Geoclue and WebKit: unit tests

[Bastien Nocera]

On Thu, 2010-06-24 at 19:44 +0200, Javier Fernandez Garcia-Boente wrote:
> On Mon, 2010-06-21 at 13:32 +0200, Javier Fernandez Garcia-Boente wrote:
> > On Thu, 2010-06-17 at 11:25 +0200, Javier Fernandez Garcia-Boente wrote:
> 
> > 
> > Anyway, getting back to the WebKit Geolocation unit tests topic, i think
> > we should provide a way of forcing one provider as the selected one,
> > even though if several are available. When programming unit tests, not
> > every functionality have to be enabled; even more, some of them could be
> > dangerous or inefficient for the automatic testing framework. The usual
> > way of programming unit tests by setting a dummy GPS coordinates before
> > executing the test and retrieving such dummy coordinates as the result
> > of the operation, so the Manual provider seems to be the more
> > appropriated for this way of use the Geolocation API.
> > 
> > Do you think this is the right way to proceed ? Implementing a way of
> > selecting a fixed provider.
> > 
> 
> New version of the patch attached. Besides that, after further
> investigations in the WebKit part, it seems that the position-changed
> notifications were not being received correctly in the WebKit unit
> tests. I initially implemented the mockup calls this way:

Could you please make sure those patches get into bugzilla? I have some
time set on the side for Geoclue hacking in the coming weeks (including
finishing the GSettings port I started), and I'll be going through the
bugzilla bugs, probably rather than the mailing-list.

Cheers