On Thu, Feb 14, 2013 at 12:43 PM, Carsten Haitzler <[email protected]> wrote:
> On Wed, 13 Feb 2013 17:02:44 +0530 Arvind R <[email protected]> said:
>
>> On Tue, Feb 12, 2013 at 4:41 PM, Carsten Haitzler <[email protected]>
>> wrote:
>> > On Fri, 1 Feb 2013 22:29:56 +0530 Arvind R <[email protected]> said:
>> >
>> > there - i did my changes to update less frequently - different to your
>> > framecount method (time based with timeouts and deletion handling). rev
>> > 83854.
>> >
>> On further probing, I found the following modification (included in
>> the post on completing the mouse interface) to really clobber the bug.
>> This does not require any minimisation/delay of update frequency, ...
>> . It works on all my test clips with all backends (of emotion). From
>> this, maybe you can figure out the actual cause of the bug - it may
>> not be solved by your timer fix - I have to check it out.
>
> oh my timer thing doesnt fix any bug - it just improves the ui to not update
> TOO often (eg 30 or 20 or 60 times a second, where each update may be very
> minimal). its an optimization for things like audio playback for example.
>
> the below smells of a workaround... do you STOP getting changed callbacks
> after
> this?
>
No. But the changed callbacks don't disrupt the plauback -:) Yes, it
IS a workaround.
>> @@ -176,9 +248,15 @@ _update_position(void *data,
>> Evas_Object *obj __UNUSED__,
>> void *event_info __UNUSED__)
>> {
>> + double pos;
>> +
>> ELM_PLAYER_DATA_GET(data, sd);
>>
>> - elm_video_play_position_set(sd->video, elm_slider_value_get(sd->slider));
>> + pos = elm_slider_value_get(sd->slider);
>> + if (fabs(sd->pos - pos) > 0.1) {
>> + elm_video_play_position_set(sd->video, pos);
>> + sd->pos = pos;
>> + }
>> }
>>
>>
>> >> Hi,
>> >>
>> >> The slider emits a "changed" signal when its internal representation
>> >> of value really changes. It does not know whether the user dragged it
>> >> or internal updates caused the value change. Most internal incremental
>> >> updates when trimmed to the current resolution of the slider, do not
>> >> cause a value change and the signal is not emitted on every "set".
>> >>
>> >> This behaviour causes a 'feedback-loop' in that the "changed" signal
>> >> from an internal update, being undifferentiated from a "user-drag",
>> >> causes a delayed reposition command to the emotion engine. The result
>> >> is a nice 'track-jump-repeat' which in some extreme cases, a
>> >> 'stuck-in-a-groove' effect.
>> >>
>> >> The foll. patch is a solution to this problem; and also reduces the
>> >> frequency of slider updates. An alternate solution would be to use
>> >> _smart_callback_del() before actually updating slider in
>> >> elc_player.c:_update_slider(), and re-installing it back - but this
>> >> could open the door to race conditions.
>> >>
>> >> Arvind
>> >> ---
>> >>
>> >> diff -uprN a/src/lib/elc_player.c b/src/lib/elc_player.c
>> >> --- a/src/lib/elc_player.c 2013-02-01 21:59:39.000000000 +0530
>> >> +++ b/src/lib/elc_player.c 2013-02-01 21:59:54.000000000 +0530
>> >> @@ -163,12 +163,35 @@ _update_slider(void *data,
>> >> ELM_PLAYER_DATA_GET(data, sd);
>> >>
>> >> seekable = elm_video_is_seekable_get(sd->video);
>> >> + if (seekable == sd->disabled)
>> >> + elm_object_disabled_set(sd->slider, !seekable);
>> >> +
>> >> length = elm_video_play_length_get(sd->video);
>> >> - pos = elm_video_play_position_get(sd->video);
>> >>
>> >> - elm_object_disabled_set(sd->slider, !seekable);
>> >> - elm_slider_min_max_set(sd->slider, 0, length);
>> >> - elm_slider_value_set(sd->slider, pos);
>> >> + if (sd->length != length) {
>> >> + sd->length = length;
>> >> + elm_slider_min_max_set(sd->slider, 0, length);
>> >> + }
>> >> +
>> >> + if (sd->frames < 0) {
>> >> + pos = elm_video_play_position_get(sd->video);
>> >> + if (fabs(sd->pos - pos) > 0.5) {
>> >> + sd->pos = pos;
>> >> + sd->frames = 30;
>> >> + elm_slider_value_set(sd->slider, pos);
>> >> + }
>> >> + }
>> >> +}
>> >> +
>> >> +static void
>> >> +_frame_decoded_cb(void *data,
>> >> + Evas_Object *obj __UNUSED__,
>> >> + void *event_info __UNUSED__)
>> >> +{
>> >> + ELM_PLAYER_DATA_GET(data, sd);
>> >> + /* do not update slider on every frame */
>> >> + if (--sd->frames < 0)
>> >> + _update_slider(data, obj, event_info);
>> >> }
>> >>
>> >> static void
>> >> @@ -176,9 +199,24 @@ _update_position(void *data,
>> >> Evas_Object *obj __UNUSED__,
>> >> void *event_info __UNUSED__)
>> >> {
>> >> + /*
>> >> + * called wnenever the value of the slider is changed.
