This is the mail archive of the libc-alpha@sources.redhat.com mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

atexit bug, RFC


I'm experiencing a bug with current CVS. Although I'm sure it's
binutils, I wanted some feedback about what works with other people.
Currently I am using gcc 2.95.3 + weak-sym patch, binutils
2.11.90.0.1, and obviously current glibc CVS. The bug is that a shared
library built against the latest glibc, when linked to by compiling
another program, fails to link missing the atexit symbol.

Both the shared lib and the program use LD=gcc, so yes, they do have all
of the start/end files, and libc_nonshared.a included.

Now, the odd part is, this test works on non-i386 (sparc and mips that I
have tested), using the same toolchain. Also, the atexit tests in glibc
pass with this glibc build. Just i386 does not appear to work.

So, can some one else confirm this with another tool chain (perhaps CVS
binutils and gcc3)? I've included a minimum test case provided by a
Debian user that shows the problem.

Thanks,
  Ben

-- 
 -----------=======-=-======-=========-----------=====------------=-=------
/  Ben Collins  --  ...on that fantastic voyage...  --  Debian GNU/Linux   \
`  bcollins@debian.org  --  bcollins@openldap.org  --  bcollins@linux.com  '
 `---=========------=======-------------=-=-----=-===-======-------=--=---'

Atexit bug


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]