Issue metadata
Sign in to add a comment
|
Should webview test team conduct tests using a different Accessibility service too? |
||||||||||||||||||||||
Issue descriptionDuring the investigation of b/64951517, it became obvious that using talkback as a screen reader does not cover all the corner cases. Especially see comment #20. We could extend our tests using a different service, but we need to find a service that is maintained and test team invest time to learn how to use it. IMO, using talkback is ok as this is the first time we see this issue.
,
Aug 31 2017
,
Aug 31 2017
I'm puzzled by the phrase, "using talkback as a screen reader". TalkBack /is/ a screen reader. What corner cases are we obligated to test which are not covered by TalkBack? Does Clank test with any accessibility services other than TalkBack? You're referring to b/64951517#comment18, which says "TalkBack is just one implementation of accessibility services, and it provides 'read-only' accessibility, i.e., it reads contents of the screen and convert to speech, but it does not modify the contents of the screen." That's incorrect. With TalkBack on, there's a different set of gestures (touch-to-explore, double-tap, 2-finger scroll, etc.) which allow you to perform all the same actions as when TalkBack is off. If this claim is the basis for our belief that testing with TalkBack is insufficient, then there's nothing we need to do here.
,
Aug 31 2017
the cornercase is the bug itself, see b/64951517 comment #1 and #22. I do not believe clank tests with any other accessibility services. for your comment on #18: IMO, one of these are correct: 1. Talkback can be used to test for the case that they are bringing in the bug, but we don't know how to test it 2. Talkback is not a full implementation as they claim in #20, and we need to test using a different accessibility service 3. Their test case is wrong If you can help me untangle this, I appreciate :)
,
Aug 31 2017
If I'm understanding correctly, the relevant thing Talkback doesn't do, but which accessibility supports, is directly modifying the state of input fields via the accessibility APIs. Talkback just defers to the IME for this (you use the regular keyboard, with speech feedback to let you know which keys you're hitting). The accessibility service inserting text into fields needs to be treated as a user gesture, because the accessibility service is presumed to be the user's agent. There's no such thing as a "full" vs "not full" implementation of accessibility services here - there's a bunch of different APIs and different services use different subsets, because they're supporting different accessibility requirements and user preferences.
,
Aug 31 2017
(for example, Voice Access supports saying "type hello" into the microphone, and it will enter "hello" into the currently focused field directly, without the IME being involved)
,
Sep 3
This issue has been Available for over a year. If it's no longer important or seems unlikely to be fixed, please consider closing it out. If it is important, please re-triage the issue. Sorry for the inconvenience if the bug really should have been left as Available. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Sep 4
|
|||||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||||
Comment 1 by sgu...@chromium.org
, Aug 31 2017