We must cancel the outstanding vibration pattern (i.e stop the device from vibrating), if the pattern is 0 or an empty list.
Created attachment 158529 [details] Patch
Looks fine to me. But, IMHO we need more detail description. It's hard to understand why you change vibrate logic like this.
Created attachment 158543 [details] Patch Thanks for the informal review Kihong. Updated changelog now.
Comment on attachment 158543 [details] Patch Can we please have a test for this?
Kenneth, thanks for looking into this. Unfortunately there is no return value in the Vibration API to check status. Same issue was discussed in bug 93957, comment 4.
Can't you do some private hooks for testing?
(In reply to comment #6) > Can't you do some private hooks for testing? I have no clue about this, but what I'm planning to do next is write unit tests for the API (in bug 93890) and make sure that vibrate/cancel callabcks are getting called correctly and the implementation is fully compliant with the spec.
EFL WK2 port has unit tests(landed in r125893) for the Vibration API now to verify the implementation.
Created attachment 159554 [details] Patch Rebased and updated changelog.
Comment on attachment 159554 [details] Patch The change looks reasonable. We wanted any test, but it would be OK for now based on the following comment. > EFL WK2 port has unit tests(landed in r125893) for the Vibration API now to verify the implementation.
Comment on attachment 159554 [details] Patch Clearing flags on attachment: 159554 Committed r126116: <http://trac.webkit.org/changeset/126116>
All reviewed patches have been landed. Closing bug.