Okay, I've run this 3 times on clean installs from stage3-athlon-xp-1.4_rc2.tbz2, and the last time used the method posted in the newsletter to capture the whole thing. The error ends in: gnu/awt/j2d/.libs/DirectRasterGraphics.o: could not read symbols: Input/output error collect2: ld returned 1 exit status make[3]: *** [libgcj.la] Error 1 make[3]: Leaving directory `/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libjava' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libjava' make[1]: *** [all-target-libjava] Error 2 make[1]: Leaving directory `/var/tmp/portage/gcc-3.2.1-r6/work/build' make: *** [bootstrap-lean] Error 2 !!! ERROR: sys-devel/gcc-3.2.1-r6 failed. !!! Function src_compile, Line 284, Exitcode 2 !!! (no error message) Not sure that the initial symbol read error noted is the same, but it's always at line 284. Let me know of course if you'd like more of the log from the most recent attempt. I see elsewhere that r8 is supposed to be in stable now, but r6 is what I consistently have building after emerge sync and emerge -u world. USE="X gtk mbox -gnome -alsa" CFLAGS="-march=athlon-xp -O3 -pipe"
Just for fun, tried installing from stage 2. Same result in the end (although different .o file with symbols complaint): java/io/.libs/ObjectStreamConstants.o: could not read symbols: Input/output error collect2: ld returned 1 exit status make[3]: *** [libgcj.la] Error 1 make[3]: Leaving directory `/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libjava' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libjava' make[1]: *** [all-target-libjava] Error 2 make[1]: Leaving directory `/var/tmp/portage/gcc-3.2.1-r6/work/build' make: *** [bootstrap-lean] Error 2 !!! ERROR: sys-devel/gcc-3.2.1-r6 failed. !!! Function src_compile, Line 284, Exitcode 2 !!! (no error message)
Looks like I duped bug 13832 here. Same Line 284 gcc fails on, anyhow.
Have uploaded full build log to bug 13832
This is about the 6th time in 4 days emerging gcc-3.2.1-r6 (worldfile update) goes wrong. I am having problems with gaim and openoffice too, they seem to be giving an similar errormessage. /var/tmp/portage/gcc-3.2.1-r6/work/build/gcc/gcj -B/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libjava/ -B/var/tmp/portage/gcc-3.2.1-r6/work/build/gcc/ -ffloat-store -g -O2 -o .libs/jv-convert --main=gnu.gcj.convert.Convert -shared-libgcc -L/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libjava -L/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libjava/.libs ./.libs/libgcj.so -L/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libstdc++-v3/src -L/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libstdc++-v3/src/.libs -lpthread -ldl -lz -L/var/tmp/portage/gcc-3.2.1-r6/work/build/gcc -L/usr/i686-pc-linux-gnu/bin -L/usr/i686-pc-linux-gnu/lib -L/usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.1 -L/usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.1/../../../../i686-pc-linux-gnu/lib -L/usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.1/../../.. -lgcc_s -lc -lgcc_s /var/tmp/portage/gcc-3.2.1-r6/work/build/gcc/gcj -B/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libjava/ -B/var/tmp/portage/gcc-3.2.1-r6/work/build/gcc/ -ffloat-store -g -O2 -o .libs/gij -shared-libgcc gij.o -L/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libjava -L/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libjava/.libs ./.libs/libgcj.so -L/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libstdc++-v3/src -L/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libstdc++-v3/src/.libs -lpthread -ldl -lz -L/var/tmp/portage/gcc-3.2.1-r6/work/build/gcc -L/usr/i686-pc-linux-gnu/bin -L/usr/i686-pc-linux-gnu/lib -L/usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.1 -L/usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.1/../../../../i686-pc-linux-gnu/lib -L/usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.1/../../.. -lgcc_s -lc -lgcc_s ./.libs/libgcj.so: undefined reference to `_dl_cpuclock_offset' collect2: ld returned 1 exit status make[3]: *** [gij] Error 1 make[3]: *** Waiting for unfinished jobs.... ./.libs/libgcj.so: undefined reference to `_dl_cpuclock_offset' collect2: ld returned 1 exit status make[3]: *** [jv-convert] Error 1 make[3]: Leaving directory `/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libjava' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/var/tmp/portage/gcc-3.2.1-r6/work/build/i686-pc-linux-gnu/libjava' make[1]: *** [all-target-libjava] Error 2 make[1]: Leaving directory `/var/tmp/portage/gcc-3.2.1-r6/work/build' make: *** [bootstrap-lean] Error 2 !!! ERROR: sys-devel/gcc-3.2.1-r6 failed. !!! Function src_compile, Line 284, Exitcode 2 !!! (no error message)
Try with CFLAGS="-march=i686 -O2 -pipe"
Went to try the other settings, and discovered that my make.conf had been reverted. The instructions in the installation doc have you edit make.conf before emerging the system in stage two, but then do an export CONFIG_PROTECT="-*", which means that by the time it gets to gcc it's doing it with -mcpu=i686 -02 -pipe. Since -mcpu should be more liberal than -march (if I read the make.conf instructions right), then this gcc bug happens either with the flags I first reported (when I was going from stage3 without stripping CONFIG_PROTECT), or with -mcpu=i686 -02 -pipe. I'll separately report the documentation bug.
Matthew, in bug 13832 suggested I might have a hard drive problem - and that's the case. Didn't realize that mkreiserfs doesn't have any sanity checks, so that reiserfsck --check may find errors (and does here) even immediately after mkreiserfs has appeared to run successfully (and presented a partition that otherwise seems to be working okay). Too bad - until now I'd been impressed by Reiser.
