glibc 2.5: tst-cancel1 failure during testing
I started building glibc 2.5 yesterday, and all is going well, builds
fine. Then the check, and I come across this:
GCONV_PATH=/tmp/glibc-build/iconvdata /tmp/glibc-build/elf/ld-linux.so.2 --library-path /tmp/glibc-build:/tmp/glibc-build/math:/tmp/glibc-build/elf:/tmp/glibc-build/dlfcn:/tmp/glibc-build/nss:/tmp/glibc-build/nis:/tmp/glibc-build/rt:/tmp/glibc-build/resolv:/tmp/glibc-build/crypt:/tmp/glibc-build/nptl /tmp/glibc-build/nptl/tst-cancel1
Didn't expect signal from child: got `Aborted'
The rest of the library tested fine so far. This appears to be this
DWARF is a debugging format.
(Ironically a MS Word DOC at a place called "freestandards")
To get to the main point of my question, is this just some extra
support that's being added into glibc at 2.5, and hence OK to ignore
this failed test case and go ahead with the install, OR, is this now
something *required* to exist in the rest of the gcc toolchain,
because it is now in glibc 2.5?
I'm hoping this is a very obvious question to someone out there,
because reading what little is on the 'Net about it doesn't tell me
alot (glibc 2.5 was just released the tail end of last month, I don't
think many if any distros are shipping it as of yet, although I could
be wrong). I posted about this in other forums, but so far no luck.
Is this trivial, or an install-stopper? Any other time building glibc
it's passed the tests, or was due to something in my environment that
a case failed (and was corrected), so I don't want to leave 2.4 behind
unless I'm certain about this failure.