home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!pipex!demon!cix.compulink.co.uk!pmoore
- Newsgroups: comp.sys.acorn.tech
- From: pmoore@cix.compulink.co.uk (Paul Moore)
- Subject: Re: Suppressing floating point errors in C
- Reply-To: pmoore@cix.compulink.co.uk
- Date: Sat, 9 Jan 1993 10:49:00 +0000
- Message-ID: <memo.852505@cix.compulink.co.uk>
- Sender: usenet@demon.co.uk
- Lines: 19
-
- mckay_g@kosmos.wcc.govt.nz ("Rob") writes (regarding my problem of being
- able to ignore SIGFPE in C):
-
- > The way I would overcome this problem is to change the FPU
- > status/control register to prevent what ever type of FP
- > errors you wish to ignore.
-
- The problem is that there is a BUG in the FP emulator, which means that
- funny things happen when you try to work with infinity/NaN (which is what
- generates the SIGFPE). As a consequence, setting the FP status/control
- register does not work properly (in the "proper" program where I need to do
- this - I have not actually tried it for this test case).
-
- I posted the FPE bug report on csa.tech, and I have sent it to Acorn, but I
- know of no work-around yet. (I won't re-post it here as it's long - Email me
- if you missed it and want to see a copy).
-
- Gustav.
-
-