On Tue, Aug 15, 2006 at 12:42:41PM +0100, Ken Moffat wrote:
>
> I'll take a look at the Makefiles later. Thanks for the
> reference.
>
Couldn't work out why that one test was using gcc instead of gcc
${BUILD32} so I used sed to fix the flags for it. But man, these
tests are fragile - this time I got a new failure in
posix/tst-nanosleep.out :
nanosleep didn't sleep long enough
Never seen that fail before. I'm starting to think we ought to
reword what we say about the glibc testsuite - it plainly hasn't
rceived vast testing on arches other than x86 outside of fedora
(who have a shedload of patches in their glibc), e.g. the problems
with 32-bit multilib locales, and this textrel test (which probably
doesn't error on fedora because the default gcc there is 32-bit).
I'll fix up textrel on ppc64 in the book, then I suppose I need to
start a fresh build - the absence of the iconvdata/tst-tables.out
failure with a broken bash makes me wonder if our fix for bash
provokes that error. I certainly respect people who write these
tests, but I'm increasingly dubious about how useful it all is.
Ken
--
das eine Mal als Tragödie, das andere Mal als Farce
_______________________________________________
Clfs-dev mailing list
[email protected]
http://lists.cross-lfs.org/cgi-bin/mailman/listinfo/clfs-dev
