Summary: | emerge glibc fails with internal compiler error in flow.c | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | folken <folken> |
Component: | [OLD] Core system | Assignee: | Please assign to toolchain <gcc-porting> |
Status: | RESOLVED WORKSFORME | ||
Severity: | blocker | CC: | ppc |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | PPC | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: | emerge --info output |
Description
folken
2004-01-16 04:29:58 UTC
Whoops Any possiblilty to try with gcc-3.3.2-r5? please try with -O2 instead of -O3. any gcc-3.3 has problems with altivec so you should disable it if you are willing to build anything with it I tried compiling gcc3.3.2-5 alas it fails for the same reason. I retried compiling glibc with -O2 and no altivec arguments; again it failes for the same reason. Just now I tried to compile an gcc 3.2.2.-2 and yet again problem with the broken (flow.c) compiler. Looks like I am trapped. I'll have to find a precompiled package with a new (old) gcc. Anybody care to sponsor theirs? :) hmm, what binutils version are you using ? binutils-2.14.90.0.8 Its gcc not binutils breaking for some reason. Could you though try to downgrade binutils - you never know :/ Also, did you merge 2.6.0 or are using 2.6 headers lately? Created attachment 26387 [details]
emerge --info output
I have encountered the same error while merging several packages, kdegames
3.2.0 being the most recent. However, running emerge --resume usually lets the
merge finish. This would suggest to me that the problem is a race condition of
some sort (or a hardware problem, but that is unlikely.) I'm using gcc
3.2.3-r4. I'm also on a 7400 (Power Mac G4 Digital Audio). Full emerge --info
output attached.
still an issue? hansmi provided me with a gcc binary package. I then deep emerged glibc binutils and pretty much the rest of the system. My CXXFLAGS are now set to -O2 -mcpu=7400 -pipe as I heard there where issues with gcc and altivec even on gcc-3.4.1. The Issue hasn't resurfaced yet, but it was extremly annoying as practically no package would emerge without the above mentioned segfault. -fno-strict-aliasing is missing from your cflags and is required if you are using gcc-3.4.1 |