Re: [PUSHED] fdo#46808, convert sdb::QueryDesigner to new style

On Wed, Apr 03, 2013 at 11:38:40AM +0200, Stephan Bergmann wrote:
> On 04/02/2013 12:54 PM, Fridrich Strba (via Code Review) wrote:

> > Thank you for your patch!  It has been merged to LibreOffice.

> > If you are interested in details, please visit

> >      https://gerrit.libreoffice.org/3031

> Lionel, can you have a look whether the new-style service
> com.sun.star.sdb.QueryDefinition and accompanying interface
> com.sun.star.sdb.XQueryDefinition make sense?

This patch seems to introduce a new published interface, which did not
exist (not even unpublished) before. I don't understand why.

As to the Java snippet, I'm not sure

> > Seeing that the only use of that service
> > (dbaccess/source/core/api/querycontainer.cxx uses XPropertySet
> > generically (and does not explicitly access any of the listed
> > properties), but also expects XContent and assumes(?) XComponent,
> > it might make more sense to drop the attributes from
> > XQueryDefinition and instead let it inherit from those additional
> > interfaces.

I don't understand the consequences of "drop the attributes from
XQueryDefinition and instead let it inherit from those additional
interfaces".

I notice that the created service is added to some internal structure:
    // first clone this object's CommandDefinition part
    Reference< XPropertySet > xCommandDefinitionPart(
    m_aContext.createComponent(
    (::rtl::OUString)SERVICE_SDB_QUERYDEFINITION ), UNO_QUERY_THROW );

   (...)

   m_xCommandDefinitions->insertByName(sNewObjectName, makeAny(xCommandDefinitionPart));

*and* other stuff accesses the listed properties, and I assume that's
somehow the created service see e.g.
 git grep PROPERTY_UPDATE_TABLENAME
But I don't understand what consequences this has for the service
definition.

> >  I'm also not sure what to make of

> > > Object oQuery = xSSFQueryDefs.createInstance(); //"com.sun.star.sdb.QueryDefinition"

> > in wizards/com/sun/star/wizards/db/DBMetaData.java.  You could push this patch for review on 
> > gerrit and ping Lionel.

I don't quite understand that Java code either. To my eyes, it seems
to access the "Command" property explicitly?

-- 
Lionel