This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] PR other/69006: fix extra newlines after diagnostics (v2)
- From: David Malcolm <dmalcolm at redhat dot com>
- To: Bernd Schmidt <bschmidt at redhat dot com>
- Cc: gcc-patches at gcc dot gnu dot org, Richard Biener <richard dot guenther at gmail dot com>
- Date: Tue, 26 Jan 2016 10:42:07 -0500
- Subject: Re: [PATCH] PR other/69006: fix extra newlines after diagnostics (v2)
- Authentication-results: sourceware.org; auth=none
- References: <56968A2E dot 3000803 at redhat dot com> <1453752783-30299-1-git-send-email-dmalcolm at redhat dot com> <56A755F7 dot 2030604 at redhat dot com>
On Tue, 2016-01-26 at 12:18 +0100, Bernd Schmidt wrote:
> On 01/25/2016 09:13 PM, David Malcolm wrote:
> > Here's an updated version of the patch.
>
> Thanks!
>
> > Instead of testing one particular kind of output via a plugin,
> > this version of the patch adds code to gcc-dg-prune to issue a
> > FAIL for any testcase containing blank lines, with a new
> > dg-allow-blank-lines-in-output
> > directive for those test cases that legimitately emit blank lines.
> > Examples of the latter include a test using -ftime-report, another
> > using -fdump-tree-cunrolli-details=stderr, and a Fortran test
> > using -fdump-fortran-original.
>
> Is the =stderr test really necessary, or does it somehow predate the
> ability to scan dumps?
It was deliberate.
Specifically: this was in gcc/testsuite/gcc.dg/unroll-2.c
Looking backwards through history, it was renamed from
gcc/testsuite/gcc.dg/unroll_1.c
in r219675 (2015-01-15)
It was changed to:
-fdump-tree-cunrolli-details=stderr
from:
-fdump-rtl-loop2_unroll=stderr
in r216238 (2014-10-14)
It was changed to:
-fdump-rtl-loop2_unroll=stderr
from:
-fdump-rtl-loop2_unroll
in r202260:
2013-09-04 Teresa Johnson <tejohnson@google.com>
* dumpfile.c (dump_finish): Don't close stderr/stdout.
* testsuite/gcc.dg/unroll_1.c: Test dumping to stderr.
which seems to have been this discussion:
https://gcc.gnu.org/ml/gcc-patches/2013-09/msg00151.html
which was part of this thread:
https://gcc.gnu.org/ml/gcc-patches/2013-08/msg00247.html
"[PATCH] Convert more passes to new dump framework"
Hence it appears to be deliberate: to give us test coverage for dumping
to stderr, during a rewrite of dumping.
> > OK for trunk in stage 4? I regard PR 69006 as a regression, and it
> > affects all diagnostics we output (unless caret-printing is
> > disabled).
>
> Yes, I think so. Ok.
Thanks.
> > + for (int row = layout.get_first_line ();
> > + row <= last_line;
> > + row++)
>
> While you're fixing the layout here, see if that doesn't all fit on one
> line.
It does. Re-running bootstrap®rtest now with that tweak; I'll commit
it if it passes.
Dave