>> The problem is (and Andrew mentioned this already on the traits  
>> list), that the following singletons are not initialized:
>>
>> RequiredSelectors initialize
>> LocalSends initialize
>> ProvidedSelectors initialize
>
> Did anyone come up with an explanation for why these classes were  
> not being automatically initialized when they were loaded.  A  
> Monticello bug?  A bootstrap bug?

it is MC's behavior: after loading, MC goes over the added method  
definitions and evaluates those that are defined on the class side  
and have selector #initialize. In the case of the above three classes  
this is a problem because they don't implement the method themselves  
but inherit it from the superclass. Hence, MC only once evaluates the  
initialize method but with the receiver of the abstract superclass,  
having no effect.
I'm not sure what's the "right" behavior, however. Was this handled  
differently by changesets?

>
> And can someone tell the the process for submitting bug fixes and  
> changes?

There are two repositories both on SqueakSource where we commit all  
changes to. This is:

- http://www.squeaksource.com/Traits.html for the core traits code.  
Changes in packages like Traits, Kernel, Monticello are committed there.
- http://www.squeaksource.com/TraitsOmniBrowser.html is the project  
Daniel created for the new browsers based on OB.

For changes for the first repository, please just commit and maybe  
send me a short message about it (I should normally also see it via  
RSS). I will then collect the changes and coordinate with the  
harvesting team.

Cheers,
Adrian