Just a heads up, that's a somewhat major change, in that it changes the
compile-time datasets data from being generated as Flash byte codes inline
(via the old DataCompiler) to passing a long string in which contains the
XML that is then parsed by Flash at runtime and converted to a dataset.
There
was a change by Tucker which was backported from legals to let the script
compiler deal with strings longer than 64K, that needs to be in also if any
compile time datasets have more than 64K of XML data.
But the change is needed if they are using binary libraries which have top
level datasets in them.
I don't think there should be any issues with compatibility after this
change, but make sure to QA carefully after this since
it does swap in a whole different code path under the covers.
On 3/30/07, Jim Grandy <[EMAIL PROTECTED]> wrote:
There's a conflict merging the -r 3794:3795 change
to WEB-INF/lps/lfc/glue/LaszloInitiator.as into branches/3.4. It looks like
you may also want r3754?
Author: hqm
Date: 2007-02-08 14:15:50 -0800 (Thu, 08 Feb 2007)
New Revision: 3754
Modified:
openlaszlo/trunk/WEB-INF/lps/lfc/glue/LaszloInitiator.as
openlaszlo/trunk/WEB-INF/lps/server/src/org/openlaszlo/compiler/CompilationEnvironment.java
openlaszlo/trunk/WEB-INF/lps/server/src/org/openlaszlo/compiler/DataCompiler.java
openlaszlo/trunk/lps/components/lzunit/lzunit.lzx
Log:
Change 20070208-hqm-2 by [EMAIL PROTECTED] on 2007-02-08 15:21:36 EST
in /cygdrive/c/users/hqm/openlaszlo/trunk
Summary: fixes for inline datasets with binary libraries
New Features:
Bugs Fixed: declared datasets were diappearing from library files
Technical Reviewer: pablo (pending)
QA Reviewer: (pending)
Doc Reviewer: (pending)
jim
On Mar 30, 2007, at 12:59 PM, Jim Grandy wrote:
OK, we'll take it.
On Mar 30, 2007, at 12:57 PM, Henry Minsky wrote:
I agree.
On 3/30/07, Adam Wolff <[EMAIL PROTECTED]> wrote:
>
> r3795 and r3797 which together fix LPP-3537
>
> This is the fix for the endless loop that can happen with
> datasets on the canvas in LzNode.prototype.makeChild, only in
> debug mode.
>
> Risks are low, and this is stopping our internal QA at the
> moment.
>
> Thanks,
> Adam
>
--
Henry Minsky
Software Architect
[EMAIL PROTECTED]
--
Henry Minsky
Software Architect
[EMAIL PROTECTED]
