< From: jwz(_at_)netscape(_dot_)com (Jamie Zawinski)
<
<< From your statements above, you believe that munpack is incorrect.
<
< If you're referring to me,
The attributions changed in mid message. Here, I was referring to Ned
Freed's response.
< that's not what I meant -- I thought that munpack was doing the right
< thing, but that whoever had encoded binary data in QP needed to use Base64
< instead.
<
< But now I agree with Ned -- you can use QP for binary data, but you just
< need to be careful to make CR, LF, and CRLF explicit, and not to use hard
< breaks (CRLF not preceeded by "=") since that's the part that has
< "textual" semantics, and a decoder is allowed to translate those to the
< local linebreak convention.
<
< It still sounds, though, like the program that is encoding binary data in
< QP is doing something wrong -- it's probably using hard breaks.
No, the program is encoding CR's as =0D. The problem is that munpack is
stripping those out after decoding them. My query to the group was: is
munpack wrong in doing so? You and Ned appear to agree with my feeling that
munpack is indeed wrong.
Tony Hansen
hansen(_at_)pegasus(_dot_)att(_dot_)com,
tony(_at_)attmail(_dot_)com
http://ourworld.compuserve.com/homepages/Tony_Hansen