Fortran.el

[Tom Browder]

On Sat, Jan 17, 2009 at 12:02 PM, Tobias Schlüter
<tobias.schlueter@physik.uni-muenchen.de> wrote:
...
> This is independent of fortran.el, as the output of your compilation is
> handled by compilation-mode.  We changed our locus output a few years back
> to coincide with the GNU standards and therefore it should just work™ (I
> verified that it does with emacs).  If you're using a not-too-old version of

Tobi, I just looked more closely at the output differences between
gfortran and g++ and see a significant difference.  That may be why
xemacs compilation-mode.el doesn't do the right thing.

Example:

gfortran:
======

Error: Unexpected data declaration statement at (1)
fmt_test.f:484.6:

      atan2d = rad2deg(angle_radians)
     1

g++:
===

xemacs_test.cc: In function 'int main()':
xemacs_test.cc:7: error: expected initializer before 'int'

Notice the g++ error line gives source file, line, and error on one
line while gfortran follows the error line with the source file and
line number on the following line.  That may be the show stopper.

I know this problem has existed for years (probably since the start of
gfortran--I'm not sure it existed with g77) but, not having to do much
fortran then, shrugged and accepted it since no one else seemed to
complain.  Now I wish it were fixed.

Is there any reason the output format is so different between the two?
  Even when using the line wrapping mode of c++ the output formats are
grossly different.

Thanks.

-Tom