STR: 1. Load the attachment. 2. Press "ctrl shift M" to switch to responsive mode. 3. Click the "simulate touch events" button. 4. Reload the page. 5. Click in the green square. => It shows: touchstart 1 touchend 1 "1" is "event.touches.length". The "touchend" event should show "0" because it should have an empty event.touches TouchList. The same STR on Firefox OS shows: touchstart 1 touchend 0 which is expected according to the spec [1]. I haven't tried Firefox/Android but I suspect this is like FxOS. [1] http://w3c.github.io/touch-events/#dfn-touchend

I tried Aurora 46 and Nightly 47.

Here is a possible patch. Hey Ryan, do you know if this code is tested ? Where should I put some tests about this ?

(note: the patch is untested yet)

Comment on attachment 8722892 [details] [diff] [review] 0001-Bug-1250691-Devtools-Responsive-mode-The-touchend-ev.patch Review of attachment 8722892 [details] [diff] [review]: ----------------------------------------------------------------- ::: devtools/shared/touch/simulator-content.js @@ +288,4 @@ > let touches = document.createTouchList(point); > let targetTouches = touches; > let changedTouches = touches; > + if (name === "touchend") { even if we don't handle "touchcancel" anywhere, do you think I should test for it here as a future-proof rule of least surprise ?

Comment on attachment 8722892 [details] [diff] [review] 0001-Bug-1250691-Devtools-Responsive-mode-The-touchend-ev.patch Review of attachment 8722892 [details] [diff] [review]: ----------------------------------------------------------------- Here's an existing test you could extend: https://dxr.mozilla.org/mozilla-central/source/devtools/client/responsivedesign/test/browser_responsiveui_touch.js ::: devtools/shared/touch/simulator-content.js @@ +288,4 @@ > let touches = document.createTouchList(point); > let targetTouches = touches; > let changedTouches = touches; > + if (name === "touchend") { Sure, let's add `touchcancel` too.

Here is a patch with a passing test (the same test is failing on master). I'll start a try tomorrow and also check this is working with my more complex workload.

Comment on attachment 8723796 [details] [diff] [review] patch v2 actually it's still the old patch

Thanks for working on this!