Further comments on other testsuites. This is with current LFS-svn (linux-3.8.0, file-5.13, and Bruce's suggestions for using the shipped info files from gcc, plus eudev instead of Bruce's continuing efforts to tame udev from systemd).
Much as expected. This time in glibc I got Error 1 from nptl/test-getpid2.out and rt/tst-cputimer1.out and they both escalated to Error 2, also posix/annexc.out and conform/run-conformtest.out which were '(ignored)' as I expected. For gcc I continue to get a lot of failures in gcc.c-torture which seems normal for me, maybe it's the absent static libs again. Bash was different this time - 40,41c39,40 < 4 < ./read2.sub: line 23: /dev/tty: No such device or address --- > > ./read2.sub: line 23: read: -3: invalid timeout specification You all know I don't believe that test suite failures are necessarily a sign of a problem (and passes certainly don't indicate everything is correct), so I just offer the glibc and bash results as examples of random variations. As for procps-ng, I'm giving it up - I'm out of ttys on the host system, supposedly (/dev/pts, /proc, /sys all mounted, I added /tools/bin to the end of the PATH) so it just reports 1 unsupported test at the end. I know when I'm beaten :) About half my desktop has built - a number of issues, but they all seem to be related to version increments which aren't in BLFS. Looks good (but apparently a 3.8.1 kernel will be better if you have other users - lwn reference https://lwn.net/Articles/540476/ but that might be for subscribers only at the moment, earleir article is 540082 quoting CVE-2013-1763). ĸen -- das eine Mal als Tragödie, das andere Mal als Farce -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
