GUI_Object binds both down & enter pointer events

[agmcleod]

melonJS/src/renderable/GUI.js

Line 99 in 9709270

me.input.registerPointerEvent("pointerenter", this, this.clicked.bind(this));

Realized due to bug in: https://groups.google.com/forum/#!topic/melonjs/204CYK6Yoe4

Given that pointerenter is the event that does not bubble for entering a section, I don't see how this should resolve to onClick. So i think we have 3 options:

1. Remove the calls to pointerenter & pointerleave
2. Keep the events, but bind to different callback functions. onEnter, onLeave
3. Change the constructor to accept options of what events should be bound.

[parasyte]

At a minimum, the pointerleave event needs to stay. GUI_Object might benefit from a new cancel action, instead of performing a click action when the pointer leaves the GUI rect.

[agmcleod]

Agreed on the cancel bit. But why do we need the pointerleave?

[parasyte]

For cases like this: blipjoy/sprung_fever#2

[agmcleod]

Ah okay, was thinking about it like with DOM events. Typically a mouseup on
a dom element will still fire on that element, but here we're doing at run
time rect calculation.

On Wed, May 27, 2015 at 3:20 PM Jay Oster [email protected] wrote:

For cases like this: blipjoy/sprung_fever#2
blipjoy/sprung_fever#2

—
Reply to this email directly or view it on GitHub
#686 (comment).

[obiot]

agree, the pointerEnter event should not trigger the onClicked event, but rather use a specific callback (onEnter, onLeave as suggested by Aaron).

if we all agree, I can modify this for inclusion in the 2.1.1 release ?

[obiot]

hi guys, see the above commit : is this what we all agreed on ?

[parasyte]

As long as the default onLeave handler cancels the current hover/click state.

[obiot]

tell me again why the onPointerLeave event should also cancel the click state ?

[parasyte]

I posted the reason in this ticket two days ago. 😞 Here's the link again: blipjoy/sprung_fever#2

If leave doesn't cancel the click, you get really broken behavior.

[obiot]

Ahha sorry, I saw it actually, but then just forgot, must be one of these "senior moments" :p

On 29 mai 2015, at 11:38, Jay Oster [email protected] wrote:

I posted the reason in this ticket two days ago. Here's the link again: blipjoy/sprung_fever#2

If leave doesn't cancel the click, you get really broken behavior.

—
Reply to this email directly or view it on GitHub.

[obiot]

we should be all set now then ! :p

i also added a test on the release flag to avoid multiple event triggering :
1bbc4e2#diff-4542aeba5835ce6807cf6e36a774d767R205

[obiot]

everybody's happy with this one ?

[agmcleod]

Yep
On Sat, May 30, 2015 at 11:37 AM Olivier Biot [email protected]
wrote:

everybody's happy with this one ?

—
Reply to this email directly or view it on GitHub
#686 (comment).

[obiot]

great, I will wait the feedback from Lamberto before closing it anyway