[Patches, fortran] ping, ping, ping, ping, ping, ping, ping and ping

[Steven Bosscher]

On Sunday 16 October 2005 12:54, Paul Thomas wrote:
> PR24311 and PR24384 - scalar (character) sources for SPREAD and MERGE
> http://gcc.gnu.org/ml/fortran/2005-10/msg00350.html
> (was http://gcc.gnu.org/ml/fortran/2005-10/msg00237.html for MERGE only)
>
> PR21459 - strings of different length in a single character(len=*)
> declaration  **** Tobi, this is really yours.
> http://gcc.gnu.org/ml/fortran/2005-10/msg00275.html

OK.

> PR18022 - assignments of the kind x(:)%i = f(y(:))    ****
> http://gcc.gnu.org/ml/fortran/2005-10/msg00271.html , which, following a
> suggestion from Richard Sandiford, has become:
>
> PR24092 - recursive derived types - part 2  -------- Steve, did you get
> back to this?
> http://gcc.gnu.org/ml/fortran/2005-10/msg00180.html

AFAICT Steve already OK'ed this.

> PR20866 - recursive statement functions. ------------ This one is fun!
> http://gcc.gnu.org/ml/fortran/2005-10/msg00159.html

Just a few suggestions for this one: First of all, recursive_stmt_fcn
needs a better comment before it, explaining what each argument is.
Second, maybe the return type should just be a bool so that you don't
need that ugly REC_ST_FN ;-)

> PR20337 - PUBLIC/PRIVATE used outside a module  ****
> http://gcc.gnu.org/ml/fortran/2005-10/msg00158.html

OK.

> PR20849 and PR20853 - incorrectly used initializers for assumed size
> arrays and external objects  ****
> http://gcc.gnu.org/ml/fortran/2005-10/msg00155.html

OK.
FWIW fixing Standard issues with this kind of small patch is obvious
if you ask me.  It would be nice, though, to add references to the
sections of the Standard explaining why something like this is wrong.
(Yes, I know, we should have done that from the start when we started
developing g95/gfortran, but at least let's add references for the 
issues that get fixed now, because apparently they're not that obvious,
or we would have done this right the first time ;-)

> PR23446 - indices of formal argument lists
> http://gcc.gnu.org/ml/fortran/2005-09/msg00400.html

This is also OK, but again a little more comment explaining why we
need this would be nice (a reference to the PR is also good).

Gr.
Steven