On Mon Nov 10 20:17:30 2003, Dustin Puryear wrote:
>We are battling the QP encoding issue. For those unfamiliar with this issue,
>when a binary that has been improperly encoded via QP encoding by an email
>client passes through Anomy it can become garbled if Anomy converts a CRLF
>to a LF (or vice-versa?).
>
Somewhere deep in the depths of time, I knocked together an alternative
design for Anomy Sanitizer's main driver which avoided this problem.
The problem is, quite simply, that the Sanitizer decodes and reencodes all
parts.
The thing is, there's no need to decode and reencode the part at all - Anomy
Sanitizer *knows* it's not going to do anything with it by this point, apart
from re-encoding it.
The problem is that Anomy Sanitizer decodes a part before considering whether
that was a good idea or not.
Whether or not the mail client should have used QP is largely irrelevant -
the point is that it has, and to decode and recode without a good reason is,
I'm afraid, a bug.
But this is, as ever, purely my opinion.
Dave