Date: Thu, 16 Jul 92 10:16:34 GMT
From: kuku AT acds DOT physik DOT rwth-aachen DOT de (Christoph Kukulies)
To: djgpp AT sun DOT soe DOT clarkson DOT edu, treece AT hal2000 DOT sabbagh DOT com
Subject: Re:  sector not found

Presently I'm hunting a bug in a large f2c/gcc compiled program.
This bug or failing behaviour shows up in a similar manner you describe,
but the 'sector not found' message only occurs on one class of 486/33 IDE
machines while the other class (different motherboard) behaves different.
(I don't want to blame the hardware, but who knows? - we have read horror
stories about RAS precharge time etc.,recently)
Anyway, I'm still hunting after this bug and I believe now, that it is not
gcc (either 1.06 or 1.07) because I'm using objects (even libf2c.a) which
are cross compiled under SCO using gcc-1.40. I tend to believe that the
problem is in go32, perhaps has something to do with unnormalized FP numbers
(just a vague thought) and exceptions/paging interaction.

BTW, a question: How is the 387 initialized? Does it trap on unnormalized
FP numbers?

--Chris