- From: Dominique Hazael-Massieux <dom@w3.org>
- Date: Tue, 12 Jan 2016 18:23:02 +0100
- To: "Kostiainen, Anssi" <anssi.kostiainen@intel.com>, "Zhang, Zhiqiang" <zhiqiang.zhang@intel.com>
- Cc: James Graham <james@hoppipolla.co.uk>, W3C Device APIs WG <public-device-apis@w3.org>, "ms2ger@gmail.com" <ms2ger@gmail.com>
On 11/01/2016 13:56, Kostiainen, Anssi wrote: > Zhiqiang - can you summarize the reason for test failures in > all.html, in particular for test cases that fail in both Chrome and > Firefox and are not manual tests? > > I believe battery-interface-idlharness.html and > battery-interface.html tests are overlapping, so we should pick one > and drop another. If promises support in idlharness.js is limited, we > might consider using battery-interface.html instead and patch it > where needed to get good test coverage. Or better, patch > idlharness.js, and use the former tests. I've found a bug in how idlharness was being used for battery, and have brought a patch: https://github.com/w3c/web-platform-tests/pull/2469 With that patch (and the the promise-throw patch to ildharness), Firefox gets 100% on the idlharness test. Chrome has 4 failures: * one due to throwing instead of rejecting on promise returning methods (in this case, getBattery()) * one due to a bad class string for the prototype of BatteryManager (not sure where that comes from) * two due to not implementing addEventListener/removeEventListener as expected. I think the 1st and last 2 bugs can be argued as not specific to Battery API; the 2nd one probably deserves more investigation though. I haven't looked at the manual test cases failures yet. Dom
Received on Tuesday, 12 January 2016 17:23:15 UTC