Bug 149111 – [GTK] timeline is not updated after few seconds when mouse hovers on controls

• Component: WebKitGTK
• Status: RESOLVED FIXED

[ChangSeok Oh]

1. go http://www.w3schools.com/html/html5_video.asp
2. click a play button
3. stay cursor on the playbutton.
4. you can see time is paused after 4 seconds.

[ChangSeok Oh]

Created attachment 261112 [details]
Patch

[Xabier Rodríguez Calvar]

Can you please tell me which diagnostics have you made so that you reach the conclusion that this is the best patch?

It is weird because we shouldn't need to check the opacity as precisely, we handle the opacity thru CSS with the classes we add to the element.

[ChangSeok Oh]

(In reply to comment #2)
> Can you please tell me which diagnostics have you made so that you reach the
> conclusion that this is the best patch?
> 
> It is weird because we shouldn't need to check the opacity as precisely, we
> handle the opacity thru CSS with the classes we add to the element.

As I described the situation in the ChangeLog, We changed not to update time if controllAreHidden is true in https://bugs.webkit.org/show_bug.cgi?id=137773.
However this affects on controls behavior very much. Let guess this scenario.
(move cursor on playbutton) -> (click play button) -> (hideControl timer fired by resetHideControlsTimer after 4 secs) -> (now controlsAreHidden is true since panel has neither -webkit-media-controls-panel.show nor -webkit-media-controls-panel.hidden) -> (playing along with visible controls due to -webkit-media-controls-panel:hover which sets opacity to 1, but time is not updated because of controlsAreHidden == 1)
So the problem here is controlsAreHidden does not actually mean "hidden" and I check the opacity is really 0.

I don't want to argue this approach is the best. Any better idea is welcome. =)

[ChangSeok Oh]

This fixes audio controls broken as well.
Try http://www.quackit.com/html_5/tags/html_audio_tag.cfm

[Xabier Rodríguez Calvar]

(In reply to comment #3)
> As I described the situation in the ChangeLog, We changed not to update time
> if controllAreHidden is true in
> https://bugs.webkit.org/show_bug.cgi?id=137773.
> However this affects on controls behavior very much. Let guess this scenario.
> (move cursor on playbutton) -> (click play button) -> (hideControl timer
> fired by resetHideControlsTimer after 4 secs) -> (now controlsAreHidden is
> true since panel has neither -webkit-media-controls-panel.show nor
> -webkit-media-controls-panel.hidden) -> (playing along with visible controls
> due to -webkit-media-controls-panel:hover which sets opacity to 1, but time
> is not updated because of controlsAreHidden == 1)
> So the problem here is controlsAreHidden does not actually mean "hidden" and
> I check the opacity is really 0.

I understand now, good catch. You're right, another check needs to be done, though I don't think checking for opacity is the best way to go, but checking for the :hover property as :hover can change opacity, display or whatever we want from CSS. IMHO, we are safer with checking for :hover.

[ChangSeok Oh]

Created attachment 261311 [details]
Patch

[ChangSeok Oh]

(In reply to comment #5)
> (In reply to comment #3)
> > As I described the situation in the ChangeLog, We changed not to update time
> > if controllAreHidden is true in
> > https://bugs.webkit.org/show_bug.cgi?id=137773.
> > However this affects on controls behavior very much. Let guess this scenario.
> > (move cursor on playbutton) -> (click play button) -> (hideControl timer
> > fired by resetHideControlsTimer after 4 secs) -> (now controlsAreHidden is
> > true since panel has neither -webkit-media-controls-panel.show nor
> > -webkit-media-controls-panel.hidden) -> (playing along with visible controls
> > due to -webkit-media-controls-panel:hover which sets opacity to 1, but time
> > is not updated because of controlsAreHidden == 1)
> > So the problem here is controlsAreHidden does not actually mean "hidden" and
> > I check the opacity is really 0.
> 
> I understand now, good catch. You're right, another check needs to be done,
> though I don't think checking for opacity is the best way to go, but
> checking for the :hover property as :hover can change opacity, display or
> whatever we want from CSS. IMHO, we are safer with checking for :hover.

Otherwise we can use querySelector to check :hover existence. However, I'm not sure if that way is better than checking opacity directly in performance aspect..

[Xabier Rodríguez Calvar]

(In reply to comment #7)
> Otherwise we can use querySelector to check :hover existence. However, I'm
> not sure if that way is better than checking opacity directly in performance
> aspect..

LGTM. I don't think, unless you correct me, that performance is going to be a problem.

[ChangSeok Oh]

Comment on attachment 261311 [details]
Patch

Thanks!

[WebKit Commit Bot]

Comment on attachment 261311 [details]
Patch

Clearing flags on attachment: 261311

Committed r190053: <http://trac.webkit.org/changeset/190053>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.