On Sunday 11 May 2008, Adam Treat wrote: > On Sunday 11 May 2008, Darin Adler wrote: > > Or you could do as Safari does and call iconForPageURL twice; the > > first time it's "display the icon if we already have it" but the > > second time it's "I heard you have a new icon, lets display the new > > icon we now know about". > > > > It would probably be cleaner if triggering the load of the icon was > > completely separate from actually fetching the icon data. > > Right. I would suggest that iconForPageURL should always return 0 unless > and until readIconForPageURLFromDisk is called which actually triggers the > load. > > It should be noted that *nothing* currently uses > readIconForPageURLFromDisk() ;) > > I guess I just couldn't believe that Safari calls iconForPageURL twice. It > is very unclear API.
Simon, I don't think QtWebKit API deals with this asynchronous nature of the IconDatabase at all. I'm not sure how to deal with it as it would seem that QWebSettings::iconForUrl() needs to be deprecated and possibly all the icon database API in QWebSettings. Looks like we need at a minimum a public API signal that the icon has been loaded (either from disk or network) and that will require a new class as QWebSettings is not a QObject. QWebIconDatabase ?? :( Cheers, Adam _______________________________________________ webkit-dev mailing list [email protected] http://lists.webkit.org/mailman/listinfo/webkit-dev
