------- Comment #47 from manu at gcc dot gnu dot org 2007-05-15 16:15 -------
(In reply to comment #46)
> I've been looking at this again recently. I have a patch that changes
> dg-error
> and dg-warning only for languages that define gcc_error_prefix and
> gcc_warning_prefix. I have tested it with C and ensured that tests for other
> languages are not affected. This will make it easy to add support for one
> language at a time.
Hi Janis, I am planning to be up and running again this weekend. I have to
start the new GSoC soon! So, could you attach or send me the patch so I can
test it myself?
> Lots of the tests that Manuel changed went from checking for errors to
> warnings
> or vice versa. I'll check those against earlier versions of GCC to see if the
> behavior really did change and when, and then report those. Some of them
> might show real bugs that ought to be fixed.
I have updated the patchset and script to a recent revision. That is, all tests
in gcc.dg that run in i686 with --enable-decimal-float are fixed by either a
patch or an automatic transformation done by the script. I'll send them to you
later.
Hmm, it didn't occur to me to run a regression hunt on the problematic tests to
check whether they are real bugs... let me know how that is working out.
Perhaps I can give you some help.
--
manu at gcc dot gnu dot org changed:
What |Removed |Added
----------------------------------------------------------------------------
Last reconfirmed|2006-03-01 02:23:32 |2007-05-15 16:15:46
date| |
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25241
