Invalid rtems_id?

[Robert S. Grimes]

Aaron J. Grier wrote:
> On Thu, Mar 27, 2008 at 04:24:27PM +0100, Thomas Dörfler wrote:
>   
>> The C standard guarantees, that an uninitialized global pointer
>> variable holds a NULL pointer, which is used as a "invalid"
>> placeholder.
>>     
>
> are you sure about that?  I was under the assumption that unix (and
> other) systems have historically zeroed .bss, but that doing so is not a
> requirement.
>
>   
The phrase "the C standard" is rather vague, but it is my understanding 
that zeroing .bss _is_ indeed specified by "the" C standard.  It is true 
that some compilers do not do this, at least by default; the "more 
honest" ones specify this in a "non-standard" section of their 
documentation.

HTH,
-Bob