[Laszlo-dev] Looking at 'No eventSender' error

[P T Withington]

We're missing.

If you make your change, then _any_ dependency that evaluates to null  
will no longer signal an error.  So I was trying to determine if  
these errors only come from conditional dependencies, or if there are  
other cases.  There is no way that I can think of to decide that,  
since we can't differentiate them in applyConstraints, and we can't  
test all possible constraints.

But, then again, I argued that this whole warning was bogus, since  
the system has always worked before without it.  So, I guess we just  
put in your change.

On 2006-11-27, at 15:33 EST, Philip Romanik wrote:

> Hi Tucker,
>
> I propose submitting the change and then creating a new jira bug to  
> track is this problem continues to occur. If it does not occur, it  
> would make sense to update the dependency generator.
>
> Phil
>
>> My point is that if the nulls all come from conditional dependencies,
>> we could try to make the dependency generator smarter.  If there are
>> other causes of null, but we skip them all in applyconstraints as you
>> suggest, then it is basically the same as removing the warning (in
>> that we will never learn about the other causes of null).  It is
>> clearly more efficient not to make the call in the null case.
>>
>> On 2006-11-27, at 14:16 EST, Philip Romanik wrote:
>>
>> > I don't know if this fixes the problem in every case, but it
>> > addresses the problem in the cases I've checked.
>> >
>> > Phil
>> >
>> >
>> >
>> >> Nice sleuthing.  Is it the case that all these warnings are coming
>> >> from that conditional dependency in basecomponent?
>> >>
>> >> Since that is an inner loop, I'd put dep[i] in a var for  
>> efficiency.
>> >
>