Le 21 juil. 2014 à 15:03, Andre Vehreschild <ve...@gmx.de> a écrit :
> Hi Dominique,
>
> thank you very much for your comments. I really appreciate them.
;-)
> Unfortunately I am contracted only for a limited number of around 6 bugs. The
> control on which bugs to pick is done by compiling a project and resolving the
> bugs that occur in order of their appearance.
If so, you should resolve (with your "client") the legal issues as soon as
possible.
> I have reworked the test:
>
> - results are now checked automatically against their expected value, and
> - unexpected results lead to an abort.
>
> I hope this will do. The test for sure can be made more elegant, but my
> Fortran
> is to limited for that. I consider it pays more to invest my time into
> resolving
> bugs, then into writing most elegant test. What do you think?
The test is fine with me (I have regtested it). You don’t need the line
+! { dg-final { cleanup-modules "m" } }
This is done automatically by the gfortran testing machinery (I have checked
it).
> The comment in interface.c is changed to:
>
> /* Only check ranks compatibility, when the actual argument is not a
> reference of an array (foo(i)). A reference into an array is assumed
> when actual->ref is non null. */
> if (actual->ts.type == BT_CLASS && CLASS_DATA (actual)->as
> && !actual->ref
> && CLASS_DATA (actual)->as->rank == symbol_rank (formal))
> return 1;
Although English is not my native language, I would prefer
/* Check ranks compatibility only when the actual argument is not a
a reference to an element of the array, i.e., when actual->ref is null. */
Someone else should give you the formal approval.
Thanks for the patch,
Dominique
> The complete patch with the changes advised is attached in pr60414_3.patch.
>
> Again, Bootstrapped and regextested on "x86_64-unknown-linux-gnu" w/o any
> regressions.
>
> Regards,
> Andre
> --
> Andre Vehreschild * Kreuzherrenstr. 8 * 52062 Aachen
> Tel.: +49 241 9291018 * Email: ve...@gmx.de
> <pr60414_3.patch>
