[plt-scheme] javascript s-exp transform

Jay,

Given these previous discussions ...

>> my ideal continuation style web app would define two co-routines, a
>> scheme
>> routine, and a javascript routine, and pass the continuation back
and
>> forth, with no continuations actually stored on the server. it would
>> also allow for the javascript routine to branch to different
>> continuations (or pick another partner for its coroutine dance),
>> similar to what send/suspend/dispatch allows you to do. 

>> kanishka
.............

On 5/24/07, Shriram Krishnamurthi <sk at ...> wrote:
> I'm afraid there is no "simple method" other than what is described
in
> that paper.  If, however, you want to try a different approach that
> perhaps feels a bit less complicated, you could consider our other
> approach to doing this, which does not use CPS at all:
>
>
http://www.cs.brown.edu/~sk/Publications/Papers/Published/pcmkf-cont-from-gen-stack-insp/

BTW, This is implemented in the `prototype-web-server' in the SVN
graveyard. Also, it will be re-integrated into the mainline server in
SVN in the coming weeks.

Jay

.................




Would you be willing to implement an experimental version of either cps
or continuation-mark version of automated program transformation on top
of the javascript s-exp syntax (from planet)? Possibly add it as helper
library to the plt server. 

example input:

(define jscode
  (js 
    (TRANSFORM 
     (define (precpsfunc action)
       (if (form-updated? action)
           (define result (validate action))
           (if (warning? result)
               (update-message-area! result))
               (resp (K-POINT (post-to-server action)))
...
)

if using the cps version, output could be a javscript table of
closures..

kanishka


       
____________________________________________________________________________________Looking for a deal? Find great prices on flights and hotels with Yahoo! FareChase.
http://farechase.yahoo.com/