 228319 [details] build.log
Created attachment 228321 [details] environment
A segfault there, with rather sane CFLAGS, suggests a hardware failure. Though, do attach *full* build log.
lowering priority per previous comment.
Created attachment 228341 [details] full build.log Sorry I didn't thought it could be necessary the whole thing...
Does it always fail in the same way? Have you tried to re-emerge it again and you get the segfault at the same time?
(In reply to comment #7) > Does it always fail in the same way? Have you tried to re-emerge it again and > you get the segfault at the same time? > I will try this evening to re-emerge it, but I am pretty shure that when I tried with different CFLAGS (-Ox) it always gave the same error. But I will re-try in about 10 hours...
Hi, voilà, I just did a new try and it failed with an other error... Both environment and build.log in attachment.
Created attachment 228643 [details] second buil.log The build.log from my second try
Created attachment 228645 [details] second environment Environment file for my second try.
These "random" segfaults are usually caused by hardware failures. This bug is INVALID then. My only suggestion is to try to ask for help in http://forums.gentoo.org Also, if you are using "ondemand" cpu governor, try to use "performance" one instead (for example, in one of my machines, my CPU seems to not like "ondemand" at all causing lots of random segfaults that are usually noticed in large compilations (like xulrunner, gtk+, glib, openoffice...)
Hi, and thanks for those suggestions. Just a last question though, does "random cpu governor" means those automatic detection CFLAGS ? That I should replace with a family specific CFLAG ?
(In reply to comment #13) > Hi, and thanks for those suggestions. Just a last question though, does "random > cpu governor" means those automatic detection CFLAGS ? That I should replace > with a family specific CFLAG ? > No, that flags are chosen by gcc and they usually are ok (I use -march=native on all my machines without problems). That "governors" are simply related with frequency and voltage scaling for energy saving. You can read much more about it in your kernel sources at /usr/src/linux/Documentation/cpu-freq