[comp.sys.atari.st] Dumas encode -- possible bug

silvert@dalcs.UUCP (Bill Silvert) (07/30/87)

I've run into a recurrent bug with multipart uuencoded files from the
net.  The last line appears truncated at the beginning.  Here is a
sample from the recent posting of turtle.arc -- this is the last part of
the fifth file, turtle.uue:

M1S 5 /#;B]'P=[_=&3\ URT $3 $DIW<:.'>,@_T1D\ )" 5:'$1O^WH Q$0e
M*$#<*0 * :$"!@< *V 2 , "X=T"%0$7O>T"FOX"]OWH%3$5)C$59WT5%7$5e
EV! 0-6#?70$) 0$ -T#<.&#?B69P * #QY3JJ,X#V/ 0*  : #$5e
 e
end

I've had other encoded files which were also perfect except for the last
line.  I haven't had a chance to test this, but has anyone else
encountered it?  Also, if anyone else has a decent last part of turtle,
please send it to me.
-- 
Bill Silvert, Modelling/Statistics Group, Marine Fish Division
Bedford Institute of Oceanography, Dartmouth, NS, Canada B2Y 4A2
UUCP: ..!{seismo|utai}!dalcs!biomel!bill (preferred address)
CDN or BITNET: silvert@cs.dal.cdn (only if UUCP address can't be used)

pes@ux63.bath.ac.uk (Smee) (08/03/87)

In article <2604@dalcs.UUCP> silvert@dalcs.UUCP (Bill Silvert) writes:
>I've run into a recurrent bug with multipart uuencoded files from the
>net.  The last line appears truncated at the beginning.  Here is a
>sample from the recent posting of turtle.arc -- this is the last part of
>the fifth file, turtle.uue:
>
>M1S 5 /#;B]'P=[_=&3\ URT $3 $DIW<:.'>,@_T1D\ )" 5:'$1O^WH Q$0e
>M*$#<*0 * :$"!@< *V 2 , "X=T"%0$7O>T"FOX"]OWH%3$5)C$59WT5%7$5e
>EV! 0-6#?70$) 0$ -T#<.&#?B69P * #QY3JJ,X#V/ 0*  : #$5e
> e
>end

Prima facie, that looks like a totally reasonable way for a uuencoded file
to end, to me.  What makes you think it's been truncated?  The last 'encoded'
line (one beginning with 'E' above) is *usually* (59/60ths of the time?)
shorter than the preceding encoded lines, and (perforce) starts with a char
other than 'M'.

For a concise description (though not the world's best read) of UUENCODED
format, see UUENCODE(5) in the Unix Programmer's Manual Reference Guide,
or try 'man uuencode' on your nearest friendly Un*x system.

Just witter on a bit now, to make my comment longer than the quote, so that
the silly mailer will accept this comment.  Wonder how much extra net
traffic *that* silly rule generates?

No, don't tell me, I don't want to know.

leilabd@cvaxa.sussex.ac.uk (Leila Burrell-Davis) (08/06/87)

In article <2604@dalcs.UUCP>, silvert@dalcs.UUCP (Bill Silvert) writes:
> I've run into a recurrent bug with multipart uuencoded files from the
> net.  The last line appears truncated at the beginning.  Here is a
> sample from the recent posting of turtle.arc -- this is the last part of
> the fifth file, turtle.uue:
>
> M1S 5 /#;B]'P=[_=&3\ URT $3 $DIW<:.'>,@_T1D\ )" 5:'$1O^WH Q$0e
> M*$#<*0 * :$"!@< *V 2 , "X=T"%0$7O>T"FOX"]OWH%3$5)C$59WT5%7$5e
> EV! 0-6#?70$) 0$ -T#<.&#?B69P * #QY3JJ,X#V/ 0*  : #$5e
>  e

This is perfectly normal behaviour by uuencode, standard or Dumas
variety. The first character on the line indicates to uudecode
how long the line will be and the E at the beginning indicates a
shorter line than the more usual M.

I dumas decoded turtle with no trouble and, though I haven't
tried to run it yet, `arc t turtle' says that all the files are
ok.
-- 
Leila Burrell-Davis, Arts E, University of Sussex, Brighton BN1 9QN, UK
Tel:  +44 273 606755  Ext. 2387          JANET:  leilabd@uk.ac.sussex.cvaxa
ARPA: leilabd%sussex.cvaxa@cs.ucl.ac.uk  BITNET: leilabd%sussex.cvaxa@uk.ac
UUCP: leilabd@cvaxa.uucp OR ...!mcvax!ukc!cvaxa!leilabd