Re: the state of the concurrency branch

[Tom Tromey]

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

>> Theoretically, it's already true on trunk that timers can run with
>> different dynamic bindings each time it runs. sit-for is called from
>> various places, timers can run within timers, and all can dynamically
>> bind variables.

Stefan> In theory you could imagine a scheme where timers inherit the dynamic
Stefan> bindings that are in place when the timer is setup (or when the process
Stefan> is started).
Stefan> It wouldn't provide the exact same behavior we have currently, butit
Stefan> might be made close enough to give the right behavior in practice in
Stefan> most cases.

I had actually implemented this for new threads, but you had me take it
out again :)

I think this would be too magical to do just for timers.
For one thing, I think a typical reason to do this kind of funny
business is to communicate from the timer (or process filter or
whatever) back to the caller.  So the scheme would have to be more
complicated.

Tom