james-server-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kenny Smith <jakarta-ja...@jakarta.apache.org>
Subject Re: fix for bug 20370
Date Wed, 04 Jun 2003 16:40:08 GMT
> I don't think we should be failing on bad data if it's pretty clear how 
> to handle it gracefully.

Hi Serge,

I understand where you're coming from, because we want James to be a 
friendly application to use. However, when dealing with communication on 
this level, there are well documented standard on how to communicate. We 
should not dilute the structure of the server to cater to clients that 
do not follow the rules.

I would suggest perhaps changing the error messages. If someone is only 
sending LR instead of CRLF, I think we should deny the command, but deny 
it with an error message that tells them they need to use CRLF and gives 
them a URL to learn more (however, that URL should not be the RFC, it 
should be a doc for the unsavvy and the impatient).

A lot of geeky people fail to read RFCs or lengthy docs (that why I 
suggest an doc for the impatient) . Most will just interact with it and 
see how it responds, then work off of their assumptions. If they 
interact with it and James tells them "You're doing A and B wrong." they 
will learn to do it right. Also... if they are an unsavvy end user and 
they get an error message saying they are doing something wrong they 
won't think "*&^^ing james... it never works." they will say "stupid 
mail client... it never works."

My unfortunatley too busy to help with the source code but overly 
opinionated 2 cents,
Kenny Smith




---------------------------------------------------------------------
To unsubscribe, e-mail: james-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: james-dev-help@jakarta.apache.org


Mime
View raw message