Re: [PATCH 06/10] add most lisp-level features

[Stefan Monnier]

> Yet one more observation here.  The Bordeaux approach lets you specify
> which bindings to capture.

AFAIK, you can easily do that manually with the "no capture" behavior
as well.

> It seems to me that some such mechanism is necessary, at the very
> least.  Otherwise, passing a local variable into a thread seems
> difficult, unless you assume that only lexical-binding:t code will
> want to make threads.

Converting to lexical-binding is fairly easy in my experience (at least
for code whose byte-compilation doesn't spit out hundreds of warnings).
So, yes, I assume that lexical-binding will be a common case for
threaded code and I'm happy to make things a bit harder for the
dynamic-binding case if it can make the whole approach simpler
and cleaner.

The concurrency code makes dynamic binding more expensive anyway
(either because of the need to unwind&rewind them during context switch,
or because accessing those vars requires fetching the current
thread-local value), so it naturally likes lexical-binding.

> Maybe everything important really will use lexical binding.

At least the `make-thread' code will likely be in a lexical-binding
file, yes.


        Stefan