Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.milw.wisc.edu!bionet!agate!ucbvax!DBTHRZ5.BITNET!BTP408
From: BTP408@DBTHRZ5.BITNET
Newsgroups: comp.sys.sgi
Subject: BUG IN THE FORTRAN COMPLILER
Message-ID: <8902032353.aa26937@SMOKE.BRL.MIL>
Date: 4 Feb 89 07:26:50 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 47
X-Unparsable-Date: 2 FEB 89 19:17-MEZ

Date:  2-FEB-1989 19:17:09.51
From: BTP408 AT DBTHRZ5
To:   UBT0::BITNET::"INFO-IRIS@BRL.MIL"
Subj: BUG IN THE FORTRAN COMPLILER
DEAR SGI-USERS,
YOU CAN SAVE SOME DAYS OF WORK BY READING THIS MESSAGE.

C THERE IS AN ERROR IN THE FORTRAN COMPILER
C IF THE "ENTRY" STATEMENT IS USED. SUM3 AND
C SUM ARE NOT EQUAL?
C ==========================================
      READ(*,*) A,B
      SUM  = GI8(A)+GI8(B)
      A1   = GI8(A)
      A2   = GI8(B)
      SUM3 = A1 + A2
      WRITE(*,*) ' SUM =            ', SUM
      WRITE(*,*) ' BUT IT SHOULD BE ', SUM3
      WRITE(*,*) ' AND NOT          ', A2
      END
C ==========================================
      FUNCTION GI1(A)
        GI1 = 0.
      RETURN
C
      ENTRY GI8(A)
        GI8 = 1.*SIN(A)
C       IF GI8 = SIN(A) IT WORKS, OR ALTERNATIVELY IF
C       GI8 IS DECLARED AS A FUNCTION.
      RETURN
      END

WE HAVE THIS PROGRAM RUNNING WITH WRONG RESULTS ON A 4D70 AND A 4D20 (IRIX 3.1).
THERE ARE NO PROBLEMS ON A VAX. IF SOMEONE OF THE SGI-PEOPLE IS READING
THIS, PLEASE DELETE THIS BUG IN THE NEXT VERSION OF THE FORTRAN COMPILER AND
GIVE ME A SHORT REPLY.

EBERHARD


E. BODENSCHATZ
THEORETISCHE PHYSIK II
UNI BAYREUTH
8580 BAYREUTH
WEST GERMANY

BTP408 AT DBTHRZ5.BITNET