Re: [Gcl-devel] Re: Server sockets with GCL?

[Dennis Decker Jensen]

Chris Hall writes:
        
        > SmallTalk module for fastcgi is also very interesting in terms
        > of fast and simple web servers. There is a white paper.
        >
        > http://mod.smalltalk.org
        >
        
        I've never used Smalltalk, but I did find this interesting:
        
        http://www.beta4.com/seaside2/
        
        Scheme-style continuations for session logic/flow control
        - I'm not sure what happens when someone presses the
        'back' button in their browser, though - I wonder if
        continuations can go backwards? ;-)

Reapplying old continuations?  AFAICT you can configure what
to do in those situations depending on the application in
context.  I still don't know how they handle old continuations.
>From what I can tell it is some sort of LRU (Least Recently Used)
algorithm parameterized for specific demands.

Chris Double has some very good tutorial docs from an example
implementation in Chicken Scheme.

http://radio.weblogs.com/0102385/

I've also kept an eye on seaside for some while now.  I wonder if
is possible to capture the runtime stack in GCL?  That is what
UncommonWeb does anyway.  I guess it cannot be done considering
GCL compiles to C being oblivious of such things. However, it
is possible to implement continuations in terms of threads,
which is what the Ruby language does, where the framework
is called Borges.  If GCL gains threads it shouldn't be too
difficult to do in pure lisp! He! ;-)

There is a blog on Seaside.

http://blogs.inextenso.com/seaside/blog/learning/rss/recent.xml|||

I've never used SmallTalk either but know it enough to read it
(Kent Beck's SmallTalk Best Practice Patterns is a good book
even though you aren't programming ST) -- the syntax is as
simple as Lisp -- Ruby I know from a little experience.
        
        
        Was the ACE article one of these, by any chance?
        
        http://www.cs.wustl.edu/~schmidt/patterns-ace.html

Thanks for the reference!

I checked the site but couldn't find it.  All I can remember is
that is was an experience report with some controversial conclusions
with regard to IO, async/sync, forking/threading, etc.

O well, I have even forgotten where I read about it...
        
        Nothing to it, right? ;-)

He!

Ciao,

Dennis Decker Jensen

"The Board views the endemic use of PowerPoint
briefing slides instead of technical papers as an
illustration of the problematic methods of technical
communication at NASA."
 -- Official report on the Columbia shuttle disaster.