This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: [v3] weak instantiations for AIX
- From: Benjamin Kosnik <bkoz at redhat dot com>
- To: David Edelsohn <dje at watson dot ibm dot com>
- Cc: libstdc++ at gcc dot gnu dot org
- Date: Fri, 11 Apr 2003 12:45:12 -0500
- Subject: Re: [v3] weak instantiations for AIX
- References: <20030411122506.40047cc2.bkoz@redhat.com><200304111733.NAA20492@makai.watson.ibm.com>
> Thanks. I'm just running a bootstrap now and I updated the
>libstdc++ testsuite in the middle of the bootstrap before it got to those
>tests, so we'll see how it goes.
Great.
> Now we can focus on the real execution failures. Some of the
>failures are locale problems. For the rest, debugging is difficult
>because of GDB and XCOFF stabs limitations. If there is information I
>manually can print from the testcases that can help us track down the
>problem, let me know.
Well, some of the tests have changed names/positions from your post here:
http://gcc.gnu.org/ml/libstdc++/2003-04/msg00031.html
I can fix up the ctype test. The filebuf_virtual fail (related to
locales, seemingly) seems to have gone away with the re-org, or perhaps
because I switched some of these tests to use the "C" locale explicitly.
I'll check in the ctype fix too, but I think the rest will be hard fails
that will require a bit more analysis. You might start by posting the
parts of the libstdc++-v3.log file that point out where the VERIFY macro
is finding issues.
best,
benjamin