>> >> + * - irespective of user drags
>> >> + * or
>> >> + * internal increments larger than the slider resolution.
>> >> + * the 2nd case causes an inadvertent repositioning of stream
>> >> + * which some engines respond to (xine).
>> >> + * This is a feedback loop that needs to be broken.
>> >> + */
>> >> + double pos;
>> >> +
>> >> ELM_PLAYER_DATA_GET(data, sd);
>> >>
>> >> - elm_video_play_position_set(sd->video, elm_slider_value_get
>> >> (sd->slider));
>> >> + pos = elm_slider_value_get(sd->slider);
>> >> + if (fabs(sd->pos - pos) > 0.25) {
>> >> + elm_video_play_position_set(sd->video, pos);
>> >> + sd->pos = pos;
>> >> + }
>> >> }
>> >>
>> >> static void
>> >> @@ -455,7 +493,7 @@ _elm_player_smart_content_set(Eo *obj, v
>> >> else elm_layout_signal_emit(obj, "elm,player,pause", "elm");
>> >>
>> >> evas_object_smart_callback_add(sd->emotion, "frame_decode",
>> >> - _update_slider, obj);
>> >> + _frame_decoded_cb, obj);
>> >> evas_object_smart_callback_add(sd->emotion, "frame_resize",
>> >> _update_slider, obj);
>> >> evas_object_smart_callback_add(sd->emotion, "length_change",
>> >> @@ -495,9 +533,13 @@ _elm_player_smart_add(Eo *obj, void *_pd
>> >> (priv->slider, _double_to_time, _str_free);
>> >> elm_slider_units_format_function_set
>> >> (priv->slider, _double_to_time, _str_free);
>> >> - elm_slider_min_max_set(priv->slider, 0, 0);
>> >> + elm_slider_min_max_set(priv->slider, 0, 1);
>> >> elm_slider_value_set(priv->slider, 0);
>> >> elm_object_disabled_set(priv->slider, EINA_TRUE);
>> >> + priv->disabled = EINA_TRUE;
>> >> + priv->length = 0.0;
>> >> + priv->frames = 0;
>> >> + priv->pos = 0.0;
>> >> evas_object_size_hint_align_set(priv->slider, EVAS_HINT_FILL, 0.5);
>> >> evas_object_size_hint_weight_set
>> >> (priv->slider, EVAS_HINT_EXPAND, EVAS_HINT_EXPAND);
>> >> diff -uprN a/src/lib/elm_widget_player.h b/src/lib/elm_widget_player.h
>> >> --- a/src/lib/elm_widget_player.h 2012-11-26 12:02:53.000000000 +0530
>> >> +++ b/src/lib/elm_widget_player.h 2013-02-01 21:59:54.000000000 +0530
>> >> @@ -33,6 +33,11 @@ struct _Elm_Player_Smart_Data
>> >> Evas_Object *rewind;
>> >> Evas_Object *stop;
>> >> Evas_Object *slider;
>> >> +
>> >> + Eina_Bool disabled; /* used for only-when-must changes */
>> >> + double length; /* --- ditto --- */
>> >> + short frames; /* used to reduce update frequency */
>> >> + double pos; /* inter-callback communication to break
>> >> slider feedback loop */
>> >> };
>> >>
>> >> /**
>> >>
>> >> ------------------------------------------------------------------------------
>> >> Everyone hates slow websites. So do we.
>> >> Make your web apps faster with AppDynamics
>> >> Download AppDynamics Lite for free today:
>> >> http://p.sf.net/sfu/appdyn_d2d_jan
>> >> _______________________________________________
>> >> enlightenment-users mailing list
>> >> [email protected]
>> >> https://lists.sourceforge.net/lists/listinfo/enlightenment-users
>> >>
>> >
>> >
>> > --
>> > ------------- Codito, ergo sum - "I code, therefore I am" --------------
>> > The Rasterman (Carsten Haitzler) [email protected]
>> >
>>
>
>
> --
> ------------- Codito, ergo sum - "I code, therefore I am" --------------
> The Rasterman (Carsten Haitzler) [email protected]
>
------------------------------------------------------------------------------
Free Next-Gen Firewall Hardware Offer
Buy your Sophos next-gen firewall before the end March 2013
and get the hardware for free! Learn more.
http://p.sf.net/sfu/sophos-d2d-feb
_______________________________________________
enlightenment-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/enlightenment-users
