Could be. I think the kernel thingy was trying to actually send key-up events
for every key that it thought was down, but maybe that was not really enough.
On 2010-08-25, at 18:29, Henry Minsky wrote:
> I think the implementation of allKeysUp in LzKeys is broken, that was one
> change I had in my fix. I added these two lines to actually clear the
> downkeys array state in the LFC, before it
> was just calling into the kernel
>
> function __allKeysUp ():void {
> this.downKeysHash = {};
> this.downKeysArray = [];
> LzKeyboardKernel.__allKeysUp();
> }
>
> So maybe I should see if that wrapper function is running and works with
> this change
>
> On Wed, Aug 25, 2010 at 5:18 PM, P T Withington <[email protected]> wrote:
>
>> It looks to me like embednew is already trying to do this, so maybe we just
>> need to figure out why it's not working?
>>
>> (Have a look at _gotFocus in embednew)
>>
>> On 2010-08-25, at 17:09, Henry Minsky wrote:
>>
>>> Hey Max,
>>>
>>> Where would I look in embed.js to find the event handlers in the wrapper
>>> that indicate that
>>> the application has lost or gained focus? I'd like to try calling in
>> using
>>> app.callMethod to
>>> clear the downkeys array in those cases, to see if that is a workable
>> patch
>>> for the
>>> 'ghost command key-down' bug.
>>>
>>> On Wed, Aug 25, 2010 at 3:16 PM, P T Withington <[email protected]
>>> wrote:
>>>
>>>> I just think we should be doing this in the wrapper in response to the
>>>> browser losing/gaining focus, not in the canvas and not based on the
>> mouse
>>>> moving.
>>>>
>>>> On 2010-08-25, at 14:46, Henry Minsky wrote:
>>>>
>>>>> I don't know that we ever got rid of this bug in DHTML actually. We
>> took
>>>>> care of issues when
>>>>> the mouse was still in the app, but this one I think was always
>> lurking.
>>>>>
>>>>> Maybe the hack should be conditionalized on DHTML runtime though.
>>>>>
>>>>>
>>>>> On Wed, Aug 25, 2010 at 2:40 PM, P T Withington <[email protected]
>>>>> wrote:
>>>>>
>>>>>> I'm not convinced this is the right solution.
>>>>>>
>>>>>> In previous attempts to solve this (
>>>>>> http://jira.openlaszlo.org/jira/browse/LPP-8283) we released the
>> stuck
>>>>>> keys when the browser lost and regained focus. I'm concerned that if
>> we
>>>>>> base this on the mouse we will get complaints that funny things happen
>>>> when
>>>>>> you push your mouse cursor out of the way...
>>>>>>
>>>>>> It's very clear that this issue is because Command-Tab takes you out
>> of
>>>> the
>>>>>> browser, so you don't see the key release. I'd like to know what made
>>>> the
>>>>>> previous solution stop working before we add another layer of mud to
>>>> this
>>>>>> diamond. :)
>>>>>>
>>>>>> On 2010-08-25, at 12:19, Henry Minsky wrote:
>>>>>>
>>>>>>> Change hqm-20100825-1vp by [email protected] on 2010-08-25
>> 12:03:54
>>>>>> EDT
>>>>>>> in /Users/hqm/openlaszlo/trunk/WEB-INF/lps/lfc
>>>>>>> for http://svn.openlaszlo.org/openlaszlo/trunk/WEB-INF/lps/lfc
>>>>>>>
>>>>>>> Summary: phantom keys marked as down in DHTML
>>>>>>>
>>>>>>> New Features:
>>>>>>>
>>>>>>> Bugs Fixed: LPP-8210
>>>>>>>
>>>>>>> Technical Reviewer: max
>>>>>>> QA Reviewer: ptw
>>>>>>> Doc Reviewer: (pending)
>>>>>>>
>>>>>>> Documentation:
>>>>>>>
>>>>>>> Release Notes:
>>>>>>>
>>>>>>> Overview:
>>>>>>>
>>>>>>>
>>>>>>> Details:
>>>>>>>
>>>>>>> Add event handler for canvas for onmousenter and onmouseleave, which
>>>>>> clear the downkeys array
>>>>>>> in the LFC and kernel Keyboard service.
>>>>>>>
>>>>>>> Tests:
>>>>>>>
>>>>>>> Run test/hello.lzx with debugger enabled in DHTML, use command-tab to
>>>>>> switch in and out of browser,
>>>>>>> verify that "enter" key works in debugger
>>>>>>>
>>>>>>> Files:
>>>>>>> M WEB-INF/lps/lfc/services/LzKeys.lzs
>>>>>>> M WEB-INF/lps/lfc/views/LaszloCanvas.lzs
>>>>>>>
>>>>>>>
>>>>>>> Changeset:
>>>>>> http://svn.openlaszlo.org/openlaszlo/patches/hqm-20100825-1vp.tar
>>>>>>
>>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Henry Minsky
>>>>> Software Architect
>>>>> [email protected]
>>>>
>>>>
>>>
>>>
>>> --
>>> Henry Minsky
>>> Software Architect
>>> [email protected]
>>
>>
>
>
> --
> Henry Minsky
> Software Architect
> [email protected]
