AUSAM/distribut.doc/tape.n

.tr ~ 
.nh
.ll 85
.lt 85
.de HD
.sp 2
.tl 'UNSW Distribution tape''\\*t'
.sp 2
.ns
..
.de FO
.bp
..
.pl 12i
.ds t Format
.wh 0 HD
.wh -4 FO
.sp 2
.cu
.ce
FORMAT OF THIS DISTRIBUTION TAPE
.sp 2
.ti +3
The distribution tape was created with 'dtp' a home grown
supa\-dupa 'tp' at 800 bpi.
Our reasons for the departure from standard Unix programs are twofold.
First, this tape contains all of our collected source, with many more
files than 'tp' can write, and with many more blocks than can be
comfortably written with 512 byte records. Second, we felt that a stand alone tape,
a la the Unix distribution was too inconvenient.
.sp 1
.ti +3
Dtp solves these problems well. It writes blocks of 16K bytes and allows
for (almost) infinite numbers of files. These files can be extracted on
a directory or individual basis, to help our colleagues with small disks.
Dtp also has a create directory mode, for extracting arbitrary chunks of
a file system.
.sp 1
.ti +3
A listing produced by dtp of the distribution tape has been included
in the package. You can scan it to isolate just what you are interested
in from the distribution.
.sp 1
.ti +3
The format of this distribution tape is as follows.
.sp 1
.nf
.in +7
BOT
Rec      1   16k bytes   Start of dtp directory.
Rec   2- m   16k bytes   Remainder of directory.
Rec m+1- n   16k bytes   Dtp dump information.
TAPE MARK
Rec   1-14   512 bytes   Straight copy of dtp program.
TAPE MARK
Rec   1-54   512 bytes   Straight copy of dtp source.
TAPE MARK
TAPE MARK
.in -7
.fi
.sp 1
.ti +3
The number 'm' above can be obtained from the listing included with the tape.
It is the number of directory records on the tape.
.sp 1
.ti +3
The record 'm+1' also contains a copy of the dtp program ('dtp' is the first
thing in the dump, so it is the first thing after the directory). This dtp
is the one we are currently using, and it uses a few extra system calls etc.
It should however extract the tape on a standard Unix system.
.sp 1
.ti +3
The 'dtp' in the second tape file is the one this tape was created with. It
will run on the standard version 6 Unix system.
.sp 1
.ti +3
The source of dtp is in the third tape file in case your system has changed
in some dramatic way from standard Unix, such that this version of dtp will
not run. The source for our version of dtp can be found somewhere in the dump.
Dtp will extract it, but you can do it by hand if you must. See the copy
of dtp(V) and dtp(VIII) enclosed.
.ds t Contents
.bp
.ce
.cu
CONTENTS OF TAPE
.sp 1
.ti +3
The dtp part of the tape was created with the command 'dtp nm0 usr'.
Thus it can be extracted partially or in full into any directory
on your system called 'usr'. If you posess a spare 'large' file system,
to extract this tape into it, mount tape on unit rmt0,
mount spare filesystem on device /dev/name and use the following commands.
.nf
.in +7
.sp 1
mkfs /dev/name count
mount /dev/name /tmp/usr
chdir /tmp
dtp xcm0
.sp 1
.in -7
.fi
The contents of the tape will be extracted. If there is a tape error (urk)
you may use the 'i' option of dtp. The error will be ignored but...
.sp 1
.ti +3
The directory 'usr' contains seven subdirectories, which you should be able
to extract on smaller devices separately. Their names are: docmention,
distribut.doc, source and sys. To extract only 'sys' for example
mount tape and disk as above and use
.nf
.in +7
.sp 1
mkfs /dev/name count
mkdir /tmp/usr /tmp/usr/sys
mount /dev/name /tmp/usr/sys
chdir /tmp
dtp xcm0 usr/sys
.fi
.in -7
.sp 1
.ti +3
If it still doesn't all fit then a similar procedure can be used for
subdirectories of these. You can find their names from the listing or
with the command
.sp 1
.ti +7
dtp lm0