Re: make check failures with glibc-2.2.3

[Andreas Jaeger]

David Ronis <address@hidden> writes:

> I just tried building glibc-2.2.3 on an i685-linux(2.4.4)-gnu(2.2.2)
> box using
> 
> 1) gcc version 3.0 20010504 (prerelease)

make check runs fine for me with this compiler but with slightly
different flags.  I guess it's the high optimization that you use,
--enable-omitfp is not recommended since it might trigger bugs in GCC.

Since all those bugs in GCC 3.0 seem to be genuine regressions from
gcc 2.95.2, it would be helpfull if you could locate the problem until
you find out what's broken and file a bug report in GCC's gnats bug
database.


> 
> and 
> 
> 2) gcc version 2.95.3 20010315 (release)


> In both cases, I configured as:
> 
> ./configure  --prefix=/usr --enable-add-ons=linuxthreads --enable-omitfp
> 
> with no CFLAGS or CXXFLAGS environment variables set.
> 
> In both cases, the builds went fine.
> 
> In both cases I got several make check errors.  Here's what grep Error
> gives on the log.
> 
> With  gcc version 3.0 20010504 (prerelease)
> 
> make[1]: *** [testgrp.out] Error 1
> make: *** [grp/tests] Error 2
> make[1]: *** [globtest.out] Error 1
> make[1]: [annexc.out] Error 1 (ignored)
> make: *** [posix/tests] Error 2
> make[1]: *** [ex1.out] Error 139
> make[1]: *** [ex2.out] Error 139
> make[1]: *** [ex3.out] Error 139
> make[1]: *** [ex4.out] Error 139
> make[1]: *** [ex5.out] Error 139
> make[1]: *** [ex6.out] Error 139
> make[1]: *** [ex7.out] Error 139
> make[1]: *** [ex8.out] Error 139
> make[1]: *** [ex9.out] Error 139
> make[1]: *** [ex10.out] Error 139
> make[1]: *** [ex11.out] Error 139
> make[1]: *** [ex12.out] Error 139
> make[1]: *** [ex13.out] Error 139
> make[1]: *** [joinrace.out] Error 139
> make[1]: *** [tststack.out] Error 139
> make[1]: *** [unload.out] Error 139
> make[1]: *** [ecmutex.out] Error 139
> make[1]: *** [ex14.out] Error 139
> make[1]: *** [ex15.out] Error 139
> make[1]: *** [ex16.out] Error 139
> make[1]: *** [ex17.out] Error 139
> make[1]: *** [tst-cancel.out] Error 139
> make: *** [linuxthreads/tests] Error 2
> make[1]: *** [tst-aio.out] Error 139
> make[1]: *** [tst-aio64.out] Error 139
> make[1]: *** [tst-clock.out] Error 139
> make[1]: *** [tst-shm.out] Error 139
> make[1]: *** [tst-timer.out] Error 139
> make[1]: *** [tst-aio2.out] Error 139
> make[1]: *** [tst-aio3.out] Error 139
> make[1]: *** [tst-aio4.out] Error 139
> make[1]: *** [tst-aio5.out] Error 139
> make[1]: *** [tst-aio6.out] Error 139
> make: *** [rt/tests] Error 2
> make[1]: *** [resolvfail.out] Error 139
> make[1]: *** [initfirst.out] Error 139
> make: *** [elf/tests] Error 2
> 
> while with  gcc version 2.95.3 20010315 (release):
> 
> make[1]: *** [testgrp.out] Error 1
> make: *** [grp/tests] Error 2
> make[1]: *** [globtest.out] Error 1
> make[1]: [annexc.out] Error 1 (ignored)
> make: *** [posix/tests] Error 2
> 
> 
> I'm not too concerned that the development compiler fared worse, but
> the last time I tried it, there was only a single glibc testsuit
> failure.  
> 
> Is it safe to install the 2.95.3 version?

Did you apply the nis patch?  Get it from:
http://www.suse.de/~aj/glibc-2.2.3-nis.diff

and recompile again.  The testgrp and globtest failures are strange and 
shouldn't
show up - I would not install this system.

The annexc failure is normal - that's why you get the "(ignored)".

Andreas
-- 
 Andreas Jaeger
  SuSE Labs address@hidden
   private address@hidden
    http://www.suse.de/~aj