This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug target/32830] shared library create by hppa64-hp11.11 can't run.
- From: "danglin at gcc dot gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: 14 Oct 2007 16:23:27 -0000
- Subject: [Bug target/32830] shared library create by hppa64-hp11.11 can't run.
- References: <bug-32830-14862@http.gcc.gnu.org/bugzilla/>
- Reply-to: gcc-bugzilla at gcc dot gnu dot org
------- Comment #8 from danglin at gcc dot gnu dot org 2007-10-14 16:23 -------
Regarding comment #4, I don't believe the warnings are a problem. hppa64-hpux
does not support undefined weak.
Regarding comment #6, '-m64' is not a valid gcc option for the hppa64-hpux
target. Thus, it appears littlestar is using a wrapper around 32 and 64 bit
gcc builds. This is likely the cause of the problem.
I linked a simple test program with the command options given in comment #6
and didn't see the problem (i.e., -lgcc_stub is added to the link command).
You can see this by adding '-Wl,-debug' to the gcc command.
Regarding '_Jv_RegisterClasses' and '__cxa_finalize', these currently don't
do anything. The stubs in libgcc_stub.a simply are there to satisfy references
to these routines generated in crtbegin.o. These result from using the
generic crt code in crtstuff.c. This could be avoided if we provided our
own crt code.
'__cxa_finalize' probably will never be used. However, '_Jv_RegisterClasses'
will be used when we complete porting libffi and libjava. Since you are
porting a native java library, you may want to look at your own implementation
of '_Jv_RegisterClasses'.
--
danglin at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|WAITING |RESOLVED
Resolution| |WORKSFORME
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32830