From: nimbus@nospicedham.XXX.invalid   
      
   Frank Kotler dit dans news:oirtl0$285$1@dont-email.me:   
      
   (...)   
   > I found another example - much the same as what Bogdan   
   posted. I'm a   
   > little confused by the Fortran parts. In the "equals"   
   example I posted,   
   > the assembly code returns the result on the FPU stack. I   
   don't see where   
   > we asked for that. The "summ_" example "seems" to ask for an   
   8 byte   
   > integer, but g77 (I found I do have it) complains if I   
   exceed a signed   
   > dword (2 billion and change). Well, I don't claim to know   
   Fortran.   
      
   I haven't had the time to look at your example in detail and   
   "reprocess" it - I think there may be mistakes in the   
   calculations re. the offsets of arhguments and stack   
   adjustment, plaus indeed the use of an FPU instruction that is   
   indeed puzzling on the surface of it.   
      
   Regarding the problem you found with g77, it's because you   
   must declare the size of integers /in the Fortran text/, e.g.   
   as : 'implicit integer (I-N)'. The compiler options which you   
   used when calling gfortran have no effect w/ g77.   
      
   I'll peek at your "equals" to-morrow and shall post a revised   
   version if I succeed - not that I doubt you yourself /don't   
   need/ me or any one to "fix" it for g77...   
      
   > This could be very important. Just in case assembly language   
   doesn't   
   > make that big comeback, Fortran might, so we should be   
   ready! :)   
      
      
   --   
   NimbUs   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)