Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!cornell!uw-beaver!sumax!polari!6sigma!blm
From: blm@6sigma.UUCP (Brian Matthews)
Newsgroups: comp.unix.wizards
Subject: Re: ENOTTY error when executing cp(1)
Message-ID: <302@6sigma.UUCP>
Date: 15 Sep 89 18:55:10 GMT
References: <507@fdmetd.uucp> <2556@taux01.UUCP> <301@6sigma.UUCP> <109@harald.UUCP>
Reply-To: blm@6sigma.UUCP (Brian Matthews)
Organization: Six Sigma CASE, Inc.
Lines: 31

In article <109@harald.UUCP> jba@harald.ruc.dk (Jan B. Andersen) writes:
|As I see it, the "correct" way to check for errors is by coding something
|like this:
|  if( some_system_call() ) {
|	printf( "errno %d has occured in some_system_call()\n", errno );
|	exit( errno );
|  }

Two problems with this:

1.  You should use perror so the user gets a reason for the failure,
    not just a number.

2.  errno isn't preserved across many stdio calls, so the exit will
    not exit with the errno generated by some_system_call.

|If you don't like if'ing the call do something like this:
|  errno = 0;
|  some_system_call();
|  if( errno ) {
|	...
|  }

My question was basically, is that guaranteed to work.  I suspected not,
but couldn't find anything to prove it.  Doug Gwyn has since posted
that you can't do this - some_system_call may set errno to any arbitrary
value, yet still succeed.
-- 
Brian L. Matthews			blm@6sigma.UUCP
Six Sigma CASE, Inc.			+1 206 854 6578
PO Box 40316, Bellevue, WA  98004