------- Comment #9 from dominiq at lps dot ens dot fr 2010-03-04 10:14 -------
>From http://gcc.gnu.org/ml/fortran/2010-03/msg00020.html
> Currently our default namelist behaviour allows extended reads.
> In other words, reading extra data beyond that specified by the array
> qualifiers in the namelist file.
I don't understand how this applies to the test case. From the manual I get
> Expanded namelist reads are permitted. This causes an error if -std=f95 is
> used.
> In the following example, the first element of the array will be given the
> value 0.00
> and the two succeeding elements will be given the values 1.00 and 2.00.
>
> &MYNML
> X(1,1) = 0.00 , 1.00 , 2.00
> /
>From this example I understand that X(1,1)=0.0, X(2,1)=1.0, and X(1,2)=2.0 if X
is a 2x2 array.
In the test in comment #1, the namelist provides 3 values for 3 elements and I
do not see any need for "expanded namelist reads". Note that there is no error
with -std=f95, only the "expected" behavior.
> We turn off this behaviour with -std=f95 or higher.
Apparently this is not true for -std=f2008:
[macbook] f90/bug% gfc -std=f2008 pr43228.f90
[macbook] f90/bug% a.out
1 4 7
52 6 8
261 0 9
only f95 and f2003.
> I am thinking the extended reads should be allowed with -std=legacy
> and have -std=gnu use the non extended reads as default.
I'ld vote for -std=legacy only.
--
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43228
