P T Withington wrote:
> [Adding Max: couple of questions for you below.]
>
> Comments:
>
> LzSprite.as:
>
> 1) I don't think you need any sprite code for getAttributeRelative. The
> makeContainerResource was necessary in the old implementation in order
> to ask the SWF runtime to calculate the transformation, now that you are
> calculating the transformation portably, you don't need to ensure there
> is a container.
Sounds good to me!
> LzSprite.js:
>
> 1) You need to check with Max whether the LzSprite.setHeight interface
> is actually meant to support percentages. I don't think it is, and if
> not, that code should be ripped out. But if it is, then you can't make
> the NaN test before that code, since `isNaN(50%) == true`.
This code should probably be removed... It was there to make
percentages work with canvas heights. It should now be handled by the
constraint system instead.
[Hey Max,
> there seems to be another skew here between the .as and .js
> implementations. If the height is set to null, the view is supposed to
> take on the size of it's resource. That seems to be in the .as sprite,
> but not the .js.]
That's because view size measurement is handled by LzView.lzs - see
LzView.setHeight()... At some point we need to remove the dead code
from the .as versions.
> 2) As with .as, I don't think there needs to be any sprite
> getAttributeRelative. The portable code already maintains accurate view
> dimensions. [Max: please verify.]
Right, except getAttributeRelative() works with height and width in .as,
but not .js. This shouldn't be a problem, since width and height can't
be transformed/scaled for subviews in .js. The API needs to be
consistent and should return something sensible for width/height though...
> 3) Please put a Debug.error in sendBehind/InFrontOf, noting that the
> interface is not yet implemented, for now. Something like:
>
> Debug.error('%s: not yet implemented', arguments.callee);
>
> would suffice.
>
> Otherwise, approved.
>
> On 2006-10-09, at 14:08 EDT, Philip Romanik wrote:
>
>> Sorry about that. Changeset attached.
>>
>>
>>
>>> Seems you forgot to attach the changeset?
>>>
>>> On 2006-10-09, at 11:28 EDT, Philip Romanik wrote:
>>>
>>> > Change change.IRUbJ6604.txt by [EMAIL PROTECTED] /cygdrive/f/laszlo/
>>> > svn/src/svn/openlaszlo/branches/legals/WEB-INF/lps/lfc/ on
>>> > 2006-10-09 11:20:29 EDT
>>> >
>>> > Summary: Move common implementation of getAttributeRelative() from
>>> > LzSprite to LzView
>>> >
>>> > New Features:
>>> >
>>> > Bugs Fixed:
>>> >
>>> > Technical Reviewer: ptw
>>> > QA Reviewer: (pending)
>>> > Doc Reviewer: (pending)
>>> >
>>> > Documentation:
>>> >
>>> > Release Notes:
>>> >
>>> > Details:
>>> > The implementation of getAttributeRelative() for dhtml was missing
>>> > width and height. Since most of the swf implementation of
>>> > getAttributeRelative() is generic, I moved it back to LzView. The
>>> > small piece of runtime specific information is still kept in
>>> > LzSprite.getAttributeRelative().
>>> >
>>> > I also included a few small changes to LzSprite.js. setWidth() and
>>> > setHeight() ignore NaN (just like they do for null or negative
>>> > numbers). Also, I added stub functions for sendBehind() and
>>> > sendInFrontOf() since these were missing from dhtml and IE6 is
>>> > complaining about them.
>>> >
>>> >
>>> > Tests:
>>> > lzpix, calendar running in dhtml, swf.
>>> >
>>> > Files:
>>> > M kernel\swf\LzSprite.as
>>> > M kernel\dhtml\LzSprite.js
>>> > M views\LaszloView.lzs
>>> >
>>> <patch.Philip.5044.tgz>
>
_______________________________________________
Laszlo-dev mailing list
[email protected]
http://www.openlaszlo.org/mailman/listinfo/laszlo-dev
