approved! Do you think there may also be some regression in the performance of the swf8 data copy routine? The bug report indicated that large dataset loads which worked in LPS 3.3 are timing out in swf8 in LPS 4.
On Sat, Jul 25, 2009 at 11:58 AM, André Bargull <[email protected]>wrote: > Change 20090725-bargull-Goy by barg...@dell--p4--2-53 on 2009-07-25 > 17:28:53 > in /home/Admin/src/svn/openlaszlo/trunk > for http://svn.openlaszlo.org/openlaszlo/trunk > > Summary: improve performance for XMLTranslator in swf9 > > New Features: > > Bugs Fixed: LPP-8350 (Large datasets don't load in OL4.4 SWF9/SWF8) > > Technical Reviewer: hqm > QA Reviewer: (pending) > Doc Reviewer: (pending) > > Documentation: > > Release Notes: > > Details: > Using the testcase attached at the bug-report for LPP-8350, I've found out > that Flash's E4X implementation has got major performance issues when using > XML#children() and XML#childIndex(). So the nextSibling() function defined > in LzXMLTranslator made the translation process from native-xml to lzx-xml > really slow. To workaround these performance issues, the node, its > index-position and the children are stored for each node while traversing > the dom-tree. This avoids calling the E4X functions multiple times and gives > significant performance improvements (e.g. load-time for "testfile6000.xml" > from the bugreport was about 50 seconds, with the patch applied the > load-time went down to 5 seconds). > > > Tests: > see bugreport > alldata and smokecheck passes in swf9 > > Files: > M WEB-INF/lps/lfc/kernel/swf9/LzXMLTranslator.as > > Changeset: > http://svn.openlaszlo.org/openlaszlo/patches/20090725-bargull-Goy.tar > > -- Henry Minsky Software Architect [email protected]
_______________________________________________ Laszlo-reviews mailing list [email protected] http://www.openlaszlo.org/mailman/listinfo/laszlo-reviews
