Implement the geolocation bubble for GTK...

Implement the geolocation bubble for GTK
This is a port of http://src.chromium.org/viewvc/chrome?view=rev&revision=42749

BUG=11246
TEST=run with --enable-geolocation and open http://bulach-pc.lon/chromium/wk/LayoutTests/fast/dom/Window/mbi.html then click the location bar icon



Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=42790

[joth, 2010-03-26 16:40:51]

Who's the GTK whizz?

[Nico, 2010-03-26 16:42:09]

erg did the content bubbles on gtk, estade is an all-purpose gtk whiz.

[Nico, 2010-03-26 16:42:56]

On 2010/03/26 16:42:09, Nico wrote:
> erg did the content bubbles on gtk, estade is an all-purpose gtk whiz.

(erg/estade: joth is in europe, hence it'd be great if you could review this
during the morning)

[bulach, 2010-03-26 16:53:06]

LGTM

...from a gtk newbie, let's wait for some whiz to take a quick look.

On 2010/03/26 16:42:56, Nico wrote:
> On 2010/03/26 16:42:09, Nico wrote:
> > erg did the content bubbles on gtk, estade is an all-purpose gtk whiz.
> 
> (erg/estade: joth is in europe, hence it'd be great if you could review this
> during the morning)

[Elliot Glaysher, 2010-03-26 17:08:53]

The actual GTK code looks fine, and works when patched into my local client when
you click Watch on the top level. So, LGTM.

I don't know if this is a known error, but when you click watch on the "watch
on" in one of the frames, it crashes with the following stack:

[4303:4303:2154649256255:FATAL:chrome/browser/geolocation/location_arbitrator.cc(80)]
Check failed: CalledOnValidThread(). 
Backtrace:
        StackTrace::StackTrace() [0xde3bca]
        logging::LogMessage::~LogMessage() [0xdf7d33]
        GeolocationArbitratorImpl::~GeolocationArbitratorImpl() [0xbd4e04]
        base::RefCounted<>::Release() [0xbd09a6]
        scoped_refptr<>::~scoped_refptr() [0xbd0339]
        GeolocationPermissionContext::~GeolocationPermissionContext() [0xbd26c1]
        base::RefCountedThreadSafe<>::DeleteInternal() [0xbd16fa]
        base::DefaultRefCountedThreadSafeTraits<>::Destruct() [0xbd112f]
        base::RefCountedThreadSafe<>::Release() [0xbd08ac]
        scoped_refptr<>::~scoped_refptr() [0xbd0289]
        (anonymous
namespace)::GeolocationConfirmInfoBarDelegate::~GeolocationConfirmInfoBarDelegate()
[0xbd2173]
        (anonymous
namespace)::GeolocationConfirmInfoBarDelegate::InfoBarClosed() [0xbd21c0]
        InfoBar::Close() [0x5a2a30]
        InfoBar::Closed() [0x5a2ac4]
        SlideAnimatorGtk::AnimationEnded() [0x5c3fef]
        Animation::Stop() [0xe6843f]
        Animation::Step() [0xe6869b]
        Animation::Run() [0xe686c1]
        DispatchToMethod<>() [0xe68b50]
        base::BaseTimer<>::TimerTask::Run() [0xe689e4]
        MessageLoop::RunTask() [0xdfa24c]
        MessageLoop::DeferOrRunPendingTask() [0xdfa2d2]
        MessageLoop::DoDelayedWork() [0xdfaa45]
        base::MessagePumpForUI::RunWithDispatcher() [0xe4aac3]
        base::MessagePumpForUI::Run() [0xe4b1aa]
        MessageLoop::RunInternal() [0xdf9a86]
        MessageLoop::RunHandler() [0xdf9932]
        MessageLoopForUI::Run() [0xdfaf3a]
        (anonymous namespace)::RunUIMessageLoop() [0x466410]
        BrowserMain() [0x46904a]

[joth, 2010-03-26 17:16:10]

On 26 March 2010 17:08, <erg@chromium.org> wrote:

> The actual GTK code looks fine, and works when patched into my local client
> when
> you click Watch on the top level. So, LGTM.
>
> great, thanks


> I don't know if this is a known error, but when you click watch on the
> "watch
> on" in one of the frames, it crashes with the following stack:
>
> yep I saw that one too last night:
http://code.google.com/p/chromium/issues/detail?id=39405

if you answer the prompts in order as they arrive, you can tip-toe around
it.

Thanks for patching it in & trying too.



