This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [Patch, Fortran] PR 48174+45304: No varags if interface is known
- From: Tobias Burnus <burnus at net-b dot de>
- To: Bob Deen <Bob dot Deen at jpl dot nasa dot gov>
- Cc: gcc patches <gcc-patches at gcc dot gnu dot org>, gfortran <fortran at gcc dot gnu dot org>
- Date: Fri, 25 Mar 2011 19:21:39 +0100
- Subject: Re: [Patch, Fortran] PR 48174+45304: No varags if interface is known
- References: <4D885C23.1070206@net-b.de> <4D88CD63.4020207@jpl.nasa.gov>
Dear Bob,
Bob Deen wrote:
Our code does not declare these routines to Fortran, therefore the
procedure interface is unknown and this patch "should" not bother
anything.
Well, it also affects:
SUBROUTINE foo()
END SUBROUTINE foo
and
REAL FUNCTION bar()
END FUNCTION bar
that is procedures without arguments, including functions which do not
return an argument. However, without looking at the debug output
(DWARF), it should not be visible for the generated program - and it
should not cause any issues.
But for the related issue... deducing the interface from usage is
likely to break things badly in my case (and doesn't sound very robust
in any case). Unless I'm misunderstanding the application of the issue?
I think deducing the interface should work fine; I think C does the
same, if no prototype has been specified.
I am not sure what it means for your test case [1], thus, one probably
should test it before. However, no patch to address that issue exists,
yet. I have added your test program as comment to PR 44471 [2]
[1] http://gcc.gnu.org/ml/fortran/2010-05/msg00330.html
[2] http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44471
Tobias