NULL Variants?

[John (J5) Palmieri]

On Wed, 2005-08-31 at 16:55 +0000, Nate Nielsen wrote:
> Nate Nielsen wrote:
> > Havoc Pennington wrote:
> > 
> >>>Nate Nielsen wrote:
> >>>
> >>>>When sending a DBUS variant across the wire, is there a way to specify a
> >>>>NULL/None/Missing value? I'm using the GLib bindings. I've tried various
> >>>>things and gotten crashes and big bad warnings.
> >>
> >>We used to have one and deleted it on purpose, I don't remember the full
> >>reason why but it's probably in the list archives.
> > 
> > 
> > And so off I went searching various archives, and found this:
> > 
> > 
> >>The NIL type:
> >>
> >>  NIL doesn't make a hell of a lot of sense as a *type*, really it's a 
> >>  value that's allowed in *some* languages to replace a value of any
> >>  type. I think we need to get rid of DBUS_TYPE_NIL since I can't make 
> >>  any sense out of it.
> > 
> > 
> > I agree that DBUS_TYPE_NILL doesn't makes less sense as a type. It makes
> > much more sense when used inside a variant. This is similar to how
> > DBUS_TYPE_DICT_ENTRY doesn't make much sense outside of an array.
> > 
> > That said, while reading through that thread in the list archives I
> > didn't find a solution. Any ideas? I need to represent missing values in
> > a variant (similar to SQL NULLs).
> 
> I could be really daft, but I don't see the reasoning of why having
> Variant NULLs is evil, or inappropriate.
> 
> If this really is missing, and there's no replacement, then I'd like to
> implement it (even though I have no experience hacking DBUS).
> 
> Havoc, John, Colin, any reasons this change wouldn't be accepted? Sorry
> for all the CC's, just interested in a semi-definitive answer.
> 
> Cheers,
> Nate

I have no objections to you contributing a patch.  Personally though
NULL Variants just makes me think that everything is going to become a
variant. Why not allow types where it is logical, like strings to be
able to be NULL?

-- 
John (J5) Palmieri <johnp at redhat.com>