The error is in kernel/dhtml/LzTextSprite.js, here:
if (! root) {
root = document.createElement('div');
root.setAttribute('id', 'lzTextSizeCache');
>>> root.setAttribute('style', 'top: 4000px;');
document.body.appendChild(root);
}
You probably need Max's input here, but I believe what is happening
is that setAttribute is being inlined by the compiler as if it were
LzNode.setAttribute, but this is DOM Element setAttribute. Max
worked around this in wafflecone I think by defining his own
__setAttr instead.
If I compare your LzTextSprite to the wafflecone, there are still
lots of differences. Maybe you just need to take more changes from
wafflecone to get to a point where the compiler and the LFC are back
in sync?
On 2007-09-14, at 15:27 EDT, Benjamin Shine wrote:
Odd, I just sent this patch for review, but the auto-generated
email message had the wrong subject line. This is the correct subject.
On Sep 14, 2007, at 12:14 PM, Benjamin Shine wrote:
Change 20070914-ben-Y by [EMAIL PROTECTED] on 2007-09-14 12:08:04 PDT
in /Users/ben/src/svn/openlaszlo/branches/legals
for http://svn.openlaszlo.org/openlaszlo/branches/legals
Summary: Work in progress on merge from wafflecone to legals [DO
NOT CHECK IN]
New Features:
Bugs Fixed:
Technical Reviewer: ptw, max
QA Reviewer: (pending)
Doc Reviewer: (pending)
Documentation:
This is a merge in progress of wafflecone r6200:6450 into legals.
It causes lzunit itself to fail in dhtml in firefox. We mustn't
check in with smokecheck failing so badly, so, I'm sending this
patch around for review. The goal is to make smokecheck succeed in
dhtml in firefox and safari.
