The author is a programmer - not a webdesigner or documentation specialist !
I was contacted more than a dozen time
by developers who are having problems with the decoding or
encoding. They all had the same problem: They did not realize
that the NNTP-protocol requires to double a dot in the first
colum when a line is sent - and to detect a double dot (and
remove one of them) when receiving a line.
So if you are directly posting to or reading from a socket
(internet conenction socket), then dont forget this, please.
I have been notified that 'some mailers' are automatically
converting the string: 'From' at the start of a line within the
body into: '>From'
Especially *nix mailers which keep their mailbox in traditional
Unix-style 'mbox-format' without Content-length seem to do it.
(Notification came from Yakov Lerner. Thank you!)
I have been notified that someone had problems with his
[NSString characterAtIndex] function when he tried to read
yEncoded messages for decoding purposes. That function did
characters conversion based on a default local character set.
Instead 'raw message data' should have been accessed.
Please keep in mind: No matter what charset your newsreader (or a
message-header) specifies: yEncoded data is ALWAYS raw data. Dont
use any conversion during encoding or decoding.
A newsgroup has been created especially for test-posts in yEncoded format: alt.binaries.test.yenc
This newsgroup is monitored and test-posts are checked for
correctness.
An autoposter will send daily short examples for developers.
The mailing list is out of order. There have been no
significant changes and discussions since the version 1.3 of the
yEnc draft.
Feel free to ask by eMail yenc@infostar.de
Discussion about yEnc is done 'in public' in the newsgroup: news.software.nntp.
As some people would prefer to include a new encoding into the MIME structure a short discussion was made in comp.mail.mime. Unfortunately it is very difficult to introduce new transfer encodings and also the rest of MIME is pretty difficult (for non MIME specialists) this topic has been moved downwards on the priority list. A short example for a single binary and a multipart binary has been created and is available in the 'Documents' section. Treat it with CARE !
There is a general approach if you want to embed a yEncoded
message into MIME:
Take the entire yEncoded message as one block and add all you
need for embedding in front of it - and behind it. The =ybegin
line with all parameters, the data and the =yend line should stay
one block. Even if information is redundant - the =y lines should
stay as they are specified (even for multiparts). This allows
other decoders to detect such blocks - and decode them properly -
even if all other information is stripped from your messages.
14. Jan. 2002: The link to the MIME embedding has been removed. We will try to create a better embedding actually.
Proposed new documents (05.Mar.2002)
These docs are not final - but published to replace at least the original draft 1.0 which might be misunderstandable and/or misleading.
Original documents
Changes: 10. June.2002:
Spelling: "grammer" --> "grammar", "Might" --> "May"
There is no copyright for the yEncoding - it is completely
released into public domain.
May it be helpful the Usenet community!
Jürgen Helbing - 10. June 2002 - Email: yenc@infostar.de