The test fails for IE6. I tried IE6/IE7/FF1.5/FF2/Opera/Safari
Otherwise, approved.
Change 20071031-maxcarlson-E by [EMAIL PROTECTED] on 2007-10-31
13:09:06 PDT
in /Users/maxcarlson/openlaszlo/wafflecone
for
<http://svn.openlaszlo.org/openlaszlo/branches/wafflecone>http://svn.openlaszlo.org/openlaszlo/branches/wafflecone
Summary: Fix lzhistory behavior when another page is visited and the
back-button is used.
New Features:
Bugs Fixed: LPP-4999 - IE resets history to #0 after visiting
another site and pressing the back button
Technical Reviewer: promanik
QA Reviewer: jcrowley
Doc Reviewer: (pending)
Documentation:
Release Notes:
Details: Applied patches based on Jes' excellent notes:
1) In embed-compressed.js, Lz.history.init was being called before
Lz._loaded. As a result, the parse function was not setting the
canvas attribute to the value that was in the URL when the page was loaded.
To solve this, I removed the onload event at the end of the page and
added the following to line 53:
dojo.flash.addLoadedListener(Lz.history.init)
2) Even after #1, something (not sure what) still depends on onload.
However, in Firefox, if you use the back button to get back to a
page, onload does not get fired (see my previous email for details).
To solve this, I defined an onunload event. If this is present, then
Firefox does not do its Javascript caching and onload will get fired
when coming back to the page. So, at the end of the file, I added:
window.onunload=function donothing() {}
Tests: See LPP-4999 for a complete testcase. Modify test.html to
use new embed-compressed.js instead of embed-fnac.js.
Files:
M lps/includes/source/embednew.js
M lps/includes/source/lzhistory.js
Changeset:
<http://svn.openlaszlo.org/openlaszlo/patches/20071031-maxcarlson-E.tar>http://svn.openlaszlo.org/openlaszlo/patches/20071031-maxcarlson-E.tar
