Hi Ricardo,
I had only problems in last weeks to "force" my RH 5 box to "upgrade"
to a working gcc-g77 compiler using the rpm contributed packages.
I don't know how to solve this (I'm still a newbie in many respects)
but I can tell you some alternatives (that worked for me :-):
1) the official RH 5 support the fortran development with
the fort77 + f2c packages:
f2c translates fortran sources in c sources (but to use these
"unpolished" sources is a hell)
fort77 is a script that drive the f2c in a way to pass sources to gcc
and produce exec,
it's a sort of "compiler script".
So you call fort77 with almost the same syntax that a normal f77
compiler and it will hide
to you all the complexities in running f2c and provide an exec.
(i.e.: # fort77 -o pippo pippo.f --> would give you the pippo exec).
Also interesting to note in this approach is the possibility to use a
little program,
f2c-polisher that you can retry from www.nag.com, that lets you polish
the c-output
from f2c so to make it human-readable (not only gcc ready).
Advantages: you use a well tested piece of sw, namely the f2c
translator.
Drawbacks: you don't use a native f77 compiler!
2) the egcs way:
egcs is a fully funcional compiler with extension to the gcc-g++-g77
arch.
for more info search egcs in www.cygnus.com (it's GPL if you wonder
about),
d/l the tarred sources and read carefully the docs on how:
a) compile the compilers (eheh ;-)
b) install them in a way they doesn't interfere with your old gcc.
it's pretty easy to do (thanx to the egcs developers).
be careful that if you d/l the biggest pkg you must allow over then 200
Mb
of free space on your hard disk during the compile process! After that
you can safely remove
sources and intermediate objects, using only the binaries produced.
I tested the egcs-g77 with my 2d-hydrodinamic-lagrangian-simulation code
and I found it works pretty well. (~12000 fortran lines ;-).
Advantages: you use a native f77 compiler.
Drawbacks: it's still developing, perhaps with some bugs... but I
haven't found yet!
So... I don't really answer your question :-(
but I've documented my way to obtain a working native fortran env
on a RH 5 box, I hope this will be useful.
ciao
fede
Ricardo Píriz wrote:
>
> Hi there!
>
> I have problems executing binaries created with g77 under RHL 5.0.
> Everything seems to go fine at compilation time, but when I try to
> execute my programs I always get a segmentation fault (even for the
> simplest "hello world" program).
>
> I have installed the following versions:
>
> libc-5.3.12-25.i386.rpm
> gcc-2.7.2.3-5.i386.rpm
> gcc-g77-2.7.2.3-5.i386.rpm
>
> Everything was working fine under RHL 4.2!
>
> Any hint would be most appreciated. Regards,
> --
> _______________________________________________________________________
> Ricardo Píriz e-mail: [EMAIL PROTECTED]
> Grupo de Mecánica del Vuelo tel: +34 91 8072133
> Isaac Newton 11, PTM 3 Cantos, 28760 Madrid, SPAIN fax: +34 91 8072199
>
> --
> PLEASE read the Red Hat FAQ, Tips, Errata and the MAILING LIST ARCHIVES!
> http://www.redhat.com/RedHat-FAQ /RedHat-Errata /RedHat-Tips /mailing-lists
> To unsubscribe: mail [EMAIL PROTECTED] with
> "unsubscribe" as the Subject.
--
Semper Voster
--------------------------------------------
| |
| [EMAIL PROTECTED] |
| |
| Via W. Tobagi, 21 |
| 20143 Milano - Italy |
| |
| Phone: +39/2/89.12.65.76 |
| |
| Member of the Linux Community :-) |
| Microsoft's just a rebooting Infinity... |
| |
--------------------------------------------
--
PLEASE read the Red Hat FAQ, Tips, Errata and the MAILING LIST ARCHIVES!
http://www.redhat.com/RedHat-FAQ /RedHat-Errata /RedHat-Tips /mailing-lists
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject.
