Re: Ping: Ping: [Patch][RFA] Warn if pointer is initialized to 'false'

[Simon Baldwin <simonb at google dot com>]

On 1 February 2010 01:10, Dave Korn <dave.korn.cygwin@googlemail.com> wrote:
> On 01/02/2010 00:28, Mark Mitchell wrote:
>>
>> We're talking about whether to issue a pedwarn or a warning, not whether
>> to warn at all. ?We certainly want to warn.
>
> ?Beg pardon, I must have misread the thread.

I think I'm misreading this review in the same way as well, now.  The
first response mentions OK, but then questions the change to call.c,
which is after all the only meaningful part of the patch.

false is a valid 'null pointer constant'.  It's just a very odd one.
It's currently a warning if used as the argument to a function that
takes a pointer, but not if directly assigned to a pointer variable,
which seems strange.

The patch introduces a warning for the latter also, since it seems
likely that it's a bug-in-waiting.  However, I  know there are also
some intentional uses of false as a function argument for a function
taking a pointer (Boost, for example, relies on this).  Perhaps some
subtlety along these lines is why direct assignment is not already a
warning.  It's also produced only if -Wconversion, something not added
with -Wall -Wextra.

Here's a more full example of how the current gcc behaves:

  // gcc -Wconversion ...
  void *f(void *p) {return p;}
  void *a = true;      // error: cannot convert 'bool' to 'void*' ...
  void *b = f(false);  // warning: converting 'false' to pointer type for ...
  void *c = false;     // without the patch, an eerie silence; with, a warning

It does seem odd that at present false is warned about if used as an
argument to a function that takes a pointer, but not if directly
assigned to a pointer variable.

Anyway, I'm not wedded to this patch.  If the risks outweigh the
advantage I'm happy to drop it.

-- 
Google UK Limited | Registered Office: Belgrave House, 76 Buckingham
Palace Road, London SW1W 9TQ | Registered in England Number: 3977902