Add Disable signal when bus_ibus_impl_set_context_engine_from_desc is called.

Add Disable signal when bus_ibus_impl_set_context_engine_from_desc is called.


BUG=#1261
TEST=Linux desktop

[fujiwara, 2011-08-14 02:41:44]

This is separated from #4853041.
Currently the engine hotkey uses bus_ibus_impl_set_context_engine_from_desc() in
bus_ibus_impl_filter_keyboard_shortcuts().
When we use engine hotkey to switch engines, the Disable signal is not called
and asked it with bug #1261.

Probably I think the calling bus_input_context_disable() in new_engine_cb() is
an idea. Or do we add bus_input_context_disable() in
bus_ibus_impl_filter_keyboard_shortcuts() ?

[Peng, 2011-08-14 04:33:36]

http://codereview.appspot.com/4875049/diff/1/bus/inputcontext.c
File bus/inputcontext.c (right):

http://codereview.appspot.com/4875049/diff/1/bus/inputcontext.c#newcode2298
bus/inputcontext.c:2298: bus_input_context_disable (data->context);
This Change is LGTM. But I have a question. In issue 1261, the problem is
Disable signal is not sent to engine. But in here, unset_engine is called before
bus_input_context_disable(). I think the engine should not get the disable
signal. right?

[fujiwara, 2011-08-24 07:32:09]

On 2011/08/14 04:33:36, Peng wrote:
> http://codereview.appspot.com/4875049/diff/1/bus/inputcontext.c
> File bus/inputcontext.c (right):
> 
> http://codereview.appspot.com/4875049/diff/1/bus/inputcontext.c#newcode2298
> bus/inputcontext.c:2298: bus_input_context_disable (data->context);
> This Change is LGTM. But I have a question. In issue 1261, the problem is
> Disable signal is not sent to engine. But in here, unset_engine is called
before
> bus_input_context_disable(). I think the engine should not get the disable
> signal. right?

Right. I didn't check the behavior.
Now I'm thinking we don't have to call bus_input_context_unset_engine and
revised the patch.

[Peng, 2011-08-24 13:13:42]

LGTM
On 2011/08/24 07:32:09, fujiwara wrote:
> On 2011/08/14 04:33:36, Peng wrote:
> > http://codereview.appspot.com/4875049/diff/1/bus/inputcontext.c
> > File bus/inputcontext.c (right):
> > 
> > http://codereview.appspot.com/4875049/diff/1/bus/inputcontext.c#newcode2298
> > bus/inputcontext.c:2298: bus_input_context_disable (data->context);
> > This Change is LGTM. But I have a question. In issue 1261, the problem is
> > Disable signal is not sent to engine. But in here, unset_engine is called
> before
> > bus_input_context_disable(). I think the engine should not get the disable
> > signal. right?
> 
> Right. I didn't check the behavior.
> Now I'm thinking we don't have to call bus_input_context_unset_engine and
> revised the patch.