From mboxrd@z Thu Jan  1 00:00:00 1970
From: Einar Karttunen <ekarttun@cs.helsinki.fi>
To: Plan9 ML <9fans@cse.psu.edu>
Message-ID: <20040310111719.GA9327@melkki.cs.Helsinki.FI>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
User-Agent: Mutt/1.4.1i
Subject: [9fans] tcs and many channels
Date: Wed, 10 Mar 2004 13:17:19 +0200
Topicbox-Message-UUID: 2861177a-eacd-11e9-9e20-41e7f4b1d025

Hello

I am hacking irc support to acme. I was wondering what would be the best
way to perform the charset conversion. Irc is quite broken, most
channels use legacy encodings, which are specified only implicitely. 

Now given a scenario with 30 channels, what would be a good way to
convert the data streams from the legacy codings to utf? The simple
solution would be to run one tcs for input and an another for output for
each channel. This would mean 60 tcs processes (and 30mb of ram).

The second approach would be to have one tcs process for each channel
for input and run a tcs process each time the user says something.
This would need half the memory, but would incur more overhead/line.

Is there a good solution to this? (Other than changing irc...)

- Einar Karttunen