This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [patch, fortran] Unify / fix error messages for matmul
- From: Janne Blomqvist <blomqvist dot janne at gmail dot com>
- To: Thomas Koenig <tkoenig at netcologne dot de>
- Cc: Fortran List <fortran at gcc dot gnu dot org>, GCC Patches <gcc-patches at gcc dot gnu dot org>
- Date: Sun, 16 Sep 2018 21:07:41 +0300
- Subject: Re: [patch, fortran] Unify / fix error messages for matmul
- References: <52d026ff-fcf7-335a-84bb-68f25ae728e8@netcologne.de>
On Sun, Sep 16, 2018 at 6:55 PM Thomas Koenig <tkoenig@netcologne.de> wrote:
> Hello world,
>
> looking through the code for run-time checking of matmul, I found that
> a) the code was not very clear, and b) due to this, at least one case
> was not checked correctly.
>
> So, here is a rewrite of that section. I have also made sure that
> the error messages now agree between library and compiler-generated
> error messages. The macros for error messages also contain the
> _ magic, so I hope that translation will be easier.
>
> I may also have fixed other bugs in that section, I didn't check
> very thoroughly.
>
> There was one suggestion in PR 37802. For code like
>
> c = matmul(a,b)
>
> where the boundaries of c were not correct, this is actually an error
> in the assignment, so I have made the error message match what we emit
> for normal array access out of bounds (like a(1:3) = b(1:2)).
>
> Regression-tested. OK for trunk?
>
Ok, thanks.
--
Janne Blomqvist