This is the mail archive of the
mailing list for the GCC project.
Re: Fix PR 16728
- From: Diego Novillo <dnovillo at redhat dot com>
- To: Benjamin Kosnik <bkoz at redhat dot com>
- Cc: "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>, "libstdc++ at gcc dot gnu dot org" <libstdc++ at gcc dot gnu dot org>
- Date: Thu, 28 Oct 2004 17:38:53 -0400
- Subject: Re: Fix PR 16728
- Organization: Red Hat Canada
- References: <email@example.com> <firstname.lastname@example.org>
On Thu, 2004-10-28 at 17:12, Benjamin Kosnik wrote:
> >When copying a pointer into another, we should also merge the flow
> >sensitive information. This first showed up on tree-cleanup-branch with
> >the more aggressive copy propagator. I'm rather surprised that it had
> >never showed up on mainline.
> Sweet. This had been broken for three months. Now it's possible to run
> the full performance testsuites on mainline...
Why aren't they run by default? Too expensive?
> >The original test is only run with 'make check-performance' on
> >libstdc++. I'd like to move/copy this test into the regular 'make
> >check', but I don't know my way around libstdc++/testsuite. Could one
> >of the libstdc++ folks give me a pointer? Where could I put this test?
> >It's a runtime test that needs '-O2 -pthread' to run.
> So is the test for this going to live in libstdc++?
It already is, I believe. The PR log suggests that the test case was
taken from one of the 'make check-performance' tests.
> Dunno. I suppose
> you'd add the smallest possible test that broke somewhere (where?), like
> the testsuite/thread tests that get run all the time.
OK, I'll try putting it there, then.