>
> 1. If you have a nice short sample like that, I'd suggest sending a
> bug report. I've found NAG to be quite good about responding to
> well-written bug reports with concise samples.
Agreed. But it was not my system-I tried to compile my program at another site.
So I had nobody to contact. Besides, my hosts planned to change to HP F90
already as I went there (Philips Eindhoven).
>From my memory, the NAG f90 was s.th. like 2.0.
> 2. But the above sample is not legal Fortran. Even if you can get
You are right. But:
> > Alas not NAG, it choked on s.th. like
^^^^^^^^^^
I was typing this too quickly, sorry. The real code was ok.
BTW, what happened in the above example was that the translated c-code contained
an else without if and thus failed to compile. I also did not test just that
short section of code. But it was that section of the module, removing this
construct avoided the problem.
Other problems were 'module procedures' for interfaces. NAG f90 claimed in some
cases that subroutines have the same argument list (it involved optional
arguments). This was false, the subroutines in question were distinct by the
type of the first or second non-optional argument. Again, for an outdated
version of NAG f90 I have no longer access to.
There was one thing I liked, however: the error messages of NAG f90 are way
better than those for the HP f90.
The whole code I ported was compiled in strict F90 mode (+langlvl=90 on HP). I
thus expected it to work pretty much out of the box and got quite embarassed.
But I do not blame it on *one* compiler vendor. Initially I avoided that
discussion by not stating the vendor, and I should have kept it that way.
I read the hint about F90 2.0 on HP-UX 11.xx by v. Hagen, I am going to follow
that one, thanks.
Klaus
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|