Possible compress bug (Re: Finding Books (quat itat itang?))

[Scot E. Wilcoxon]

In article <479 at snow.warwick.UUCP> kay at warwick.UUCP (Kay Dekker) writes:
>This article came to us via warwick, ukc, mcvax, seismo, ut-sally, im4u
>and caip; it originated somewhere on the ARPAnet.  Is this editor-mangle,
>gateway mangle or news-system mangle?
>In article <1633 at caip.RUTGERS.EDU> daemon at caip.UUCP writes:
>>From: Randall B. Neff <NEFF at SU-SIERRA.ARPA>
>[I omit heaps of stuff here]
>>If you know what you want and know the ISBN (from 0 or distributor's catalogs)
>>the store can very likely get it for you.  With the right relr:3,
>>you 3,3,3ablewn,e your
>>oril ghead of  of  
>>usepular.ED.ED.-Iopere cannot 3
>>
>>uth oe firee er
...

It looks like compress 4.0 mangle.  I just had a C source get mangled
the same way. I think it looks like part of the decompression table is
wrong so stuff is garbled.  The problem has been seen to happen when
uncompressing many files with one command (ie, "uncompress *"). [Thanks to
the Unix Users of Minnesota members who uncovered the bug at a meeting 
last week]  Some large files also have been lost, but I don't yet know
if those were done singly or with other files.

In my case I was uncompressing some 3.0 stuff, so I know the bug was in the
uncompress.  Also, a backup of the file was uncompressed fine by 4.0 when
it was done alone.

I did notify the 4.0 author, and he hadn't heard of this before.
Perhaps someone can find a repeatable example for him?
The program is wonderful and, unfortunately in this case, quite robust.
-- 

Scot E. Wilcoxon  Minn. Ed. Comp. Corp.       quest!mecc!sewilco
45 03 N / 93 08 W   (612)481-3507   {ihnp4,philabs}!mecc!sewilco
"The appeal of nuclear energy has decayed rapidly in the UUSR"