More weirdness. If I add a subview to a stretched view, I get different results on how that subview is displayed depending on whether a mouse event is defined or not. Can anyone explain what's happening here?

<canvas debug="true">

    <simplelayout axis="x" spacing="10"/>

    <!-- Mouse over event -->
    <view height="400" width="100" stretches="height"
          bgcolor="silver"
          onmouseover="Debug.write('mouseover')">
        <view height="1" width="100" bgcolor="green"/>
    </view>

    <view height="400" width="100" stretches="height"
          bgcolor="silver"
          onmouseover="Debug.write('mouseover')">
        <view height="2" width="100" bgcolor="red"/>
    </view>

    <!-- No mouse over event -->
    <view height="400" width="100" stretches="height"
          bgcolor="silver">
        <view height="1" width="100" bgcolor="green"/>
    </view>

    <view height="400" width="100" stretches="height"
          bgcolor="silver">
        <view height="2" width="100" bgcolor="red"/>
    </view>

</canvas>

Probably related to LPP-3726, LPP-4661, LPP-4684

Hm, other weird things happen when stretching a resource. I'll add another test case when I can repro.

OK, this is the bug that I'm really trying to fix. It appears that if you subsequently change the height of a stretched view with a resource from 0, the view will not accept any mouse events. Test case:

<canvas debug="true">

    <view name="myview" height="0" stretches="height"
          resource="v_scroll_track.png"
          bgcolor="green"
          onclick="Debug.write('click', this)"
          onmouseover="Debug.write('mouseover', this)"
          onmouseout="Debug.write('mouseout', this)"/>


    <!-- After setting height to 200, the view gets no mouse events. -->
    <button x="100" text="set height to 200"
            onclick="myview.setHeight(200)" />

</canvas>

r7933 | ptw | 2008-01-31 07:25:53 -0500 (Thu, 31 Jan 2008) | 34 lines
Changed paths:
   M /openlaszlo/trunk/WEB-INF/lps/lfc/debugger/platform/swf/LzDebug.as
   M /openlaszlo/trunk/WEB-INF/lps/lfc/kernel/swf/LzMakeLoadSprite.as
   M /openlaszlo/trunk/WEB-INF/lps/lfc/kernel/swf/LzSprite.as

Change 20080125-ptw-x by ptw@dueling-banjos.local on 2008-01-25 15:41:02 EST
    in /Users/ptw/OpenLaszlo/ringding-2
    for http://svn.openlaszlo.org/openlaszlo/trunk

Summary: Yet another try at solving stretches vs. clickable

Bugs Fixed:
LPP-5372 'Mouse events available for first 100 pixels on a stretched view'

Technical Reviewer: hqm (message://<8c61fad60801301500w262fa19aq92cfcb6bd08011c@mail.gmail.com>)
QA Reviewer: pkang (pending)

Details:
    LzMakeLoadSprite: DON'T call updateResourceSize (which apparently
    papered over bugs in LzSprite, and now does the wrong thing).
    Don't inline parts of the code from LzSprite#set{Width,Height},
    which leads to skew; call them.

    LzSprite: Add an informative debugging name. Added a bunch of
    devnotes to document the secret inner workings of the Flash player
    that Adam apparently thougth were patently obvious. Initialize
    resource{width,height}. Make sure you have a clip to scale before
    you try to scale it. Correct computation of scale for stretches
    when there is no resource supplied. Correct the computation for
    scaling of button (clickable) to compensate for the parent clip
    being stretched. Simplify computation of updateResourceSize.

    LzDebug: Make the movieclip debug printer more informative. Print
    _width and _height of movieclip when inpsecting.

Tests:
    Test cases from LPP-5372, 4661 (original and one in comments), 3726

(trunk 4 local build r7937)

1st example in bug:
Runs successfully in swf and dhtml.

2nd example in bug:
Runs successfully in swf and dhtml.
WARNING: Due to limitations in the DHTML runtime, stretches will only apply to resources in this view, and doesn't affect child views.

3rd example in bug:
Runs successfully in swf and dhtml.