Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!gem.mps.ohio-state.edu!ginosko!cs.utexas.edu!uunet!mcsun!hp4nl!star.cs.vu.nl!maart
From: maart@cs.vu.nl (Maarten Litmaath)
Newsgroups: comp.unix.wizards
Subject: O_NONBLOCK (was: Re: Strange SUN behaviour.)
Message-ID: <3790@solo6.cs.vu.nl>
Date: 24 Oct 89 01:01:03 GMT
References: <218@oasis.mrcu> <219@oasis.mrcu>
Organization: V.U. Informatica, Amsterdam, the Netherlands
Lines: 20

yh87@mrcu (Paul Johnson) writes:
\...
\My program was using fcntl(F_SETFL,...) to set the FNDELAY flag on
\descriptor 0 (non-blocking mode).  [...]  I had assumed
\that something like this would be local to the process which set the
\flag.  [...]

I consider this a misfeature too.  The worst thing: POSIX 1003.1 seems to
bless the sharing of the flag!  (The flag is called O_NONBLOCK and its
semantics differ slightly from FNDELAY.)
This behavior requires the shell to unset the flag after each command
(which is what csh does under SunOS [34].x[.y]).
Bletch!
Whenever a file descriptor is dup(2)ed (possibly implicitly through fork(2)),
it should get a *copy* of its parent's O_NONBLOCK (and O_APPEND) flag
(possibly copy-on-write).
Should I put on my asbestos suit, Doug?
-- 
A symbolic link is a POINTER to a file, | Maarten Litmaath @ VU Amsterdam:
 a hard link is the file system's GOTO. | maart@cs.vu.nl, mcsun!botter!maart