[Laszlo-dev] For Review: Change 20090305-hqm-o Summary: make onerror event work again for loading a bad resource

[André Bargull]

LzSprite.as (swf9) shouldn't part of this change (there is only a 
debug-statement added).
Otherwise approved!

May you file some improvement requests / bugs:
- to remove the debug-error in "LzLoader#returnData(..)" for 
load-failures (there is no debug-error in swf9/dhtml; debug-error should 
be restricted to real LFC errors and not used to inform about load 
failures of user-content)
- ResponderMEDIA.respondImpl() shouldn't call respondWithExceptionSWF() 
for "normal" http-errors (eg. 404), instead respondWithErrorSWF() should 
be used
  (It's not really useful to include a Java stacktrace in the response 
swf-file, is it?)
- "lz.Idle.onidle.delegateList" still lists the loadCheckerDel-delegate 
(see kernel/swf/LzMediaLoader.lzs) after the onerror-event was 
dispatched (swf8, proxied) => memory leak!
- and we need to decide on (not) sending onloadratio/onframeloadratio 
for onerror:
  o events are send: swf8 (proxied), dhtml
  o events aren't send: swf8 (unproxied), swf9

On 3/5/2009 6:15 PM, Henry Minsky wrote:
> Change 20090305-hqm-o by hqm at badtzmaru.home on 2009-03-05 12:05:57 EST
>     in /Users/hqm/openlaszlo/trunk6
>     for http://svn.openlaszlo.org/openlaszlo/trunk
>
> Summary: make onerror event work again for loading a bad resource
>
> New Features:
>
> Bugs Fixed: LPP-7494
>
> Technical Reviewer: andre
> QA Reviewer: philr
> Doc Reviewer: (pending)
>
> Documentation:
>
> Release Notes:
>
> Details:
>
> + the old swf8 datacompiler is still needed to compile "error swfs" 
> which are sent in lieu of the requested
> resource,
>
> -- for swf8, the datacompiler support had rotted, uncomment the _m, 
> _t, globals shortcuts in LzDatapointer.
>
> -- for swf9, we need to make Responder.java do something useful if the 
> swf runtime > 8; I believe sending
> a 404 may be the best we can do, that will signal an IOErrorEvent in 
> swf9's Loader.
>
> Tests:
>
>  http://localhost:8080/trunk6/test/lfc/legals/multiframeresource.lzx
>
> look at the "should fail" test case on the right
>
> SWF8
> PROXIED  gets onerror
> SOLO     onerror after five seconds ; canvas.mediaerrortimeout = 4.5secs
>          In SOLO don't know if we can detect the failure to load any 
> faster than via timeout
>
> SWF9
> PROXIED  gets onerror
> SOLO     gets onerror
>
> DHTML
> SOLO     onerror, but also onloadratio:1 and onframesloadratio:1
> PROXIED  onerror, but also onloadratio:1 and onframesloadratio:1
>
>
> Files:
> M      WEB-INF/lps/lfc/kernel/swf9/LzSprite.as
> M      WEB-INF/lps/lfc/services/LzBrowser.lzs
> M      WEB-INF/lps/lfc/data/LzDatapointer.lzs
> M      WEB-INF/lps/server/src/org/openlaszlo/server/LPS.java
> M      
> WEB-INF/lps/server/src/org/openlaszlo/servlets/responders/Responder.java
>
> Changeset: 
> http://svn.openlaszlo.org/openlaszlo/patches/20090305-hqm-o.tar
>