X-Authentication-Warning: delorie.com: mail set sender to djgpp-bounces using -f
Date: Thu, 2 Sep 2004 18:30:56 -0400
Message-Id: <200409022230.i82MUuqE006471@envy.delorie.com>
From: DJ Delorie <dj AT delorie DOT com>
To: djgpp AT delorie DOT com
In-reply-to: <20040902221157.GB7970@webhome.cz> (message from Egon Eckert on
	Fri, 3 Sep 2004 00:11:57 +0200)
Subject: Re: scanf: strange behavior?
References: <20040902144152 DOT GH6858 AT webhome DOT cz> <2pp00iFnf7nsU1 AT uni-berlin DOT de> <41374c9c$0$173$cc7c7865 AT news DOT luth DOT se> <20040902184005 DOT GA23943 AT webhome DOT cz> <41378A42 DOT 5040801 AT acm DOT org> <200409022144 DOT i82LiDLe005751 AT envy DOT delorie DOT com> <20040902221157 DOT GB7970 AT webhome DOT cz>
Errors-To: nobody AT delorie DOT com
X-Mailing-List: djgpp AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk


> -- I'm becoming more and more confident, that it's actually
> inconsistent with the current implementation.
> 
> To return EOF, the input has to end *before* the first item
> converted/assigned, because when it has to end otherwise?  It has to
> end sometime...  No word about conversion items pending.

The documentation is vague (heck, the standard is a little vague too).
The problem is that it doesn't say what happens when they're
successful conversions AND input ends prematurely.  The normal case is
that you have a conversion for all your format specifiers before input
ends.  The other two cases are a conversion error (wrong input) or end
of string (no input).

The appropriate test cases are thus, for "%d %d":

    ""
    "x"
    "5"
    "5 x"
    "5 5"
    "5 5 x"

The first and third are the problem cases here.