Re: client requests ending \012

[Keith Moore <[e-mail]>]

> The difficulty is with imposing a fix for a local problem as a global
> product default.

except that it's not purely a local problem - the same clients could
(and often do) crop up anywhere.  and the vendor ends up dealing with
this support issue again and again.  much easier to just 'fix' the 
problem.

I'd say that the difficulty is one of trying to get people to accept
near-term pain to minimize long-term pain.  and also the difficulty of
replacing many broken clients when you can "fix" the problem on a small 
number of servers.

Keith

this may actually be a "flaw" in the traditional IETF notion of 
interoperability testing being sufficient - merely being able to 
interoperate doesn't mean that the implementations conform to the 
standard, or more importantly, that the implementations behave 
robustly in corner cases (like LF . LF in the middle of a message).