[Laszlo-dev] For Review: Change 20080225-hqm-q Summary: add data to SWF SOLO POST requests, to ensure that Flash does not turn them into GET requests

Henry Minsky henry.minsky at gmail.com
Mon Feb 25 10:10:12 PST 2008


Change 20080225-hqm-q by hqm at badtzmaru.local on 2008-02-25 12:53:53 EST
     in /Users/hqm/openlaszlo/trunk/WEB-INF/lps/lfc
     for http://svn.openlaszlo.org/openlaszlo/trunk/WEB-INF/lps/lfc

Summary: add data to SWF SOLO POST requests, to ensure that Flash does  
not turn them into GET requests

New Features:

Bugs Fixed: LPP-5368

Technical Reviewer: andre
QA Reviewer: max
Doc Reviewer: (pending)

Documentation:

Release Notes:

Details:

SWF has a bug whereby an empty POST request (one with no data in it's  
post body) is turned into a
  GET request. This behavior used to be masked by the way we forced  
all query args into the POST
body for post requests.

When a change was made to  separate  and preserve the query string  
from the post body, people
started reporting bugs where their POST requests became GET requests.

This patch forces the "__lzbc__" cache-breaking timestamp arg into the  
body of SOLO post requests.

I am wondering if I should also force this value into Proxied  
requests, to ensure that
the data which is sent to the back-end is consistent between SOLO and  
proxied requests. This
patch only forces the __lzbc__ arg into SOLO POST  requests.


Tests:

test.lzx:
<canvas debug="true" height="400" >
     <dataset name="dsGroups" request="true" type="http"  
src="test.jsp" querytype="POST"/>
     <datapointer name="dpGroups" xpath="dsGroups:/Response">
         <handler name="ondata" args="d"><![CDATA[
             Debug.write(d);
         ]]></handler>
     </datapointer>
</canvas>

test.jsp:
<%System.out.println(request.getMethod());%>
<Response>
	<Method><%=request.getMethod()%></Method>
</Response>

also, test/lfc/data/alldata.lzx
amazon.lzx
calendar.lzx

Files:
M      data/LzHTTPDataProvider.lzs

Changeset: http://svn.openlaszlo.org/openlaszlo/patches/20080225-hqm-q.tar


More information about the Laszlo-dev mailing list