[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
compiling pcl on Coral Common Lisp
- To: Kevin_Crowston@XV.MIT.EDU
- Subject: compiling pcl on Coral Common Lisp
- From: alms@cambridge.apple.com (Andrew L. M. Shalit)
- Date: Thu, 17 Aug 89 17:43:48 EDT
- Cc: commonloops.pa@Xerox.COM, Kevin_Crowston@XV.MIT.EDU
- In-reply-to: Kevin_Crowston@XV.MIT.EDU's message of 17 Aug 89 14:52:40 EDT <0.43520.29067.17244.10413@XV.MIT.EDU>
- Redistributed: commonloops.pa
From: Kevin_Crowston@XV.MIT.EDU
Date: 17 Aug 89 14:52:40 EDT
Sender: Kevin_Crowston@XV.MIT.EDU
Quite some time ago, I mentioned I had problems compiling
PCL on a Mac with an accelerator board. Through the kindness
of TSI, the board manufacturer, I have discovered that
Allegro, when running on a 68020, assumes the existance of
an FPU chip. Putting in the FPU chip solves the problem.
this isn't quite right. Mac Allegro never assumed that 68020 implied an
FPU chip. Rather, it assumes that unimplemented instruction traps
get processed correctly. On a machine without an FPU, executing an FPU
instruction should cause an unimplemented instruction trap, which Lisp
than catches, redirecting things to SANE. Unfortunately, some accellerator
board manufactures play with the exception handling so that unimplemented
instruction traps don't occur properly.
In any case, putting in an FPU chip solves the problem.