[patch] [ccla] Dojo touch event layer for both mobile and PC

[evan]

We'll use this ticket to track touch event layer which is designed to recognize various user gestures with appropriate events published.

It also offers unified event APIs for developers to catch and respond to the raised events across browsers not only on PC but also on mobile devices such as iPhone, iPad and Andriod etc.

Please refer to the following materials:

​1. Implementation proposal

​2. Live demo(only tested on iPhone v.3GS+)

​3. Demo video(demo.wmv)

[evan]

Breakdown the original patch(from Siqi Zhong, IBM CCLA) into three - please apply to dijit, dojo and dojox separately.

[evan]

Demo for touch0.1.patch, experimenting basic touches & gestures(tap, rotation) with dijit Slider and Charting, from Si Qi Zhong(IBM, CCLA), Ruan Qi (IBM, CCLA)

[Chris Mitchell]

Has there been any further feedback on this ticket?

I've reviewed this latest patch, and it looks very clean with the new event module refractor from kris, but looks like the events patch is out of date with trunk. -Chris

[evan]

Replying to chrism:

Has there been any further feedback on this ticket?

I've reviewed this latest patch, and it looks very clean with the new event module refractor from kris, but looks like the events patch is out of date with trunk. -Chris

Yep, the discussion is currently going on in mailing list. Kris has already fixed it in git, just uploaded a latest ​path version there

[evan]

Refactored based on ​new event framework, also added support for isolating mouse/touch handlers

[evan]

touch0.3-all-in-one.patch includes ​new event/listen by adding normalization works for W3C touch events to dojo.listen, also includes a formal impl of Tap/Swipe? gestures from Eduardo(IBM, CCLA). The usages look like

1. Unified W3C touch events working well across touch devices
   

dojo.connect(node, "ontouchstart"|"ontouchmove"|..., func);
dojo.connect(node, "onorientationchange", func);

2. Abstracted touch/mouse events working well on desktop and mobiles - future dijit/dnd changes
   

dojo.connect(node, dojo.touch.press | move | release, func);

3. Gestures working well on desktop and mobiles:
   

dojo.connect(n, dojo.gesture.tap | taphold | doubletap, func);
dojo.connect(n, dojo.gesture.swipe | swipeup | swipedown | swipeleft | swiperight, func);

[evan]

AMD issue fixed working well with trunk, gesture event bubbling issue still exists

[evan]

Fixed with [24634] - by getting the touch & gestures code into trunk for 1.7

[evan]

(In [25050]) Refs #12176 making gesture bubbling compatible with iOS v.3

[evan]

(In [25701]) Refs #12176 clean up resources when a gesture event is disconnected or the global gesture instance is unRegistered.

[evan]

(In [25706]) Refs #12176 !strict missed dojo.gesture.destroy() and related doh test cases and space fix

[bill]

Can you make touch_gestures.html be run by runTests.html? Or is there a reason it isn't? (Does it work on IE?)

Actually, the automated test doesn't seem to work on FF (FF6/mac). It throws DOH assertions. The other problem with the test though is that those assertions aren't reported because you are running doh.is() tests inside of a dojo.connect() callback. (But you can see it in firebug. It's in the gestureBubbling() test on on.emit(innerDoH, 'mouseup', {}).

[Chris Mitchell]

The touch demo is broken on trunk.

[Evan]

In [26541]:

Fixes #12176, fixes regressions due to #13828 changes

[Evan]

In [26542]:

Refs #12176 added test_touch.html into dojo/doh harness

[Evan]

Replying to bill:

Can you make touch_gestures.html be run by runTests.html? Or is there a reason it isn't? (Does it work on IE?)

Actually, the automated test doesn't seem to work on FF (FF6/mac). It throws DOH assertions. The other problem with the test though is that those assertions aren't reported because you are running doh.is() tests inside of a dojo.connect() callback. (But you can see it in firebug. It's in the gestureBubbling() test on on.emit(innerDoH, 'mouseup', {}).

Yep, this was fixed with [26542], [26512] and [26540]

[bill]

I don't think it's quite fixed, as I was trying to say above you can't just put doh.is() etc. calls inside of a callback like you are doing in (for example) swipe.html:

h2 = on(inner, swipe.end, function(e){
	endExecuted = true;
	doh.assertEqual('swipe.end', e.type);
	doh.assertEqual(inner, e.target);
	doh.assertEqual(200, e.dx);
	doh.assertEqual(-120, e.dy);
});

If the assertions fail they will throw exceptions, but the execeptions will be swallowed.

[Evan]

In [26549]:

Refs #12176, refined doh cases a bit so that assertions won't be swallowed.

[Evan]

In [26550]:

Refs #12176, also refined test_touch.html doh cases a bit so that assertions won't be swallowed.

[Bill Keese <bill@…>]

In 755fbbc3be5cb56555d1f237d3c6418569418b45/dojo:

Error: Processor CommitTicketReference failed

Unsupported version control system "git": Can't find an appropriate component, maybe the corresponding plugin was not enabled?