[Nepomuk] Nepomuk Core - Questions & Patches

[Sebastian Trüg]

Hi Vishesh,

FINALLY SOMEONE LOOKS AT THIS STUFF BESIDES ME!!!!!!!!!!!!!!!!
THANK YOU! :)

On 05/07/2010 01:15 PM, Vishesh Handa wrote:
> 1. The *Nepomuk::ResourceData::hasProperty( const QUrl& uri )* function
> currently uses the MAINMODEL to answer. It think it should use the
> cache, as it would be faster. The function *hasProperty( const QUrl& p,
> const Variant& v ) *also uses the cache. A patch has been provided.

yes, please commit.

> 2. The *Nepomuk::ResourceData::removeProperty( const QUrl& uri )*
> function doesn't update the cache. The setProperty function does. It
> should be symmetrical. I think.

also correct.

Can you backport this one?

> 3. In *Nepomuk::ResourceFilterModel::updateProperty( const QUrl&
> resource, const QUrl& property, const Node& value )* shouldn't the line
> if( it.hasNext() ) be a while( it.hasNext() ) ? I haven't tested it out,
> but I think this is what should happen -
> 
> Let's say I have a resource with a uri and certain properties.
> 
> uri a rdf:Resource .
> uri a rxx:A .
> uri a rxx:B .
> 
> and I call the function *updateProperty*( uri, rdf:type, rxx:C ). The
> current implementation would get all the above 3 statements (I'm not
> sure in what order) . Suppose it got /"uri a rdf:Resource/" first. It
> would then proceed to remove the statement and add the /"uri a rxx:C"
> /statement. This doesn't seem right.
> 
> IMO, it should either delete all matching statements and add the one
> specified in the arguments or it should take an oldValue and newValue
> and accordingly change. KDevelop says it isn't used anywhere, so this
> isn't really that important. :-/

You are correct. But as you said the method is not used anyway, so we
better remove it. right?

Cheers,
Sebastian