[XESAM] Search API Update Proposals

[Mikkel Kamstrup Erlandsen]

2007/8/8, Thiago Macieira <thiago at kde.org>:
>
> Mikkel Kamstrup Erlandsen wrote:
> >> > In any case, we're discussing null types in D-Bus itself. This will
> >> > probably require a new major message version...
> >>
> >> So using the 'correct' 'bv' is faster. That is interesting. Especially
> >> since quite a few NULL fields may appear.
> >> So i am still in favor of 'aa(bv)'
> >
> >Is this really true? We need to parse the Variant no matter what...
> >
> >Hmmm, in any case this is binding dependant.
>
> If you have a flag saying "nothing to see here, move along", why are you
> looking into the variant anyways?
>
> Anyways, that's a feature of QtDBus: it skips parsing into variants until
> someone tries to read it. And since no one will, it'll never be parsed.
> It'll just be skipped over.


Ok, we do not have to parse it if b=false, but as it is both possible to
introspect supported fields and actively set the fields to return per hit I
assume it will be the uncommon case that fields are not set.

I think we should optimize towards the case where the fields are indeed set.
If many fields are unset your search results are likely to be bad anyways,
and it does not really matter if your search is a split second faster...

Cheers,
Mikkel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.freedesktop.org/archives/xdg/attachments/20070808/9d5696e3/attachment.html