This is the mail archive of the
mailing list for the GCC project.
Re: What is the most suitable way to exclude test cases for specific targets
- From: Janis Johnson <janis187 at us dot ibm dot com>
- To: =?iso-8859-1?Q? Bj=F6rn=20Haase ?= <Bjoern dot M dot Haase at web dot de>
- Cc: gcc at gcc dot gnu dot org
- Date: Tue, 4 Jan 2005 10:26:07 -0800
- Subject: Re: What is the most suitable way to exclude test cases for specific targets
- References: <firstname.lastname@example.org>
On Sun, Jan 02, 2005 at 03:35:25PM +0100, =?iso-8859-1?Q? Bj=F6rn=20Haase ?= wrote:
> after scanning the archive of the gcc ML, I have stepped over the statement, that explicit use of
> expect scripts, e.g. the use of .x files, is deprecated. For this reason:
> "If not using .x files, what is the most appropriate method for skipping test cases completely
> for selected targets ?"
I've been meaning to document the dg- directives used in GCC tests and
explain how to mark tests as unsupported or expected failures. There
have been other similar questions recently, so this is definitely needed
and I'll try to get something submitted by the end of the week.
> My present aim is to get the test suite functional for the avr platform. Currently the real compiler
> bugs are masked out by the huge number of unsupported test cases that show up
> as failures. (I'd first like to concentrate on the tests for c and proceed with the g++ tests later on.)
> I have had a look at the test case sources so far in order to find templates for dg magic comments.
> Unfortunately, I did not find anything suitable so far. There exists thread suggesting to use "dg-skip-if".
> The method described there, however, seemed not to work.
I recently found out that dg-skip-if wasn't looking at the correct set
of compiler flags and fixed that on 2004-12-22. Please let me know what
problems still exist.