>>>>> "DS" == Dan Sugalski <[EMAIL PROTECTED]> writes:
DS> I don't think Larry's weighed in on this one way or another lately.
DS> My bet is thread-shared data will get a shared attribute:
DS> my $foo : shared;
DS> possibly with some sort of dynamic sharing allowed. (i.e. you stuff a ref
DS> to an unshared scalar into a shared array so the unshared scalar becomes
DS> shared)
I'm not so sure. If we are not going to do automatic locks, promoting
variables to shared behind the coders back is promoting deadlocks.
Mr. Chairman,
I haven't seen any concrete discussion of how to proceed at this point.
I've seen mention of Configure, threading, modular structure, etc.
Is it time for a detail design?
Should the sub-lists be started? Or are there preliminary steps decisions
still to be taken?
Here are some slices. Though I'm not sure if they can be started in
parallel since there is some feedback between them.
API - external and internal
Representation - SVs and family, how will they look internally
how opaque are they internally, etc.
OpTree - Representation and implementation
Threading (crosses all the lines and should wait until these have become
slightly gelled).
<chaim>
--
Chaim Frenkel Nonlinear Knowledge, Inc.
[EMAIL PROTECTED] +1-718-236-0183