Path: utzoo!yunexus!telly!ddsw1!mcdchg!rutgers!tut.cis.ohio-state.edu!BAYES.ARC.NASA.GOV!self
From: self@BAYES.ARC.NASA.GOV (Matthew Self)
Newsgroups: gnu.gcc.bug
Subject: possible bug in optabs.c
Message-ID: <8811232225.AA12447@bayes.arc.nasa.gov>
Date: 23 Nov 88 22:25:06 GMT
Article-I.D.: bayes.8811232225.AA12447
Sender: daemon@tut.cis.ohio-state.edu
Distribution: gnu
Organization: GNUs Not Usenet
Lines: 25


  I spotted a fishy looking line in optabs.c in gcc 1.30.  Unless
ftruncdf2 is a special case, this is probably a typo.  It's probably
never mattered since most all FPU's can do ftruncdf with machine
instructions and never use the library call.

*** optabs.c-orig	Wed Nov 23 14:05:21 1988
--- optabs.c	Wed Nov 23 14:05:31 1988
***************
*** 1552,1558 ****
      ftrunc_optab->handlers[(int) DFmode].insn_code = CODE_FOR_ftruncdf2;
  #endif
    ftrunc_optab->handlers[(int) SFmode].lib_call = "_ftruncsf2";
!   ftrunc_optab->handlers[(int) DFmode].lib_call = "_ftruncsf2";
  
  #ifdef HAVE_andqi3
    if (HAVE_andqi3)
--- 1552,1558 ----
      ftrunc_optab->handlers[(int) DFmode].insn_code = CODE_FOR_ftruncdf2;
  #endif
    ftrunc_optab->handlers[(int) SFmode].lib_call = "_ftruncsf2";
!   ftrunc_optab->handlers[(int) DFmode].lib_call = "_ftruncdf2";
  
  #ifdef HAVE_andqi3
    if (HAVE_andqi3)