[gclist] Distributed Garbage Collection Revisited 1

[H.C.C.Rodrigues]

Matthew,

>>                   ___________________
>>                  /                  |
>>                  |                  |
>>                  V                  |
>>          R1    ScionA1              |
>>          |       |                  |
>>          V       V                  |
>>        ObjA2<--ObjA1                |
>>                  |                  |
>>                  V                  |
>>                StubB1               |
>>                  |       Space A    |                    
>>                  |      ----------  |
>>                  V       Space B    |
>>          R2    ScionB1              |
>>          |       |                  |
>>          V       V                  |
>>        ObjB2<--ObjB1                |
>>          |       |                  |
>>          |       V                  |
>>          \---->StubA1               |
>>                  |                  |
>>                  \-------------------
>
>there is no point in time at which the trace would miss both R1 and
>R2, thereby negligently collecting the entire cycle.  I've laid this
>out in fairly excruciating detail in previous messages, so I won't do
>so again here.

Thank you for your message.
I was not thinking that the cycle could be negligently colleted. I was
thinking (from your messages, not from your paper) that would be a barrier in
ObjB1 (like Gustavo and Giuliano) that would start a back search. From what
you said now, I understand taht ObjB1 will answer 'live' because it was
acessed.

Helena.