Re: Patch for fields of `struct buffer'

[Tom Tromey]

>>>>> "Stefan" == Stefan Monnier <address@hidden> writes:

Tom> I will probably keep the current approach (the above shows "1"), since
Tom> the hair is all isolated in the thread-creation function, and since this
Tom> is how I expected it to work in the first place.

Stefan> BTW, I think the option 1 of sharing between threads might lead to
Stefan> complications in some implementations, since we now have variables which
Stefan> are neither global nor specific to a single thread but are shared by
Stefan> some subset of all threads.

Aha.  I was unclear here.

In my implementation, variables let-bound in the parent thread are just
re-let-bound in the new thread.  The bindings are not shared.

I hadn't even considered this idea.

Stefan> We can provide ways to transfer the value (but just the value) of some
Stefan> variables from a parent to his child.

Perhaps copying CL is the way to go, then.

Tom