Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!purdue!haven!adm!smoke!ibd!heilpern
From: heilpern@ibd.BRL.MIL (Mark A. Heilpern )
Newsgroups: comp.unix.wizards
Subject: Re: reading an "IBM tape" on UNIX
Message-ID: <260@ibd.BRL.MIL>
Date: 5 May 89 14:53:55 GMT
References: <23398@shemp.CS.UCLA.EDU> <4847@macom1.UUCP>
Reply-To: heilpern@brl.arpa (Mark A. Heilpern (IBD) <heilpern>)
Organization: Ballistic Research Lab (BRL), APG, MD.
Lines: 38

In article <4847@macom1.UUCP> larry@macom1.UUCP (Larry Taborek) writes:
>From article <23398@shemp.CS.UCLA.EDU>, by tamir@oahu.cs.ucla.edu:
>> I have received a tape that was written on an IBM mainframe
>

>
>> if there is a way to read this tape on UNIX.
>
>
>> /*     DEFINE TFB-512 AS 199 (BLK 60000                           */
>> /*     FORMAT 199 F (B 4096                                       */
>

>I have read successfully Mag tapes created off an IBM mainframe.
>I use the dd command to read them.  You will probably have to
>play around with this some, varying the parameters but try 
>something like this:
>
>dd if=/dev/rmt0 of=stuff0 conv=swab cbs=512
>
>where rmt0 is the device driver for the mag tape device that you
>are reading from and stuff0 is the file you are going to put the
[lots deleted]

An easier method, if you are unfamiliar with the use of dd, is to
use the public domain program 'ansir', ansi-read.
I believe source for this exists on uunet.uu.net.
It can read files in a variety of different formats, including both
ascii, ebcdic, and binary, in the format of quite a few machines.
One bug in it I've found however: While trying to read files in
DEC format (from the rsx-11 operating system), 'ansir' was unable
to retrieve the file-name information. I later found out 'ansir'
had placed 14 extra bytes at the head of each file-- the name, user/group
codes, etc, in a compressed manner. It was rather difficult writing
the (basically simple) program to decode the name from this.
The difficulty was in a lack of information.

	--M.