Clean up geolocation_dispatcher_host.h

Clean up geolocation_dispatcher_host.h

It was pulling internal geolocation headers into resource_message_filter.cc

BUG=none
TEST=none

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=49271

[joth, 2010-06-08 12:40:40]

Nice easy review (I hope!) :)

[andreip3000, 2010-06-08 14:59:52]

Looks good, one question below:

http://codereview.chromium.org/2742002/diff/3001/4001
File chrome/browser/geolocation/geolocation_dispatcher_host.cc (right):

http://codereview.chromium.org/2742002/diff/3001/4001#newcode135
chrome/browser/geolocation/geolocation_dispatcher_host.cc:135:
geolocation_renderer_ids_.insert(render_view_id);
Maybe it is obvious but I am still learning the codebase: before you were
calling RegisterDispatcher, which was checking that the call was made on the
right thread and, if not, it was   dispatching it to the right thread. Now, you
just assert we're always on the correct thread. Why is this?

[joth, 2010-06-08 15:12:26]

Thanks

no problem about the Q...

On 8 June 2010 15:59, <andreip@chromium.org> wrote:

> Looks good, one question below:
>
>
> http://codereview.chromium.org/2742002/diff/3001/4001
> File chrome/browser/geolocation/geolocation_dispatcher_host.cc (right):
>
> http://codereview.chromium.org/2742002/diff/3001/4001#newcode135
> chrome/browser/geolocation/geolocation_dispatcher_host.cc:135:
> geolocation_renderer_ids_.insert(render_view_id);
> Maybe it is obvious but I am still learning the codebase: before you
> were calling RegisterDispatcher, which was checking that the call was
> made on the right thread and, if not, it was   dispatching it to the
> right thread. Now, you just assert we're always on the correct thread.
> Why is this?
>
>
RegisterDispatcher was only ever called from OnRegisterDispatcher which in
turn is only called from OnMessageReceived (through the message map) which
we know - and assert - is only called on the IO thread. as it is hooked into
resource message filter which by design is only running on IO thread

At some point in the past this was either not the case, or we weren't sure
it would be, but in the current (stable) incarnation is certainly appears to
be as I describe.






>
> http://codereview.chromium.org/2742002/show
>

[andreip3000, 2010-06-08 15:15:01]

LGTM

I see, thanks for the explanation.


On 2010/06/08 15:12:26, joth wrote:
> Thanks
> 
> no problem about the Q...
> 
> On 8 June 2010 15:59, <mailto:andreip@chromium.org> wrote:
> 
> > Looks good, one question below:
> >
> >
> > http://codereview.chromium.org/2742002/diff/3001/4001
> > File chrome/browser/geolocation/geolocation_dispatcher_host.cc (right):
> >
> > http://codereview.chromium.org/2742002/diff/3001/4001#newcode135
> > chrome/browser/geolocation/geolocation_dispatcher_host.cc:135:
> > geolocation_renderer_ids_.insert(render_view_id);
> > Maybe it is obvious but I am still learning the codebase: before you
> > were calling RegisterDispatcher, which was checking that the call was
> > made on the right thread and, if not, it was   dispatching it to the
> > right thread. Now, you just assert we're always on the correct thread.
> > Why is this?
> >
> >
> RegisterDispatcher was only ever called from OnRegisterDispatcher which in
> turn is only called from OnMessageReceived (through the message map) which
> we know - and assert - is only called on the IO thread. as it is hooked into
> resource message filter which by design is only running on IO thread
> 
> At some point in the past this was either not the case, or we weren't sure
> it would be, but in the current (stable) incarnation is certainly appears to
> be as I describe.
> 
> 
> 
> 
> 
> 
> >
> > http://codereview.chromium.org/2742002/show
> >
>