This is the mail archive of the
mailing list for the GCC project.
Re: Testsuite patch for PR testsuite/29093
- From: Janis Johnson <janis187 at us dot ibm dot com>
- To: Steve Ellcey <sje at cup dot hp dot com>
- Cc: gcc-patches at gcc dot gnu dot org, geoffk at apple dot com
- Date: Wed, 11 Oct 2006 14:46:09 -0700
- Subject: Re: Testsuite patch for PR testsuite/29093
- References: <200610112134.OAA14109@hpsje.cup.hp.com>
On Wed, Oct 11, 2006 at 02:34:54PM -0700, Steve Ellcey wrote:
> The test gcc.dg/debug/dwarf2/dwarf-file1.c fails on systems that use
> DWARF2 and also define HAVE_AS_DWARF2_DEBUG_LINE. Those systems do not
> put the string "File Entry:" in the assembly language file.
> They do put out the string ".file", so my idea is to check for either of
> these strings in the test. I think this should be OK unless systems
> that don't define HAVE_AS_DWARF2_DEBUG_LINE might also put out ".file".
> I don't see anyway for them to do that looking at the code but I don't
> have a system where HAVE_AS_DWARF2_DEBUG_LINE isn't defined so I didn't
> check it with an example.
> I tested the change on IA64 Linux and the test passes there with this
> change. This is not a regression (the test never passed on platforms
> which define HAVE_AS_DWARF2_DEBUG_LINE) but I believe testsuite changes
> are allowed during regression fix period, correct?
> If so, OK for checkin on the main line?
It looks fine to me but I don't know about DWARF2 output. If no one
objects, this can go into mainline in 24 hours.
Testsuite fixes are always appropriate except during the freeze just
before a release. If this test fails on release branches, the fix
can go into them as well.