This is the mail archive of the
fortran@gcc.gnu.org
mailing list for the GNU Fortran project.
Fw: namelist modifications - de-whitespaced and completed.
- From: "Paul Thomas" <paulthomas2 at wanadoo dot fr>
- To: <fortran at gcc dot gnu dot org>
- Date: Sat, 2 Apr 2005 10:04:58 +0200
- Subject: Fw: namelist modifications - de-whitespaced and completed.
Steve,
Maybe. I haven't problems with others patches or with sending
patches to fortran@. Fortunately, the rejected portions were
very small and easy to do by hand.
$ diff --version
diff (GNU diffutils) 2.8.7
$ patch --version
patch 2.5.8
I notice that in my script that applies the patches I am using the -l
option. I think that is the key to why I am not seeing the rejections.
I'll make absolutely certain that the final version patches without any hand
crafting or options associated with white-space.
I've read through trans-io.c, io.h, transfer.c, and lock.c.
I'm part way through list_read.c. I haven't seen anything
that is setting off alarms. We'll need Paul B. or Steven B.
to look over trans-io.c, because I don't understand the
tree-ssa backend stuff.
OK - I packaged the tree-ssa part so that it would be compact and would not
give them much extra work.
In my first quick read of the diff,
I notice that you included ctype.h in list_read.c or write.c
and used toupper. You don't need the header and toupper
should be TOUPPER. This is a GCC builtin function.
OK - noted.
Thanks a lot, Steve. I am checking and cross checking the test cases. I
have them all written to check the things namelist should do; testing the
errors is a bit more of a hassle than I expected.
On this latter, is it sufficient to check that the errors are caught, or is
it necessary to check the messages too?
Regards
Paul T