[Stackless] For interested - an interlacing compiler for C that allows continuations

Aaron Watters aaron at reportlab.com
Sun Sep 8 15:01:54 CEST 2002


Paul Sheer wrote:

>
> BUT it does require whole program analyses. for exmaple, any
> functions which need to be considered must be present and
> preceeding in the C file. Also, a file is considered as
> a whole program. A global variable has to be placed in the
> structure that represents the namespace for a tasklet,
> hence all global variables must be seen by the compiler---
> they must be declared within the same single source file.

I don't know about PhD theses, but this restriction might
be bad for practical software engineering purposes since it
seems to disallow things like dynamically linked libraries
or in process COM objects and the like.

I would vote for removing it if you can (and still get
the committee to sign off :) ).

How about going the other direction and emulate tasklets
in processes?

I've been beginning to wish for process/pseudo tasklets.
The bad thing about my http experiments so far is that if one
cgi script crashes the whole server dies.  It would be cool if
I could write stackless style multistage programs that reside
in separate processes and communicate using channels.  Then if
one crashes, it's more like conventional http/cgi, big deal.

What would be even cooler would be if I could take a
tasklet/process which has gone dorment and pickle it, storing
it away in a database for the time when the activity wakes up.
If you could do this with a C program in a portable manner
that would be great.

All these things are emulated using various object oriented
approaches, but they all require increasing levels of
(global) code obfuscation too, as usual...

   -- Aaron Watters

===
"Where's the any key?" -- the immortal homer (simpson)


_______________________________________________
Stackless mailing list
Stackless at www.tismer.com
http://www.tismer.com/mailman/listinfo/stackless



More information about the Stackless mailing list