- From: Maksim Lebedev <alessarik@gmail.com>
- Date: Fri, 22 Aug 2014 17:24:53 +0400
- To: public-pointer-events@w3.org
- Message-ID: <CAMDeA0ZMCBTn1syDXMxDf1c0JA54g+i53RCtZFtjXbwWFtYbFQ@mail.gmail.com>
Hi folks. I am very doubt by tests of asyncronoun events. pointerevent_gotpointercapture_before_first_pointerevent-manual.html pointerevent_lostpointercapture_is_first-manual.html According specification of pointer events got/lostpointercapture events should be async. But latest changes made something, which was confusing me: https://github.com/w3c/web-platform-tests/commit/e4c7116d8cb7289b84ab2feddd7178fda23b4f5e https://github.com/w3c/web-platform-tests/commit/24577bb71789010738a0b1cd9d371a07f0f2cfe1 After this changes we can find that test execute setPointerCapture (or releasePointerCapture) and after that timer is setted for changing isAsync to true. Also we can find that on gotpointercapture event (or on lostpointercaptere event) tests checks isAsync. In what sequence this items should be executed? Timer with (isAsync = true) at first. And after this gotpointercapture should be fired. Opposite. gotpointercapture should be fired at first, And only about this Timer with (isAsync = true) There is specification for timers: http://www.whatwg.org/specs/web-apps/current-work/multipage/webappapis.html#timers I can assume that current version of tests are incorrect. And author of changes notifyed that IE didn't pass tests: https://github.com/w3c/web-platform-tests/pull/1113#issue-38207053 At this point, if tests are incorrect, then tests should be changed for correct state. I can propose to delete creating timer and simply only set isAsync to true after setPointerCapture (releasePointerCapture) Please, let me know your opinion. Thanks, Maksim Lebedev.
Received on Sunday, 24 August 2014 08:34:23 UTC