[Pharo-project] Apparently composed trait methods loses their categories...

[Stéphane Ducasse]

Hi guys

In 20, which T1>>m1 is defined under cat1
when we do 
        T3 uses: T1 
                T3>>m1 is 'as yet unclassified' and this breaks several tests.

Any idea of a change that could have modified the system?

Stef