Closed Bug 970220 Opened 10 years ago Closed 10 years ago

After invoking the setPointerCapture method on an element, subsequent pointer events for the specified pointer must be targeted at that element. Additionally, the relatedTarget property of all such pointer events must be set to null.

Categories

(Firefox for Metro Graveyard :: General, defect)

All
Windows 8
defect
Not set
normal

Tracking

(Not tracked)

RESOLVED FIXED
Firefox 31

People

(Reporter: alessarik, Assigned: alessarik)

References

Details

(Whiteboard: [pointer-events] p=0)

Attachments

(1 file, 5 obsolete files)

http://www.w3.org/wiki/PointerEvents/TestAssertions#Test_Assertions_for_setPointerCapture have a test 13.3 which have to be passed on MetroFireFox.
Actual results:
If we will use trunc version FF we will get relatedTarget property with pointer on DOM element.
Expected results:
According the specification we should have relatedTerget property as NULL.
Whiteboard: [triage]
No longer blocks: metrov1backlog
Whiteboard: [triage] → [pointer-events]
Before we land any of this we should fix this bug 968148
Attached patch For_test_13.3_updated.patch (obsolete) — Splinter Review
Some changes related with current version and future patches
Attachment #8373189 - Attachment is obsolete: true
Depends on: 968148
Blocks: metrobacklog
Whiteboard: [pointer-events] → [pointer-events] p=0
Attached patch change_related_target (obsolete) — Splinter Review
Comment was added in a head of patch
Attachment #8377101 - Attachment is obsolete: true
Patch should be commit after bug 970199
(In reply to Maksim Lebedev from comment #4)
> Patch should be commit after bug 970199

Did you mean bug 968148 ?
Comment on attachment 8378183 [details] [diff] [review]
change_related_target

>     PROFILER_LABEL("Input", "DispatchPointerEvent");
>-    newPointerEvent =
>+    nsAutoPtr<WidgetPointerEvent> newPointerEvent =
>       new WidgetPointerEvent(aMouseEvent->mFlags.mIsTrusted, aMessage,
>                              aMouseEvent->widget);

On non-windows build it cause build errors: see
dom/events/nsEventStateManager.cpp:4094:49: error: conversion from 'mozilla::WidgetPointerEvent*' to non-scalar type 'nsAutoPtr<mozilla::WidgetPointerEvent>' requested
https://tbpl.mozilla.org/php/getParsedLog.php?id=34917029&tree=Try#error0
Attachment #8378183 - Flags: feedback-
(In reply to Oleg Romashin (:romaxa) from comment #5)
> (In reply to Maksim Lebedev from comment #4)
> > Patch should be commit after bug 970199
> 
> Did you mean bug 968148 ?

I mean 970199, because some of changes in this patch is from patch in bug 970199
Attached patch change_related_target (obsolete) — Splinter Review
> On non-windows build it cause build errors: see
> dom/events/nsEventStateManager.cpp:4094:49: error: conversion from
> 'mozilla::WidgetPointerEvent*' to non-scalar type
> 'nsAutoPtr<mozilla::WidgetPointerEvent>' requested

I am very suprised that classes are compiled in different ways on different platforms. I hope this modified patch will fix this compile error.
Attachment #8378183 - Attachment is obsolete: true
Attachment #8381961 - Flags: review?(bugs)
Attachment #8381961 - Flags: review?(bugs) → review+
Attached patch change_related_target_ver5.diff (obsolete) — Splinter Review
Changes according with last version of sources
Attachment #8381961 - Attachment is obsolete: true
Attachment #8403926 - Flags: review?(mbrubeck)
Attachment #8403926 - Flags: review?(bugs)
Attachment #8403926 - Flags: feedback?(oleg.romashin)
Attachment #8403926 - Flags: feedback?(nicklebedev37)
Attachment #8403926 - Flags: review?(bugs) → review+
Attachment #8403926 - Flags: feedback?(oleg.romashin) → feedback+
Keywords: checkin-needed
Keywords: checkin-needed
File bug for this issue:
https://tbpl.mozilla.org/php/getParsedLog.php?id=37506653&tree=Try#error0
Status: UNCONFIRMED → NEW
Ever confirmed: true
Comment on attachment 8403926 [details] [diff] [review]
change_related_target_ver5.diff

Clearing review request because I'm not a DOM peer and because the test failure will need to be fixed before this lands.  However, the change looks good per the spec.
Attachment #8403926 - Flags: review?(mbrubeck)
> failure will need to be fixed before this lands.  However, the change looks
This patch does not break the tests by default, there are another patch which enabling pointer event properties by default (especially touch action) which break some of gtests..
1. We need new bug against gtest issue
2. Maxim plz don't include patch which is enabling all prefs by default into bug specific try builds.
Can I checkin this version of patch?
TRY build have several issues, but looks like its not related with my patch.
Comment on attachment 8403926 [details] [diff] [review]
change_related_target_ver5.diff

Review of attachment 8403926 [details] [diff] [review]:
-----------------------------------------------------------------

Maksim, could you please join the declaration and initialization of the newPointerEvent variable.
Attachment #8403926 - Flags: feedback?(nicklebedev37) → feedback+
(In reply to Nick Lebedev [:nl] from comment #16)
> Maksim, could you please join the declaration and initialization of the
> newPointerEvent variable.
I have such variant in first versions of patches,
but it have issue with compilation. Please see comment 6
+ update commit message
Attachment #8403926 - Attachment is obsolete: true
Keywords: checkin-needed
https://hg.mozilla.org/mozilla-central/rev/5805f3907c9a
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → FIXED
Target Milestone: --- → Firefox 31
You need to log in before you can comment on or make changes to this bug.