[Laszlo-user] swf9 method parameters mismatch causes hangs

[P T Withington]

Thanks for the bug report.  As Henry has already said in the bug, we  
have some ideas how we can improve the swf9 generated code so that  
this error can be caught at compile time.

We also have some work scheduled to improve the debugging of errors in  
swf9.  Basically, we want to try to grab the stack trace that we can  
extract from the debug player and fit it in to our debug backtrace,  
which hopefully can make the backtrace more meaningful to the LZX  
programmer.  It's an interesting problem mapping from LZX to as3 and  
then trying to map back...

On 2009-04-06, at 13:58EDT, Gregory Denton wrote:

> The forum thread is not me, but same problem.
>
> I was running flash player 9 32-bit version under firefox 64-bit and
> decided to upgrade to the flash 10 dev version but could not get that
> to work (tried "wrapping" and linking in plugin dirs in various way
> with no luck, hard to find info on net, anyone know how?). Wasted most
> of Friday on this, but at least upgraded now.
>
> So, I'm running with flash 10 64-bit version (lz SOLO and compiled
> with -Dcompiler.swf9.catcherrors=true) and that does show the .lzx
> file line numbers on hang errors except that the file/line displayed
> is up the call stack several levels, which still necessitates a
> search.
>
> Here is the bug:
> http://www.openlaszlo.org/jira/browse/LPP-8030
>
> Thanks.
>
> On Fri, Apr 3, 2009 at 5:36 AM, P T Withington <ptw at pobox.com> wrote:
>> I'm guessing the thread in the forums is you too?  I posted some  
>> suggestions
>> there:
>>
>> http://forum.openlaszlo.org/showthread.php?t=13292
>>
>> I would have hoped the Flash back end would have detected this  
>> error at
>> compile time, but apparently not?  Please file a bug with your  
>> example case.
>>  I think we could do better here...
>>
>> On 2009-04-02, at 22:18EDT, Gregory Denton wrote:
>>
>>> I'm upgrading my large app to swf9 and having a very difficult  
>>> time of
>>> it. I've done all the things in the wiki and have a clean compile.  
>>> The
>>> app hangs whenever there is a call to a method where the caller  
>>> has an
>>> extra arg that is not in the method args list. This is not mentioned
>>> in the wiki and there is no debug message printed! (It is easy,  
>>> for me
>>> at least :-(, to sometimes forget the signature for methods called  
>>> in
>>> a dynamic-like manner.) My primitive debugging technique is to put  
>>> in
>>> Debug.writes, very time consuming with the depth of calls. Is this a
>>> known problem? Any way to put out a message at the point it hangs?
>>>
>>> http://wiki.openlaszlo.org/Runtime_Differences
>>>
>>> A simple test case:
>>>
>>> <canvas debug="true">
>>>   <simplelayout axis="y"/>
>>>   <view name="foo" width="100" height="100" bgcolor="red">
>>>       <method name="meth">
>>>           Debug.write('NOT PRINTED IF swf9!!!');
>>>       </method>
>>>   </view>
>>>   <button text="foo" onclick="foo.meth(0); Debug.write('NOT PRINTED
>>> IF swf9!!!');"/>
>>> </canvas>
>>>
>>> Thanks.
>>
>>