Side navigation
#14717 closed bug (cantfix)
Opened January 23, 2014 04:29AM UTC
Closed January 28, 2014 12:33AM UTC
onbeforeunload doesn't work on iOS 6-7
Reported by: | m_gol | Owned by: | m_gol |
---|---|---|---|
Priority: | low | Milestone: | |
Component: | event | Version: | 2.1.0-rc1 |
Keywords: | Cc: | ||
Blocked by: | Blocking: |
Description
iOS doesn't seem to support the onbeforeunload
event but window.onbeforeunload
is null
which makes our test suite believe it's supported and run the test. :(
This is the only test that's failing in iOS6, fixing/workarounding it would make it possible to register iOS6 in TestSwarm. (iOS7 fails more tests so that would come later).
I have no idea how to blacklist this test in Mobile Safari without resorting to UA sniffing.
Attachments (0)
Change History (6)
Changed January 23, 2014 04:33AM UTC by comment:1
component: | unfiled → event |
---|---|
milestone: | None → 1.11/2.1 |
owner: | → m_gol |
priority: | undecided → low |
status: | new → assigned |
Changed January 23, 2014 04:46AM UTC by comment:2
Changed January 27, 2014 05:53PM UTC by comment:3
resolution: | → fixed |
---|---|
status: | assigned → closed |
Tests: Blacklist beforeunload test in iOS
iOS has the window.onbeforeunload field but doesn't support the beforeunload
handler making it impossible to feature-detect the support.
Fixes #14717
Closes gh-1496
Changeset: 988d99ad278c2c9dc16c68f86bddc4df99dcd928
Changed January 28, 2014 12:21AM UTC by comment:4
I'm thinking we may want to close this actual ticket as cantfix
and we probably need somewhere to document this. That part is stumping me though, because a note in the API docs is likely to be overlooked. Perhaps Google will find this ticket.
Changed January 28, 2014 12:33AM UTC by comment:5
milestone: | 1.11/2.1 |
---|---|
resolution: | fixed |
status: | closed → reopened |
Changed January 28, 2014 12:33AM UTC by comment:6
resolution: | → cantfix |
---|---|
status: | reopened → closed |
You're right, it's not fixed per se, I just blacklisted a failing test so that the build passes. Definitely cantfix.
The problem is a support test would have to effectively trigger
unload
and ifbeforeunload
is not supported there's no way to cancel the unload. :/