This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Re: Parentheses expressions as actual arguments of elemental procedures
- From: dominiq at lps dot ens dot fr (Dominique Dhumieres)
- To: fortran at gcc dot gnu dot org
- Cc: burnus at net-b dot de, d at domob dot eu, paul dot richard dot thomas at gmail dot com, mikael dot morin at tele2 dot fr
- Date: Tue, 11 Nov 2008 11:34:24 +0100
- Subject: Re: Parentheses expressions as actual arguments of elemental procedures
Mikael,
With the patch in
http://gcc.gnu.org/ml/fortran/2008-11/txt00003.txt
the original test for pr35681 and variants work without regression.
> Now it's working with operators.
> Still to be worked on: vector subscripts, functions and
> array-constructors as actual argument.
I have some tests with vector subscripts of the kind
'call foobar (y(1), y)' or 'call nonassign (x(2:3), x(1:2))' that do not
generate temporaries with the above patch while they did with the patch in
http://gcc.gnu.org/ml/fortran/2008-11/txt00001.txt
I think these cases are invalid, but that creating temporaries for them
won't hurt. However, unless you have example(s) of valid cases, this
should probably not delay the fix for the valid cases.
I don't understand what you mean by "functions and array-constructors".
Could you give some examples?
Thanks for the patches.
Dominique