Bug 85369 – [EFL] [DRT] keyboard-related tests do not pass

• Component: WebKit EFL
• Status: RESOLVED FIXED

[Mikhail Pozdnyakov]

PlatformKeyboardEventEfl implementation should be fixed:
- event.keyIdentifier, event.keyCode, event.charCode should have proper values 
- PlatformKeyboardEvent::getCurrentModifierState should be implemented

in order to unskip:
fast/events/js-keyboard-event-creation.html

fast/events/key-events-in-input-button.html
fast/events/key-events-in-input-text.html
fast/events/keydown-numpad-keys.html
fast/events/onsearch-enter.html
fast/events/option-tab.html
fast/events/special-key-events-in-input-text.html
fast/forms/button-enter-click.html
fast/forms/button-spacebar-click.html
fast/forms/enter-clicks-buttons.html
fast/forms/input-search-press-escape-key.html
fast/forms/input-text-enter.html
fast/forms/textinput-not-fired-on-enter-in-input.html
http/tests/misc/isindex-with-no-form.html

[Mikhail Pozdnyakov]

Created attachment 142660 [details]
patch

Fixes the left testcases from the list.

[Gyuyoung Kim]

Comment on attachment 142660 [details]
patch

View in context: https://bugs.webkit.org/attachment.cgi?id=142660&action=review

> LayoutTests/ChangeLog:3
> +        [EFL] PlatformKeyboardEventEfl implementation fix

This patch implements new functionality, fix potential bug and refactoring code. Please do not mix multiple patches into a patch.

> Source/WebCore/platform/efl/EflKeyboardUtilities.cpp:228
> +        return "\x1b";

To be consistent with other condition, let's use String("").

> Source/WebKit/efl/WebCoreSupport/EditorClientEfl.cpp:444
> +    // From GTK. Don't allow text insertion for nodes that cannot edit.

It looks this condition is clear. So, it seems to me "From GTK" is not needed.

[Mikhail Pozdnyakov]

(In reply to comment #2)
> (From update of attachment 142660 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=142660&action=review
> 
> > LayoutTests/ChangeLog:3
> > +        [EFL] PlatformKeyboardEventEfl implementation fix
> 
> This patch implements new functionality, fix potential bug and refactoring code. Please do not mix multiple patches into a patch.

Thanks for review. Splitting it..

[Mikhail Pozdnyakov]

Created attachment 143483 [details]
patch v2 (drt changes)

[Gyuyoung Kim]

Comment on attachment 143483 [details]
patch v2 (drt changes)

View in context: https://bugs.webkit.org/attachment.cgi?id=143483&action=review

> Tools/DumpRenderTree/efl/EventSender.cpp:455
> +        modifiers |= EvasKeyModifierShift;

As we discussed on irc, I think we need to know why shift should be adjust into modifiers before landing this patch.

[Mikhail Pozdnyakov]

(In reply to comment #5)
> (From update of attachment 143483 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=143483&action=review
> 
> > Tools/DumpRenderTree/efl/EventSender.cpp:455
> > +        modifiers |= EvasKeyModifierShift;
> 
> As we discussed on irc, I think we need to know why shift should be adjust into modifiers before landing this patch.

Probable reason is that VK_ codes do not distinguish between lows and caps, and having of SHIFT modifier is the only way to understand the case

[Gyuyoung Kim]

Comment on attachment 143483 [details]
patch v2 (drt changes)

View in context: https://bugs.webkit.org/attachment.cgi?id=143483&action=review

> Tools/DumpRenderTree/efl/EventSender.cpp:454
> +    if ((character->length() == 1 ) && ('A' <= charCode && charCode <= 'Z'))

I'm not sure if this is webkit style, personally, I would prefer operand to operator in comparison condition. For example,
(charCode >= 'A' && charCode <= 'Z')

>>> Tools/DumpRenderTree/efl/EventSender.cpp:455
>>> +        modifiers |= EvasKeyModifierShift;
>> 
>> As we discussed on irc, I think we need to know why shift should be adjust into modifiers before landing this patch.
> 
> Probable reason is that VK_ codes do not distinguish between lows and caps, and having of SHIFT modifier is the only way to understand the case

If Virtual key codes only can distinguish them, I'm ok. But, in my opinion, it is good to add comment for this to here.

[Gyuyoung Kim]

(In reply to comment #7)

>> If Virtual key codes only can distinguish them
Oops, wrong typing.

