[Patch, Fortran] PR fortran/34162: Internal procedure as actual argument

[Tobias Burnus]

Dear Daniel,

Daniel Kraft wrote:
> Currently regtesting on GNU/Linux-x86-32. Ok for trunk if no regressions?

OK - thanks for the patch.

	* gfortran.dg/internal_dummy_1.f03: Renamed from .f90, add -std=f2003.

(Renaming is not really needed as .f{90,95,03,08} does not really signifies
the Fortran standard but just that the file is in free format. I personally
restrict myself to .f90 - especially as some other compilers do not support
the .f03/.f08 extension. Though, if you prefer .f03 - go ahead.)

> Property changes on: gcc/testsuite/gfortran.dg/internal_dummy_1.f03
> ___________________________________________________________________
> Added: svn:mergeinfo

I have no idea why you gets this - but I think there should be no
svn:mergeinfo on the trunk. Consider using "svn propdel" before
committal.

 * * *

Paul Thomas wrote:
> > Good to know. Thus, implementing it essentially means to change some
> > checks - and writing a few test cases. Had we known this earlier, gfortran
> > had implemented it already a couple of years ago ...
>
> There was a thread on this a good number of years back - it was known
> that it just worked because the removal of the error had been tried.

I think that was before my time. At least when PR34162 was filled on
2007-11-20, it did not get any replies that implementing is a trivial task.

Now, you only need to tell me that one only needs to switch a knob to get
submodules working - then I see the chance that gfortran will have all
*new* Fortran 2008 features in 4.6 ;-)
Cf. http://gcc.gnu.org/wiki/Fortran2008Status
(However, before gfortran is a *full* Fortran 2008 compiler, we still need
to implement all the remaining Fortran 2003 features, which will take a
while...)

Tobias