From:	Sinan_Unur AT mail DOT com (A. Sinan Unur)
Newsgroups:	comp.os.msdos.djgpp
Subject:	qsort vs QuickSort (was Re: Peculiar behavior of program.)
Date:	29 Jun 2001 08:05:39 -0700
Organization:	http://groups.google.com/
Lines:	26
Message-ID:	<426926da.0106290704.58345944@posting.google.com>
References:	<Pine DOT SUN DOT 3 DOT 91 DOT 1010625154252 DOT 2920H-100000 AT is> <3b37e2d6 DOT 287121289 AT news DOT primus DOT ca> <9h9ich$jhe$1 AT nets3 DOT rz DOT RWTH-Aachen DOT DE> <3b3b43a3 DOT 210697691 AT news DOT primus DOT ca> <9hfqbt$g8b$1 AT nets3 DOT rz DOT RWTH-Aachen DOT DE> <3b3be472 DOT 251871236 AT news DOT primus DOT ca>
NNTP-Posting-Host:	128.253.251.163
X-Trace:	posting.google.com 993827141 24215 127.0.0.1 (29 Jun 2001 15:05:41 GMT)
X-Complaints-To:	groups-support AT google DOT com
NNTP-Posting-Date:	29 Jun 2001 15:05:41 GMT
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp
Reply-To:	djgpp AT delorie DOT com

invalid AT erehwon DOT invalid (Graaagh the Mighty) wrote in message news:<3b3be472 DOT 251871236 AT news DOT primus DOT ca>...

> >Just because it's called qsort() by no means implies that the
> >implementation is QuickSort. In the case of DJGPP, it isn't.
> 
> Oops, I forgot: just because it's called libc by no means implies that
> it's ANSI C. :P

ANSI C imposes no such requirement. The algorithm does not matter.
What matters is that the implementation behaves in the following way:

(quoting from http://www.dinkumware.com/htm_cl/stdlib.html#qsort)

"The function sorts, in place, an array consisting of nelem elements,
each of size bytes, beginning with the element whose address is base.
It calls the comparison function whose address is cmp to compare pairs
of elements. The comparison function must return a negative value if
e1 is less than e2, zero if the two are equal, or a positive value if
e1 is greater than e2. Two array elements that are equal can appear in
the sorted array in either order."

Sinan.

--
A. Sinan Unur
http://www.unur.com/

- Raw text -