Re: Qsort defects (in C-library)

2013-02-18 Thread Christopher Faylor
On Mon, Feb 18, 2013 at 10:05:11PM +0100, Corinna Vinschen wrote: >On Feb 18 08:51, Dennis de Champeaux wrote: >> >> >> // I hope this is the proper mailing list > >Unfortunately it's not. Qsort is not implemented in Cygwin itself, but >rather in newlib, the underlying C lib. The right mailing

Re: Qsort defects (in C-library)

2013-02-18 Thread Corinna Vinschen
On Feb 18 08:51, Dennis de Champeaux wrote: > > > // I hope this is the proper mailing list Unfortunately it's not. Qsort is not implemented in Cygwin itself, but rather in newlib, the underlying C lib. The right mailing list is newlib AT sourceware DOT org. It's also kind of good style not to

Qsort defects (in C-library)

2013-02-18 Thread Dennis de Champeaux
// I hope this is the proper mailing list Qsort defects Qsort was described in a 1993 paper by Bentley & McIlroy. Let me start with the good news. Someone replaced a 2nd recursive call at the bottom by tail recursion. Here the bad news. Someone modified the code with an "improvement", which h