Sender:	nate AT cartsys DOT com
Message-ID:	<3692BEEE.79E6FF53@cartsys.com>
Date:	Tue, 05 Jan 1999 17:39:58 -0800
From:	Nate Eldredge <nate AT cartsys DOT com>
X-Mailer:	Mozilla 4.08 [en] (X11; I; Linux 2.0.35 i486)
MIME-Version:	1.0
To:	djgpp AT delorie DOT com
Subject:	Re: Announce: Allegro 3.1
References:	<aCwlkLBnYnk2Ewy7 AT talula DOT demon DOT co DOT uk>
Reply-To:	djgpp AT delorie DOT com

Shawn Hargreaves wrote:

> <complaint mode>
> 
> Ok, so this is directed at the people who maintain gcc rather than
> anyone on this list, but it is something that really annoys me: why
> can't they leave the warning options alone between compiler versions? By
> all means add new warning switches that can be optionally enabled, but
> for anyone who maintains and distributes large source packages, it is
> very irritating when different compilers will complain about a different
> range of things (some of which are really anal and IMHO not actual
> errors at all). It is a particular nuisance when, as with some of the
> PGCC complaints, the option to disable the warning in one compiler
> version is itself not recognised by other versions...
> 
> </complaint>

With all due respect:

For the vast majority of users, it is best that -Wall and other global
warning switches reflect the best available warnings.  It makes sense to
me that anyone who wants to support more than one version should stay
away from -Werror in the distribution Makefile.  This admittedly won't
barf on warnings that could be due to the user messing up the build, but
will keep it working.  IMHO, warnings are for developers, not for users
compiling.

And how is a compiler to recognize an option to disable a warning which
won't exist until a future version?  (Unless with Emacs' M-x
time-forward ;-)  [see etc/future-bug])
-- 

Nate Eldredge
nate AT cartsys DOT com

- Raw text -