Re: Patch for fields of `struct buffer'

[David De La Harpe Golden]

On 31/01/11 19:04, Tom Tromey wrote:
> > > > > > "David" == David De La Harpe Golden<address@hidden>  writes:
>
> Tom>  One easy one is whether a new thread should inherit thread-local
> Tom>  bindings from its parent thread.  Our initial implementation did
> Tom>  inherit, but later I found out that this is not common in the Lisp
> Tom>  world.
>
> David>  I'd suggest going with CL afaik-now-most-common-practice and not
> David>  inherit. Isn't that easier implementation-wise anyway?
>
> Not notably, especially since this code is already written.
> (I did notice a bug on the branch ... but whatever, it is easy to fix.)

FWIW - turns out the initial model in SBCL was inherit, but a transition 
to non-inherit was made in late 2005 at SBCL 0.9.6 [1][2].  ISTR more 
wide-ranging discussions of the issue elsewhere, but this seems to be 
the relevant sbcl-devel thread:

"[sbcl-devel] dynamic variables and threads" on 2005-10-12
http://sourceforge.net/mailarchive/forum.php?thread_name=200510121507.15694.mega%40hotpop.com&forum_name=sbcl-devel

> Or, instead of the locks-and-condition-variables approach, go the CSP
> route and have channels and channel-select.  This idea is making a
> comeback, maybe it would fit well in Emacs Lisp.

I do think that some message-passing type system is a nice high-level 
model, even in the multithreading case rather than multiprocessing.
But isn't that something one might then build on top given the lowlevel 
pthredy stuff?  I suppose you could not fully expose the lowlevel stuff 
to lisp.

[1] http://www.sbcl.org/manual/Special-Variables.html#Special-Variables
[2] http://www.sbcl.org/all-news.html#0.9.6