From: skb AT xmission DOT com (Scott Brown)
Newsgroups: comp.os.msdos.djgpp
Subject: LFNs, timeslices and the preprocessor
Date: Sun, 22 Oct 2000 02:21:23 GMT
Organization: (none)
Lines: 21
Message-ID: <39f23421.749522025@198.60.22.3>
NNTP-Posting-Host: slc259.modem.xmission.com
X-Trace: news.xmission.com 972181710 16556 166.70.2.5 (22 Oct 2000 02:28:30 GMT)
X-Complaints-To: abuse AT xmission DOT com
NNTP-Posting-Date: 22 Oct 2000 02:28:30 GMT
X-No-Archive: Yes
X-Newsreader: Forte Free Agent 1.21/32.243
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

A few little questions for y'all....

#1 - Is __dpmi_yield what I want to make my program stop being such a
pig when running in a Win9x DOS box?  The current incarnation (built
with BC++ 3.0) is *supposed* to run without being too noticeable, but
in reality it makes Windows pretty much unusable until it finishes.
I'm porting it to DJGPP for other reasons, but making it better
behaved would be a nice bonus.

#2 - Is there a reliable way to detect whether long filenames are
supported?  My program relies on LFNs to function correctly, and could
do a lot of damage if it thinks it has LFN support when it doesn't.

#3 - Slightly off-topic portability question.  Reading the CLC FAQ
suggests that a construct like "#if sizeof(int)" doesn't work (it does
work in BC, apparently just to annoy people trying to port stuff from
BC to GCC).  What else would be a good way to test the size of a type
at compile time (e.g. to make sure that long is exactly 32 bits)?  I
could check it with sizeof at runtime, but that would be nasty.

Thanks.