NetNews Usenet Archive 1992 #30/NN_1992_30.iso/spool/comp/programm/3308

Text File | 1992-12-11 | 1.6 KB | 39 lines

Newsgroups: comp.programming
Path: sparky!uunet!gossip.pyramid.com!decwrl!amdcad!angelo!roberts
From: [e-mail] (Dave Roberts)
Subject: Re: Binary to ASCII encodings (was Re: A better mousetrap)
Message-ID: <[e-mail]>
Sender: [e-mail] (NetNews)
Nntp-Posting-Host: angelo
Organization: Advanced Micro Devices, Santa Clara, CA
References: <[e-mail]> <[e-mail]>
Date: Fri, 11 Dec 1992 20:13:18 GMT
Lines: 26

In article <[e-mail]> [e-mail] (Tim Goodwin) writes:

>uuencode uses a fixed 3-byte -> 4-character encoding, for a 1.333
>swell factor.  (There is also some additional overhead, for example the
>line feed at the end of each line, and the "M" at the beginning of
>each.)

Just a quibble.  The "M" at the beginning of the line isn't an "M" per
say.  It's actually a length of the number of character in the line
that has been encoded in the same fashion as the six bits of normal
data are.  It so happens, however that most (all?) of the programs
that do uuencoding use the same line length.  If you look at the last
line of the encoded file, however, (the one before the "end" line)
you'll see that it probably isn't as long as the others and doesn't
have an "M", but rather something else as the first character.

In other words, if you're writing a decoder, don't look at the first
character of a line as an "M", look at it as a length.  It makes
everything general, and if anybody writes an encoder that uses a line
length different than "M" (which I suppose is legal), yours won't be
the decoder that breaks.

Dave Roberts
[e-mail]