If virtual key codes only can distinguish between lows and caps when there is shift modifier in modifier, I'm ok.

[Mikhail Pozdnyakov]

(In reply to comment #8)
> (In reply to comment #7)
> 
> >> If Virtual key codes only can distinguish them
> Oops, wrong typing.
> 
> If virtual key codes only can distinguish between lows and caps when there is shift modifier in modifier, I'm ok.

Here is the actual reason described:

From d15ea3fef52046f49dbca9a8f7633799255c7750 Mon Sep 17 00:00:00 2001
From: "[e-mail]" <[e-mail]@268f45cc-cd09-0410-ab3c-d52691b4dbfc>
Date: Thu, 29 Nov 2007 00:43:14 +0000
Subject: [PATCH]         Reviewed by Adam Roben.

        * DumpRenderTree/mac/EventSendingController.mm:
        (-[EventSendingController keyDown:withModifiers:]):
        Send capital letters through as lowercase letters with the shift key down
        rather than sending them as if they were highly unusual "capital letter keys".

[Gyuyoung Kim]

(In reply to comment #7)
> > Tools/DumpRenderTree/efl/EventSender.cpp:454
> > +    if ((character->length() == 1 ) && ('A' <= charCode && charCode <= 'Z'))
> 
> I'm not sure if this is webkit style, personally, I would prefer operand to operator in comparison condition. For example,
> (charCode >= 'A' && charCode <= 'Z')

Any opinions ?

[Raphael Kubo da Costa (:rakuco)]

Comment on attachment 143483 [details]
patch v2 (drt changes)

View in context: https://bugs.webkit.org/attachment.cgi?id=143483&action=review

> Tools/DumpRenderTree/efl/EventSender.cpp:442
> -    int charCode = JSStringGetCharactersPtr(character)[0];
> +    const char charCode = JSStringGetCharactersPtr(character)[0];

Why change this?

>> Tools/DumpRenderTree/efl/EventSender.cpp:454
>> +    if ((character->length() == 1 ) && ('A' <= charCode && charCode <= 'Z'))
> 
> I'm not sure if this is webkit style, personally, I would prefer operand to operator in comparison condition. For example,
> (charCode >= 'A' && charCode <= 'Z')

I second Gyuyoung's comment, and there's also an extra space after "1".

>>>> Tools/DumpRenderTree/efl/EventSender.cpp:455
>>>> +        modifiers |= EvasKeyModifierShift;
>>> 
>>> As we discussed on irc, I think we need to know why shift should be adjust into modifiers before landing this patch.
>> 
>> Probable reason is that VK_ codes do not distinguish between lows and caps, and having of SHIFT modifier is the only way to understand the case
> 
> If Virtual key codes only can distinguish them, I'm ok. But, in my opinion, it is good to add comment for this to here.

Adding a comment would be good for future reference.

[Mikhail Pozdnyakov]

(In reply to comment #11)
> (From update of attachment 143483 [details])
> View in context: https://bugs.webkit.org/attachment.cgi?id=143483&action=review
> 
> > Tools/DumpRenderTree/efl/EventSender.cpp:442
> > -    int charCode = JSStringGetCharactersPtr(character)[0];
> > +    const char charCode = JSStringGetCharactersPtr(character)[0];
> 
> Why change this?
Did it to avoid unnecessary conversions each time we compare 'charCode' with a symbol. Will remove it from the patch as it is not related to actual fix of DRT.

[Mikhail Pozdnyakov]

Created attachment 143814 [details]
patch v3 (fixed review comments)

[Mikhail Pozdnyakov]

(In reply to comment #10)
> (In reply to comment #7)
> > > Tools/DumpRenderTree/efl/EventSender.cpp:454
> > > +    if ((character->length() == 1 ) && ('A' <= charCode && charCode <= 'Z'))
> > 
> > I'm not sure if this is webkit style, personally, I would prefer operand to operator in comparison condition. For example,
> > (charCode >= 'A' && charCode <= 'Z')
> 
> Any opinions ?

Let's do this way. Fixed.

[Raphael Kubo da Costa (:rakuco)]

Comment on attachment 143814 [details]
patch v3 (fixed review comments)

Looks good, thanks.

[WebKit Review Bot]

Comment on attachment 143814 [details]
patch v3 (fixed review comments)

Clearing flags on attachment: 143814

Committed r118404: <http://trac.webkit.org/changeset/118404>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.