[Bug fortran/32737] internal compiler error: Segmentation fault

["burnus at gcc dot gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>]

------- Comment #8 from burnus at gcc dot gnu dot org  2007-07-12 18:41 -------
> Concerning the compiling sequence: I am aware that the program interpfxlv 
> needs the three modules (parametri, thom_mod and spline) to work, but I was
> wondering if there is a way to tell gfortran to follow the right sequence on
> it's own.

I'm not aware of any compiler able to do so. Most compilers are single pass and
need the information right away.
One could create a script which spits out an ordered file list: Simply grepping
for the MODULE name and for the USE statements, building a tree and starting to
write the filenames starting from the leafs of that tree.

> The problem is: why do I get the final internal compiler error
> once I have the three modules in the directory?

Ups, I missed that. An internal compiler error should never ever happen; but it
happens for valid usage this really bad. (However, I cannot reproduce it here
on x86-84/Linux; if the following does not work, some gfortraner with Windows
has to debug it.)

Could you try with a newer gfortran:
http://gcc.gnu.org/wiki/GFortranBinaries#Windows
just to rule out that is has been fixed in meanwhile.

Otherwise:

a) Does it work if you use the "right" order (see comment #6)

b) If you compile it one by one in alphabetical order (with .mod present),
where does it break?

c) If you have by chance a valgrind or a debugger (breakpoint "fancy_abort" or
"internal_error"), a backtrace or valgrind output would be highly welcome too.
(Run gfortran -v -c file.f95 and use the line which contains "f951" in the
debugger/valgrind.)


-- 

burnus at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-valid-code


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=32737