[racket-dev] [plt] Push #25575: master branch updated

From: Robby Findler (robby at eecs.northwestern.edu)
Date: Wed Oct 31 10:13:57 EDT 2012

Maybe in a while (lets see if my change works out).


On Wed, Oct 31, 2012 at 9:10 AM, Matthias Felleisen
<matthias at ccs.neu.edu> wrote:
> Is it appropriate then to deprecate the library and
> to move the code eventually into racket/control-examples?
> On Oct 31, 2012, at 10:08 AM, Robby Findler wrote:
>> On Wed, Oct 31, 2012 at 8:53 AM, Matthias Felleisen
>> <matthias at ccs.neu.edu> wrote:
>>> At some point someone (Asumu?) mentioned that this is the only
>>> use of coroutine in our code base. Is this correct?
>> grep suggests that there are no more (none on planet either).
>> FWIW, this is still "morally" a coroutine in the sense that it starts
>> some work and pauses in the middle of the work, either resuming the
>> work later or aborting the work (depending on what happened during the
>> pause). And it is cooperative, pausing the work only all spots where
>> it is safe to do so.
>> The reason I went away from the racket/engine library is that I wanted
>> to have the work happen on a specific thread (the eventspace handler
>> thread for the drracket eventspace) and racket/engine doesn't do that.
>> I considered using continuations but the work is in a simple tail
>> recursive loop (well, except for a for-each outside that) so I just
>> rewrote it directly.
>> Robby

Posted on the dev mailing list.