>
>
[4303:4303:2154649256255:FATAL:chrome/browser/geolocation/location_arbitrator.cc(80)]
> Check failed: CalledOnValidThread().
> Backtrace:
>        StackTrace::StackTrace() [0xde3bca]
>        logging::LogMessage::~LogMessage() [0xdf7d33]
>        GeolocationArbitratorImpl::~GeolocationArbitratorImpl() [0xbd4e04]
>        base::RefCounted<>::Release() [0xbd09a6]
>        scoped_refptr<>::~scoped_refptr() [0xbd0339]
>        GeolocationPermissionContext::~GeolocationPermissionContext()
> [0xbd26c1]
>        base::RefCountedThreadSafe<>::DeleteInternal() [0xbd16fa]
>        base::DefaultRefCountedThreadSafeTraits<>::Destruct() [0xbd112f]
>        base::RefCountedThreadSafe<>::Release() [0xbd08ac]
>        scoped_refptr<>::~scoped_refptr() [0xbd0289]
>        (anonymous
>
>
namespace)::GeolocationConfirmInfoBarDelegate::~GeolocationConfirmInfoBarDelegate()
> [0xbd2173]
>        (anonymous
> namespace)::GeolocationConfirmInfoBarDelegate::InfoBarClosed() [0xbd21c0]
>        InfoBar::Close() [0x5a2a30]
>        InfoBar::Closed() [0x5a2ac4]
>        SlideAnimatorGtk::AnimationEnded() [0x5c3fef]
>        Animation::Stop() [0xe6843f]
>        Animation::Step() [0xe6869b]
>        Animation::Run() [0xe686c1]
>        DispatchToMethod<>() [0xe68b50]
>        base::BaseTimer<>::TimerTask::Run() [0xe689e4]
>        MessageLoop::RunTask() [0xdfa24c]
>        MessageLoop::DeferOrRunPendingTask() [0xdfa2d2]
>        MessageLoop::DoDelayedWork() [0xdfaa45]
>        base::MessagePumpForUI::RunWithDispatcher() [0xe4aac3]
>        base::MessagePumpForUI::Run() [0xe4b1aa]
>        MessageLoop::RunInternal() [0xdf9a86]
>        MessageLoop::RunHandler() [0xdf9932]
>        MessageLoopForUI::Run() [0xdfaf3a]
>        (anonymous namespace)::RunUIMessageLoop() [0x466410]
>        BrowserMain() [0x46904a]
>
>
> http://codereview.chromium.org/1360007
>

[Evan Stade, 2013-09-09 20:43:18]

https://codereview.chromium.org/1360007/diff/4001/chrome/browser/gtk/content_...
File chrome/browser/gtk/content_blocked_bubble_gtk.cc (right):

https://codereview.chromium.org/1360007/diff/4001/chrome/browser/gtk/content_...
chrome/browser/gtk/content_blocked_bubble_gtk.cc:205:
gtk_widget_grab_focus(button);
why are these two calls both necessary?

[joth, 2013-09-09 20:51:10]

On 2013/09/09 20:43:18, Evan Stade wrote:
>
https://codereview.chromium.org/1360007/diff/4001/chrome/browser/gtk/content_...
> File chrome/browser/gtk/content_blocked_bubble_gtk.cc (right):
> 
>
https://codereview.chromium.org/1360007/diff/4001/chrome/browser/gtk/content_...
> chrome/browser/gtk/content_blocked_bubble_gtk.cc:205:
> gtk_widget_grab_focus(button);
> why are these two calls both necessary?

Holy timewarp code review! :)
I'll defer to erg as this patch represents the peak (and sum) of my GTK
development experience.

[Evan Stade, 2013-09-09 21:05:56]

On 2013/09/09 20:51:10, joth wrote:
> On 2013/09/09 20:43:18, Evan Stade wrote:
> >
>
https://codereview.chromium.org/1360007/diff/4001/chrome/browser/gtk/content_...
> > File chrome/browser/gtk/content_blocked_bubble_gtk.cc (right):
> > 
> >
>
https://codereview.chromium.org/1360007/diff/4001/chrome/browser/gtk/content_...
> > chrome/browser/gtk/content_blocked_bubble_gtk.cc:205:
> > gtk_widget_grab_focus(button);
> > why are these two calls both necessary?
> 
> Holy timewarp code review! :)
> I'll defer to erg as this patch represents the peak (and sum) of my GTK
> development experience.

it came up in https://codereview.chromium.org/23463013/

I'll assume the former is not necessary. Thanks!