On Wed, Apr 29, 2009 at 12:45:58PM -0700, Jim Gifford wrote: > > Here's what I'm proposing, moving the current development branch in RC > status, no changes. Open up trunk for the 1.3.x. > Belated results of the testsuites on x86_64-64 I've built my full current desktop (xorg-7.4, a bit of gnome-2.24, a little kde-4.2.2) without any noticeable regressions - this was using 2.6.29.2 headers, with gdbm and man-db instead of man.
I just let everything continue and blew away the directories, so no ideas about any of the failures... For glibc I copied gconv-modules, but it failed with a segfault in rt/tst-timer.out. Maybe that's a pure64 thing. Binutils had 4 unexpected failures in ld : Running /building/binutils-2.19.1/ld/testsuite/ld-bootstrap/bootstrap.exp ... FAIL: bootstrap with --static ... Running /building/binutils-2.19.1/ld/testsuite/ld-elf/elf.exp ... FAIL: static preinit array FAIL: static init array FAIL: static fini array Gcc had 5 unexpected failures in libmudflap, all in libmudflap.c++/pass41-frag.cxx Coreutils is hard to parse - it seemed ok, but then fell apart i n the gnulib tests: make[5]: Entering directory `/building/coreutils-7.1/gnulib-tests' gcc -m64 -std=gnu99 -I. -I../lib -I. -I. -I.. -I./.. -I../lib -I./../lib -g -O2 -MT read-file.o -MD -MP -MF .deps/read-file.Tpo -c -o read-file.o read-file.c read-file.c:136: fatal error: opening dependency file .deps/read-file.Tpo: Permission denied compilation terminated. make[5]: *** [read-file.o] Error 1 Libtool had 4 failures: 28: Runpath in libtool library files FAILED (runpath-in-lalib.at:61) 37: localized compiler messages FAILED (localization.at:50) 47: enforced lib prefix FAILED (need_lib_prefix.at:170) 73: Run tests with low max_cmd_len FAILED (cmdline_wrap.at:43) Automake had 3 failures: FAIL: check10.test FAIL: cond35.test FAIL: lex3.test Findutils had 1 failure in locate: FAIL: old_prefix: locate failed, `tmp/locatedb': No such file or directory Grep reported 3 of 14 tests failed, ISTR that is normal on this version Module-init-tools no longer reports success or failure, I assume it was ok For a change, Vim actually reported 'ALL DONE' without any error - that's a first for me. Overall, the build works and has no regressions (the kernel seems unresponsive when c++ apps are compiling, and sometimes I get a storm of mouse events, but that's just the way it is with 2.6.29 and gcc-4.3). If I get time, I'll try a default-32 ppc64 (i.e. only 32-bit userspace after the toolchain) - knowing my luck I'll probably bork it. Unfortunately, I no longer have time to attempt to maintain multilib desktop scripts, so I don't have any interest in building a full multilib clfs on that dog. Maybe I'll retry x86_64-64, to see if I can shed any light on the failures, but I don't think that is very likely. ĸen -- das eine Mal als Tragödie, das andere Mal als Farce _______________________________________________ Clfs-dev mailing list [email protected] http://lists.cross-lfs.org/listinfo.cgi/clfs-dev-cross-lfs.org
