Re: [C++ patch] accept __null as sentinel

[Mark Mitchell <mark at codesourcery dot com>]

Michael Matz wrote:
> Hi,
>
> On Fri, 29 Apr 2005, Mark Mitchell wrote:
>
>
> > > Maybe.  But note that the case of a literal 0 will still be warned (as
> > > also for 0L), no matter what.  It's only __null, a GCC specific thing,
> > > which I don't want to warn about, in a warning only implemented by
> > > GCC.  We _know_ using __null will work.
> >
> > But the user (probably) didn't write "__null" -- they wrote "NULL".  
> > So, what we really know is that using GCC's <stddef.h> will work.  But
> > if you're trying to write code portable to non-GCC compilers, then the
> > warning still has value for you.
>
>
> That's why a suggested to put it under -pedantic. 

The documentation for -pedantic says:

"Issue all the warnings demanded by strict ISO C and ISO C++; reject all 
programs that use forbidden extensions, and some other programs that do 
not follow ISO C and ISO C++."

Failing to pass a proper NULL pointer to a function with 
__attribute__((sentinel) is not a failure to conform to ISO C++. 
Whereas -Wall is documented as:

"This enables all the warnings about constructions that some users 
consider questionable, and that are easy to avoid (or modify to prevent 
the warning), even in conjunction with macros."

So, -Wall would be a better place, or -Wformat=2, as Joseph suggests, if 
we think the warning is too annoying by default.

--
Mark Mitchell
CodeSourcery, LLC
mark@codesourcery.com
(916) 791-8304