Path: utzoo!attcan!uunet!husc6!ncar!hao.ucar.edu!cmiller From: cmiller@hao.ucar.edu (Charlie Miller) Newsgroups: comp.sources.bugs Subject: GNU gcc compiler bug? Keywords: GNU gcc Message-ID: <1051@ncar.ucar.edu> Date: 28 Nov 88 21:29:06 GMT Sender: news@ncar.ucar.edu Reply-To: cmiller@hao.ucar.edu (Charlie Miller) Organization: High Altitude Observatory/NCAR, Boulder CO Lines: 56 Has anyone else noticed the following bug in the GNU gcc compiler? Of course, maybe there has been an installation error or some other problem, anyway, when compiling a simple program like: (compile command line: gcc -S test.c) #includemain() { double i; i = cos(.5); } the .s file looks like this: #NO_APP .globl fltused .text .even LC0: .double 0r0.5 .even .globl _main _main: link a6,#-8 movel LC0+4,sp@- movel LC0,sp@- jbsr _cos fmoved fp0,a6@(-8) L1: unlk a6 rts As you can see, the compiler sets up for a software library call, makes the call, then tries to get the result out of the M68881 floating point co- processor. I have just installed the gcc compiler, as a base for installing g++. The compiler version is 1.27. I am using an Integrated Solutions M68020/68881 based computer running UNIX bsd 4.3. The gcc compiler seems to work fine on other floating point operations like divide and multiply, making use of the m68881. If anyone has seen this problem before or has any information please email me. Thanks. =========================================================================== Some folks look for answers... -Charles E. Miller USPS Mail: High Altitude Observatory P.O. Box 3000 Boulder CO 80307 Internet: cmiller@hao.ucar.edu UUCP: {arizona,decvax,ihnp4}!hao!cmiller ===========================================================================