This is the mail archive of the fortran@gcc.gnu.org mailing list for the GNU Fortran project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [patch, fortran] PR35420, PR36421, PR35422 g0, gw.d, and a0 format descriptors


Jerry DeLisle wrote:
> 2008-06-07 Jerry DeLisle <jvdelisle@gcc.gnu.org>
> 
>	PR fortran/36420, 36422
>
Please use one line per PR.


+	    return gfc_notify_std (GFC_STD_F2008, "F2008 Feature: 'G0' in "

For consistency, please use "Fortran 2008:".
(And update the third test case.)

Actually, you could also change:
   gfc_error ("F2003 Feature: ROUND= specifier at %C not implemented");


+      if (t == FMT_G && u == FMT_ZERO)
+	{
+	  if (compile_options.allow_std < GFC_STD_F2008
+	      || dtp->u.p.mode == READING)
+	    {
+	      fmt->error = zero_width;
+	      goto finished;

I'm not quite sure whether one should do run-time standard conformance
checking: Assume this occurs in a rarely used branch of the code
then the program all of a sudden stops working and no source code can be
found (e.g. commercial application). Additionally, it is not clear from
the error message that this is a -std= problem.
Therefore I personally like compile-time -std= checks, but dislike
run-time standard-conformance checks. I leave it to you whether you keep
the check. (If you remove the GFC_STD_F2008 check, you need to change the
write into a read in test case 2.)

Otherwise the patch is OK.

Tobias


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]