Bug 193811 – iOS: inputmode="none" disables hardware keyboard's globe key

• Component: DOM
• Status: RESOLVED FIXED

[Ryosuke Niwa]

Setting inputmode="none" makes it impossible to change the keyboard type in hardware keyboard.

[Ryosuke Niwa]

Created attachment 360084 [details]
Unsupports inputmode=none for now

[Ryosuke Niwa]

<rdar://problem/47406553>

[Wenson Hsieh]

Comment on attachment 360084 [details]
Unsupports inputmode=none for now

Should we additionally remove the inputMode attribute from the HTMLElement interface?

[Wenson Hsieh]

(In reply to Wenson Hsieh from comment #3)
> Comment on attachment 360084 [details]
> Unsupports inputmode=none for now
> 
> Should we additionally remove the inputMode attribute from the HTMLElement
> interface?

Oops, sorry — what I meant to say was, should we additionally remove support for inputmode="none" when parsing/canonicalizing the inputmode attribute? (i.e. the logic in InputMode.cpp).

[Ryosuke Niwa]

(In reply to Wenson Hsieh from comment #4)
> (In reply to Wenson Hsieh from comment #3)
> > Comment on attachment 360084 [details]
> > Unsupports inputmode=none for now
> > 
> > Should we additionally remove the inputMode attribute from the HTMLElement
> > interface?
> 
> Oops, sorry — what I meant to say was, should we additionally remove support
> for inputmode="none" when parsing/canonicalizing the inputmode attribute?
> (i.e. the logic in InputMode.cpp).

Perhaps but I'm worried that such a change poses a more compatibility risk than not. Since inputmode is a hint, I think it's safer to ignore it. I could be proven wrong though...

[Wenson Hsieh]

(In reply to Ryosuke Niwa from comment #5)
> (In reply to Wenson Hsieh from comment #4)
> > (In reply to Wenson Hsieh from comment #3)
> > > Comment on attachment 360084 [details]
> > > Unsupports inputmode=none for now
> > > 
> > > Should we additionally remove the inputMode attribute from the HTMLElement
> > > interface?
> > 
> > Oops, sorry — what I meant to say was, should we additionally remove support
> > for inputmode="none" when parsing/canonicalizing the inputmode attribute?
> > (i.e. the logic in InputMode.cpp).
> 
> Perhaps but I'm worried that such a change poses a more compatibility risk
> than not. Since inputmode is a hint, I think it's safer to ignore it. I
> could be proven wrong though...

Fair enough. I guess we'll see!

[Ryosuke Niwa]

Comment on attachment 360084 [details]
Unsupports inputmode=none for now

Clearing flags on attachment: 360084

Committed r240497: <https://trac.webkit.org/changeset/240497>

[Ryosuke Niwa]

All reviewed patches have been landed.  Closing bug.

[Shawn Roberts]

The following test is a flaky crash on MacOS Debug

http/tests/resourceLoadStatistics/prune-statistics.html

Flakiness Dashboard: 

https://webkit-test-results.webkit.org/dashboards/flakiness_dashboard.html#showAllRuns=true&tests=http%2Ftests%2FresourceLoadStatistics%2Fprune-statistics.html

Probable cause: 

Can reproduce crash starting with Debug build 240497 to current 240715 with the following command: 

run-webkit-tests http/tests/resourceLoadStatistics/prune-statistics.html --iterations 500 -f --debug --exit-after-n-failures=1 --no-retry-failures

Crash log:

https://build.webkit.org/results/Apple%20High%20Sierra%20Debug%20WK2%20(Tests)/r240710%20(6440)/http/tests/resourceLoadStatistics/prune-statistics-crash-log.txt

Thread 0 Crashed:: Dispatch queue: com.apple.main-thread
0   com.apple.JavaScriptCore      	0x000000010d3aeb10 WTFCrash + 16 (Assertions.cpp:255)
1   com.apple.JavaScriptCore      	0x000000010d3aeb29 WTFCrashWithSecurityImplication + 9
2   com.apple.WebKit              	0x0000000105a2a262 WTF::RefCountedBase::ref() const + 66 (RefCounted.h:43)

[Ryosuke Niwa]

I think you commented on a wrong bug? This change only affects iOS code so it can't possibly regress a macOS debug test.

[Shawn Roberts]

I apologize, I commented on this bug in error.

(In reply to Ryosuke Niwa from comment #10)
> I think you commented on a wrong bug? This change only affects iOS code so
> it can't possibly regress a macOS debug test.