System is dual processor Sun Ultra2 (sparc64 sun4u SMP), 'emerge -u system' current as of 16.xii.02. After 'emerge rsync' on 17 Dec, 'emerge -p -u system' wants to update portage and [ebuild U ] sys-libs/glibc-2.3.1-r2 [2.2.5-r7] [ebuild U ] sys-devel/binutils-2.13.90.0.16 [2.13.90.0.4-r0] [ebuild UD] sys-apps/hdparm-5.2-r3 [5.3-r1] [ebuild U ] sys-devel/egcs64-sparc-19980921-r2 [19980921-r1] There seem to be problems with this, on my system at least. So far, 1. In ebuild.sh @ 395, the line reads rm-f /usr/bin/cc, when it should be "rm -f /usr/bin/cc" Since this fails, and (for me, at least, cc was a link), the next line wipes out the compiler. 2. After fixing that, try an 'ebuild glibc-... compile' It doesn't. For me, at least, you get In file included from ../sysdeps/unix/sysv/linux/sparc/sparc32/setgroups.c:2: ../sysdeps/unix/sysv/linux/i386/setgroups.c: In function `setgroups': ../sysdeps/unix/sysv/linux/i386/setgroups.c:57: `__NR_setgroups32' undeclared (first use in this function) ../sysdeps/unix/sysv/linux/i386/setgroups.c:57: (Each undeclared identifier is reported only once ../sysdeps/unix/sysv/linux/i386/setgroups.c:57: for each function it appears in.) As a sanity check, I have verified that ebuild glibc-2.2.5-r7.ebuild compile still goes OK. 3. At this point, I think I'll hold off updating anything for a day or so...
The glibc compile problem is due to the change of the keywords for sparc64 to sparc. The contents of the ebuild was not updated. The ebuild file has now been fixed, emerge rsync + emerge glibc will now work. The portage error/ fiasco is well known and ppl have jumped on it, again emerge rsync + emerge portage. Don't "fix" the ebuild.sh as it will zap gcc! It has been a funny day for gentoo, anyway, that about fixes it all. Thanks for your time, Stefan
