[gclist] some questions about Generational GCs.

[STARYNKEVITCH Basile]

Hello all,

I'm coding a precise (mostly) copying genrational GC for an internal
C++ application. (This means that I can define common C++ classes for
all GC-ed objects and require that all -local or global- roots be
explicited thru ad hoc C++ classes).

I have some questions to share with you - I do have some ideas about
them:

1) how to implement finalized objects: Some objects need to be
   explicitly destroyed (ie have their C++ destructor called) by the
   GC. I'm thinking about using special "generations" for them, and
   explicitly maintaining a collection grouping them.

2) In a major GC cycle, can I condemn some (eg half of) the old
   generations. I believe that yes (with perhaps sometimes potential
   leak with cycles of old objects...).

3) an unrelated question - just out of curiosity, does Java requires a
   conservative GC? By reading the Java VM spec, it seems to me that
   yes (the JavaVM GC can't know what words on the Java stack are
   references?) but I am not sure of it.

Regards.

N.B. Any opinions expressed here are only mine, and not of my organization.
N.B. Les opinions exprimees ici me sont personnelles et n engagent pas le CEA.

---------------------------------------------------------------------
Basile STARYNKEVITCH   ----  Commissariat à l Energie Atomique 
DTA/LETI/DEIN/SLA * CEA/Saclay b.528 (p111f) * 91191 GIF/YVETTE CEDEX * France
phone: 1,69.08.60.55; fax: 1.69.08.83.95 home: 1,46.65.45.53
email: Basile point Starynkevitch at cea point fr