Summary: | emerging wxGTK fails with internal errors? | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Starling <gentoo-bugs> |
Component: | [OLD] GCC Porting | Assignee: | Please assign to toolchain <gcc-porting> |
Status: | RESOLVED DUPLICATE | ||
Severity: | blocker | ||
Priority: | High | ||
Version: | unspecified | ||
Hardware: | x86 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Starling
2005-07-17 09:51:11 UTC
(In reply to comment #0) Gah, it's emerge wxGTK, not emerge wxgtk to reproduce. Yes I did test that. Using gcc-3.4.4-hardened (since that is what it looks like you are using), and your CFLAGS, to compile wxGTK-2.6.1, worked for me. Does it always fail in the same place for you? (In reply to comment #2) > Using gcc-3.4.4-hardened (since that is what it looks like you are using), and > your CFLAGS, to compile wxGTK-2.6.1, worked for me. Does it always fail in the > same place for you? Let me just try the emerge again... ./bk-deps i686-pc-linux-gnu-g++ -c -o basedll_cmdline.o -I.pch/wxprec_basedll -D__WXGTK__ -DwxUSE_GUI=0 -DWXMAKINGDLL_BASE -DwxUSE_BASE=1 -fPIC -DPIC -DWX_PRECOMP -DNO_GCC_PRAGMA -DGTK_NO_CHECK_CASTS -Ilib/wx/include/gtk-ansi-release-2.6 -I/var/tmp/portage/wxGTK-2.6.1/work/wxWidgets-2.6.1/include -I/usr/include/gtk-1.2 -I/usr/include/glib-1.2 -I/usr/lib/glib/include -D_REENTRANT -I/usr/X11R6/include -D_FILE_OFFSET_BITS=64 -D_LARGE_FILES -D_LARGEFILE_SOURCE=1 -O2 -O2 -march=athlon-xp -fomit-frame-pointer -pipe -pthread -I/usr/include/SDL -D_REENTRANT -Wall -Wno-ctor-dtor-privacy /var/tmp/portage/wxGTK-2.6.1/work/wxWidgets-2.6.1/src/common/cmdline.cpp i686-pc-linux-gnu-g++: Internal error: Killed (program cc1plus) Please submit a full bug report. See <URL:http://bugs.gentoo.org/> for instructions. No it does not. o.o In fact, when I go into the build directory manually and type make repeatedly, the internal error doesn't usually happen the second time it tries to compile the file. But then it happens on another file! wtf? Interesting. In my kernel logs, when gcc fails with this persnickety internel error, I see lines like this: Aug 1 00:43:28 localhost grsec: denied resource overstep by requesting 4096 for RLIMIT_CORE against limit 0 for /usr/libexec/gcc/i686-pc-linux-gnu/3.4.4/cc1plus[cc1plus:172] uid/euid:0/0 gid/egid:0/0, parent /usr/i686-pc-linux-gnu/gcc-bin/3.4.4/i686-pc-linux-gnu-g++[i686-pc-linux-g:30300] uid/euid:0/0 gid/egid:0/0 Unfortunately, the grsec people said that was only reporting an error, not causing one. They said that without grsec the above problem would just dump core every time, and not be logged. Actually, by setting 'ulimit -c unlimited', the above error (requested a core larger than the limit of 0) does not occur. So I suppose this isn't very helpful, but maybe someone more knowledgeable than me will find it useful. I have a core for the gcc, but no way to analyze that core since you just can't debug on hardened. All the address PIEing, and randomizing messes up gdb, and I don't have time to compile a totally separate instance of gcc without any of those gdb messing up thingies. Just poking around, it looks like the Red Hat people (who ownz0rz gcc) are having trouble with this too! https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=123119 I don't know why this only happens with wxGTK still. Heck, not even sure at this point: it might be some catastrophic failure. My hard disk is fairly new though. Well, if it is not happening in the same location everytime, that could mean its a hardware issue. Can you try running memtest86+ for a couple hours and see if that reports any errors back? |