[Laszlo-dev] Need some help [#LPP-5447] DHTML: inputtext and clickable - OpenLaszlo Jira

[P T Withington]

On 2009-07-22, at 17:46EDT, Max Carlson wrote:

> P T Withington wrote:
>> [[#LPP-5447] DHTML: inputtext and clickable - OpenLaszlo Jira](http://jira.openlaszlo.org/jira/browse/LPP-5447 
>> )
>> What is the difference between LzInputTextSprite/ 
>> __setglobalclickable and LzMouseKernel.setGlobalClickable?  In  
>> particular, the latter just turns off the whole click tree by  
>> hiding the root of the tree (which seems should be pretty  
>> efficient) and the latter turns off only divs that have the class  
>> `lzclickdiv` by modifying the CSS style rule associated with that  
>> class (which seems might be more expensive).  Why the two different  
>> methods?
>
> LzInputTextSprite/__setglobalclickable() is used by inputtexts to  
> disable clickable.  Since inputtexts currently bind into the click  
> div (see LzInputTextSprite.__show) the clickdiv itself can't be  
> hidden - or else the inputtext would be hidden also.

Aha!  But if we don't reparent the input text (the current proposed  
solution), we can dispense with that, and just use the more efficient  
mechanism of turning off the tree at the root.

> LzMouseKernel.setGlobalClickable() is used by the html tag to  
> disable the click tree so it doesn't interfere with an iframe's.
>
>> Can you explain the purpose of __lastShown, __focusedSprite, and  
>> __lastfocus?  I can't follow what these are trying to do.
>
> __lastshown tracks the last inputtext to be shown by __show(), and  
> is used to hide the currently showing inputtext.
> __focusedSprite tracks the last inputtext to be focused, and is used  
> to work around bugs in firefox's focus management and prevent  
> spurious/extra onfocus/blur events from being sent.
> __lastfocus holds a reference to the last inputtext to be selected -  
> by select() or setSelection().  It's used as a callback (see  
> setTimout()) to work around a bug in IE where a field can't be  
> selected immediately after it's focused.
>
> Complicated, no?  Since you're in there, can you please add these  
> notes as comments?  Thanks!

Yes, but I'm thinking some of this will go away if we don't have to  
reparent the inputtext element into the click tree.  Which makes me  
wonder why we were doing that in the first place?  Are we going to  
break something else by not reparenting?

If I understand correctly, all the show/hide/focus stuff has to do  
with the input element not being where it would normally be.  If we  
just leave it be, will we still need all these?