g77 egcs bug ?

[ALLEON Guillaume]

N8TM@aol.com wrote:

> g77 -c -g gemmh.f
> g77: Internal compiler error: program f771 got fatal signal 11
>
> I am using egcs-1.1.1 [egcs-2.91.60 19981201] (but I think it is the
> same with
> egcs-1.1.2). My platform is : SunOS 5.6 i86pc.
>
> The silly piece of code was this (a stupid wrapper !)
>
>       SUBROUTINE _GEMMH_ ( TRANSA, TRANSB, M, N,
>      &                     K, ALPHA, A, LDA,
>      &                     B, LDB, BETA, C, LDC )
> c
>       implicit none
> c
>       CHARACTER*1 TRANSA, TRANSB
>       INTEGER     M, N, K, LDA, LDB, LDC
> c
>       double complex     A(LDA,*)
>       double complex     B(LDB,*)
>       double complex     C(LDC,*)
>       double complex     ALPHA
>       double complex     BETA
> c
>       CALL _GEMM_ ( TRANSA, TRANSB, M, N,
>      &              K, ALPHA, A, LDA, B, LDB,
>      &              BETA, C, LDC )
> c
>       RETURN
>       END
>
> the interesting thing is that it is OK without -g
>
> Guillaume
>
> My g77's have no trouble compiling this on i[56]86 after I remove the html
> characters, either egcs-1.1.2 or egcs-19990502, but I don't have SunOS.
>
> Tim
> tprince@computer.org

You are right, this is perfectly working on Linux systems running either
egcs-1.1.1 or
egcs-1.1.2. I got this problem only on SunOS based systems.

Guillaume