From owner-freebsd-hackers  Tue Oct 17 22:10:43 1995
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id WAA27172
          for hackers-outgoing; Tue, 17 Oct 1995 22:10:43 -0700
Received: from rover.village.org (rover.village.org [198.137.146.49])
          by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id WAA27164
          for <freebsd-hackers@FreeBSD.ORG>; Tue, 17 Oct 1995 22:10:39 -0700
Received: from localhost (localhost [127.0.0.1]) by rover.village.org (8.6.11/8.6.6) with SMTP id XAA16946; Tue, 17 Oct 1995 23:10:05 -0600
Message-Id: <199510180510.XAA16946@rover.village.org>
To: jdp@polstra.com (John Polstra)
Subject: Re: getdtablesize() broken? 
Cc: terry@lambert.org, freebsd-hackers@FreeBSD.ORG
In-reply-to: Your message of Tue, 17 Oct 1995 15:23:00 PDT
Date: Tue, 17 Oct 1995 23:10:03 -0600
From: Warner Losh <imp@village.org>
Sender: owner-hackers@FreeBSD.ORG
Precedence: bulk

: > In addition, the select(2) call in BSD reserves the right to modify the
: > timeval structure to indicate the remaining time to allow the use of
: > the timeout as an even outcall mechanism for logical multithreading.
: 
: Yes.  I wish more implementations of select actually did that.

As someone who ported an application that depended on select(2) not
doing that, I can tell you that only Linux will really change the
value of timeval in a select call.  If that is not correct, I'd like
to know who else does change it (rather than merely reserve the right
to change it).

Warner