[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Axiom-developer] openpty patch again
From: |
Gabriel Dos Reis |
Subject: |
Re: [Axiom-developer] openpty patch again |
Date: |
02 Dec 2006 04:29:07 +0100 |
Waldek Hebisch <address@hidden> writes:
| Gabriel Dos Reis wrote:
| > Waldek Hebisch <address@hidden> writes:
| >
| > | Normally when viewman dies sman re-spawns a new copy...
| >
| > This is a design flaw in sman and friends -- I came across it yesterday
| > and thought "wait, this is an unbounded replication; nobody wants
| > that!".
| >
|
| This behaviour is semi-reasonble: Axiom creators treated Axiom and
| supporting programs as an Internet server where clients come, request
| some computation and go away while the core parts (AXIOMsys, sman,
| session, viewman, nagman) run for long time.
Note however the specific beahviour we are talking about is an
*uncontrolled* process-spawning. A server will have a mechanism to
decide/detect when a death is requested as opposed to accidental.
| Anyway, if you look for design flaws, you will probably find many.
Well, I was not specifically looking for design flaws :-)
| One example that I found few days ago: 'AXIOMsys' at various moments
| sends 'EndOfOutput' command to 'session' and then go on to process
| information sent by 'session'. But session also sends commands
| to 'AXIOMsys' and for some commands expects that 'AXIOMsys' will
| return some data as an answer. There were no way for 'session'
| to distinguish returned data from 'EndOfOutput' command.
:-(
-- Gaby