Date: Thu, 4 Sep 1997 17:32:35 -0700 (PDT) Message-Id: <199709050032.RAA05031@adit.ap.net> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: rtr AT rd DOT bbc DOT co DOT uk (Richard Russell), djgpp AT delorie DOT com From: Nate Eldredge Subject: Re: Curious floating point exception errors Precedence: bulk It's difficult to know without the source. Please post it, stripped to the minimum amount that reproduces the problem, and how to reproduce it. Thanks. At 02:08 9/4/1997 GMT, Richard Russell wrote: > >I have a program which compiles with gcc and runs perfectly on a Sun SPARC. >When compiled with DJGPP on a PC it fails with an FPESIG exception. >Something very odd is happening because inserting a diagnostic printf will >sometimes, by itself, cure the problem. Even more'ing the output can make >a difference, which suggests that perhaps the absolute address at which the >code is loading is somehow affecting things. > >I might be inclined to suspect the PC's hardware except that the problem >is repeatable and occurs on more than one machine. I can find no bugs >in the C code, such as uninitialised arrays. > >gcc -v gives: >Reading specs from c:/dgjpp/lib\specs >gcc version 2.7.2.1 > >Any ideas ? > >Richard. > > > > Nate Eldredge eldredge AT ap DOT net