tcl-websh-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From dav...@dedasys.com (David N. Welton)
Subject Re: "cleanups": 4726
Date Wed, 12 Dec 2001 17:22:54 GMT

        Simon says, in an email:

        There is more than just the context that can by lying around in a
        mod_websh interpreter after a request (global vars, additional
        namespaces). How should we handle this situation ?

        variant a)

          try to detect the interpreters state before the request is handled,
          but after evaling the initialization. Then restore this state before
          you pass the interpreter back to the pool.

          Pro: developer doesnot have to worry, mod_websh does it right
               in any case
          Con: expensive

        variant b)

          the developer has to think

          Pro: simple, fast (;) solution
          Con: dangerous (session cross-talk)

        For now, we chose variant b). I think that we should invesrtigate
        variant a) furthermore, and time it.

I keep thinking about this one...  I think it might be interesting to
make a more global package based on this thought.  I recently made a
comp.lang.tcl posting to that effect as well.

Basically, the plan would be to have two checkpoints.  An entry
checkpoint, and an exit checkpoint.  Everything created between the
two gets deleted when the exit checkpoint is hit.  Naturally, some
commands to save certain things would be necessary, as well.  Maybe
also configuration options for things such as whether to destroy
variables and/or channels...

Thoughts?
-- 
David N. Welton
   Consulting: http://www.dedasys.com/
Free Software: http://people.debian.org/~davidw/
   Apache Tcl: http://tcl.apache.org/
     Personal: http://www.efn.org/~davidw/

Mime
View raw message