Date:	Sun, 10 Jun 2001 15:32:24 +0300 (IDT)
From:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender:	eliz AT is
To:	Tim Van Holder <tim DOT van DOT holder AT pandora DOT be>
cc:	djgpp-workers AT delorie DOT com
Subject:	Re: Compiler options for djdev build
In-Reply-To:	<CAEGKOHJKAAFPKOCLHDIKEALCEAA.tim.van.holder@pandora.be>
Message-ID:	<Pine.SUN.3.91.1010610152710.7564D-100000@is>
MIME-Version:	1.0
Reply-To:	djgpp-workers AT delorie DOT com
Errors-To:	nobody AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

On Sun, 10 Jun 2001, Tim Van Holder wrote:

> Well, in order to "fix" this, you change the type used - i.e. you make
> todigit() and friends take an int, not a char.  But you're not passing an
> integer value, you're actually passing a character value - as such,
> declaring the argument as int seems wrong.

It's not wrong, it's very natural in C, if you think about a function 
invocation as a kind of expression (which it is).  All integral types 
shorter than int are always promoted to int's in an expression.

> The warning is only useful when converting legacy C code

The DJGPP library has quite a lot of such legacy code, most of it going 
back to the BSD library, originally used by DJ for v1.x.  The timezone 
code (djtzs*.zip) is also ``legacy code'', as well as edebug and fsdb.

> As you say, converting a function to varargs may trigger breakage, but how
> likely is that to happen?

I don't know.  But it seems a good idea to pay such a small price for 
added safety.

- Raw text -