Newsgroups: comp.sys.acorn Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!think.com!cass.ma02.bull.com!mips2!thed.uk22.bull.com!kelvin From: kelvin@thed.uk22.bull.com (Kelvin Hill) Subject: Re: Bad files from info-server? Organization: Bull HN Informations Systems Inc. Date: 30 Jun 91 11:38:20 GMT Message-ID: References: <11369@castle.ed.ac.uk> Sender: news@mips2.ma30.bull.com (Usenet News Manager) In <11369@castle.ed.ac.uk> gtoal@castle.ed.ac.uk (G Toal) writes: >Albert has already posted once saying that if he gets files sent >to him which are not already compressed, he .tar.Z's them. So feed >your file through compress -d and tar xv in that order. The files *are* Spark archives. I ran them both thru a visual check and !Sparkinfo. They are just damaged archives. What I am trying to figure out is wether they were damaged before Albert got them, or if it's the info server damaging them. >The delays aren't from Newcastle -- they're from Britains extremely >damp piece of wet string at nsfnet-relay which goes down more often >than Linda Lovelace :-( Aha, I thought it might be that... thanks for the confirmation. >G >PS Why would you expect CR's? This isn't DOS, you know... In any Sparkive I would expect to see a fairly even spread of all binary values present, otherwise the compression is not doing the full job. If there is one particular binary value missing, then it's a reasonable assumption that somewhere, something has selectively stripped that value out. For example, a DOS to UNIX file conversion/file transfer could have done it. There are lots of LFs in the files, but not one single CR... comments? Kelvin. -- Kelvin J. Hill - BULL HN Information System Ltd, Hounslow, England, UK. Internet - kelvin@thed.uk22.bull.com | UUCP - kelvin@cix.compulink.co.uk "" kelvin@kelvin.uk22.bull.com | AMPRnet - kelvin@g1emm.ampr.org