You correctly summarized the problem.  There a lots of places in the  
LFC where we were sloppy about our Javascript because the Flash  
player is very lenient.  Phil R has been working in this area, so you  
should coordinate with him.  He has a 'bag of tricks' on how to  
rewrite code to not provoke these errors while still maintaining the  
original semantics.

One approach is to turn on warnings in the LFC (buildlfcdebug -- 
debug, or put #pragma "warnUndefinedReferences" in the file you want  
to focus on), and then exercise the api's in swf.  The swf debugger  
will then print warnings for all the case you will need to fix for  
dhtml.

Careful inspection of the code is another approach.  Theoretically  
better, but more difficult.

Finally, I have on my list improving the DHTML back end so that it  
can catch these kinds of errors, print a warning as the swf back end  
does, and then proceed on.

On 2006-09-23, at 19:08 EDT, Benjamin Shine wrote:

> In porting calendar to legals, I am running into
> «TypeError#11| TypeError: this.p has no properties»
> (Nope, I don't have a simple test case yet, and my version of
> calendar, which demonstrates this problem, isn't checked in yet.)
>
> This is the same problem this thread is talking about, right? As Adam
> explained it to me, Flash doesn't mind saying
> foo.bar.something
> when foo.bar is null, but DHTML does not permit that.
>
> Adam suggested I dig in, write unit tests and track down all
> instances of not checking for null in LzDatapointer. I'm up for it!
> Has anyone else started on this?
>
> On Sep 19, 2006, at 6:08 AM, P T Withington wrote:
>
>> There seem to already be plenty of places in LzDatapointer where p
>> can be null, so I think really the code just has to be gone over and
>> bullet-proofed.
>
>
>
> _______________________________________________
> Laszlo-dev mailing list
> [email protected]
> http://www.openlaszlo.org/mailman/listinfo/laszlo-dev



_______________________________________________
Laszlo-dev mailing list
[email protected]
http://www.openlaszlo.org/mailman/listinfo/laszlo-dev

Reply via email to