No, you are absolutely right. I didn't originate the diff, but I
should have looked more closely before forwarding on (sorry --
juggling too much today). The original version (with the temporary f)
is no different -- there's nothing between the assignment of f and
its use.
So we're back to reverting the original change and losing the dynamic
behavior, unless someone would like to volunteer to put together a
more comprehensive fix. It doesn't seem worth it to keep the original
coordinate value, however, so perhaps we should just revert the change.
I do agree that the update() call in the original change was a good
addition, so perhaps we should leave that in.
jim
On May 30, 2006, at 1:53 PM, Philip Romanik wrote:
> I don't see how this helps any. The first time the layout is run
> (assumming on x), the x coordinate of the views are changed. When
> the layout axis is changed to y, the proposed change is fetching
> the x coordinate from the view and setting it to the same value.
> Isn't this statement,
>
> s.setAttribute(o, s[o])
>
> just a nop (or am I missing something)?
>
> Phil
>
>
>
>> Bruce has added a different proposed change to LPP-2091. Would you
>> and Max mind taking a look and sending your comments to the list?
>>
>> But I think there's an improvement to be made to that diff.
>> Instead of
>>
>> var f;
>> if (o=="y") {f=s.y;}
>> else {f=s.x;}
>>
>> s.setAttribute( o , f);
>>
>> I think you could just do
>>
>> s.setAttribute(o, s[o]);
>>
>> jim
>>
>> On May 30, 2006, at 12:08 PM, Philip Romanik wrote:
>>
>> > The change has been backed out, but the behavior is still broken.
>> > In a static situation, it works correctly. Running this,
>> >
>> > <canvas height="100">
>> > <view >
>> > <simplelayout axis="x" spacing="10"/>
>> > <view bgcolor="blue" y="10" height="30" width="50"/>
>> > <view bgcolor="blue" y="20" height="30" width="50"/>
>> > <view bgcolor="blue" y="30" height="30" width="50"/>
>> > </view>
>> > </canvas>
>> >
>> > works as expected. However, when you dynamically change the axis,
>> > the old value for coordinate is not reset to its previous value. I
>> > now understand why the coordinate of the other axis was set to 0.
>> > In order to fix this properly the object needs to keep track of the
>> > old coordinates.
>> >
>> > This example shows how the layout is broken:
>> >
>> > <canvas height="600" debug="true">
>> > <view>
>> > <method event="onclick">
>> > sl.axis = (sl.axis == "x") ? "y" : "x";
>> > sl.update();
>> > </method>
>> > <simplelayout id="sl" axis="x" spacing="10"/>
>> > <view id="v1" bgcolor="blue" height="30" width="50"/>
>> > <view id="v2" bgcolor="blue" height="30" width="50"/>
>> > <view id="v3" bgcolor="blue" height="30" width="50"/>
>> > </view>
>> > </canvas>
>> >
>> > When you click on the view, the axis changes from x to y. The x
>> > coordinates maintain the value they had when simplelayout first
>> ran.
>> >
>> > Also, the call to update() was removed in the setAxis method (as
>> > part of backing out the change). The user has to manually call
>> > update() in order to see the changes.
>> >
>> >
>> > Phil
>> >
>> >
>> >> Approved, pending fixes to line endings in layouttest.lzx...
>> Sorry I
>> >> missed this the first time!
>> >>
>> >> --
>> >> Regards,
>> >> Max Carlson
>> >> OpenLaszlo.org
>> >>
>> >> Jim Grandy wrote:
>> >> > Change 42573 by [EMAIL PROTECTED] on 2006/05/29 20:51:18
>> >> *pending*
>> >> >
>> >> > Summary:Back out most of change 42068, since it causes
>> >> > simplelayout
>> >> > misbehavior
>> >> >
>> >> > Bugs Fixed:LPP-2091
>> >> >
>> >> > Technical Reviewer: max
>> >> > QA Reviewer: promanik
>> >> > Doc Reviewer:
>> >> >
>> >> > Release Notes:
>> >> >
>> >> > Tests:
>> >> >
>> >> > Affected files ...
>> >> >
>> >> > ... //depot/lps-3.3/lps/components/utils/layouts/
>> simplelayout.lzx#1
>> >> > edit ... //depot/lps-3.3/test/components/layouttest.lzx#1 edit
>> >> >
>> >>
>> >
>
_______________________________________________
Laszlo-dev mailing list
[email protected]
http://www.openlaszlo.org/mailman/listinfo/laszlo-dev
