Booz 1.02 -- Barebones Ooz
                                  a
                         Zoo Extractor/Lister
                                  by
                             Rahul Dhesi

Booz 1.02 is a small, memory-efficient Zoo archive extractor/lister. 
It is not fancy.  It does not recognize the advanced features
available in current versions of Zoo, such as long filenames,
directory names, comments, and multiple file generations.  Extraction
always uses a short MS-DOS format filename and all extracted files go
into the current directory.

But Booz 1.02 is simple and portable and can be implemented in about
fifteen minutes on any system with a reasonably good C compiler that
provides **IX-compatible read(), write(), and lseek() functions.
And Booz 1.02 can extract and list all archives created by all
currently-existing versions of Zoo.

At compilation time, conditional compilation selects one of three
levels of sophistication for Booz 1.02.  The three options are as
follows.

The Tiny option:  Booz compiled with the Tiny option is very frugal
and should compile and run under CP/M and other systems with limited
memory.  Tiny Booz always extracts an entire archive at a time and
does not attempt to prevent an exctracted file from overwriting
another.  Tiny Booz requires you to type the entire filename of the
archive being extracted, including the ".zoo" extension.

The Small option:  Booz compiled with the Small option is a little
more sophisticated than Tiny Booz.  Small Booz assumes the default
extension ".zoo" for an archive if it is not specified.  It will let
you specify which files should be extracted from an archive.  Small
Booz accepts these wildcard symbols in the names of files to be
extracted:  "?" stands for any one character and "*" stands for any
sequence of zero or more characters.  Combinations of these are
permitted so, for example, "*c?t*h" matches filenames ending with "h"
and containing "c" and "t" separated by one character.  Small Booz is
nearly as memory-efficient as Tiny Booz and it very likely that it
will compile and run under CP/M.

The Big option:  Booz compiled with the Big option does everything
that Small Booz does and in addition it can give the directory
listing of a Zoo archive or test its integrity.  Despite its name,
Big Booz is still quite frugal.  Under MS-DOS, when compiled with
the Microsoft C compiler version 3.00, it runs in about 50 kilobytes
of free memory and occupies about 9.5 kilobytes on disk.


                         COMPILING BOOZ 1.02

1.
Make sure that the two macros OPEN and CREATE are correctly defined
for your system in file `oozio.h'.  Some sample macros are provided. 
The macros must be defined to open files in binary mode (i.e.,
without newline conversions).

The macro OPEN is supplied a filename and it must open the file for
reading and return an open file descriptor, or -1 if the open fails. 
It is used to open the archive being extracted or listed, and to test
the existence of a file about to be extracted.

The macro CREATE is supplied a filename and it must create a new file
for writing and return an open file descriptor, or -1 if the create
fails.  It is used for creating each file that is extracted.

2.
If your C library does not provide the unlink() function (which
deletes a file given its name), define an empty function by that
name, or define a suitable (possibly empty) macro by that name in
file `oozio.h'.

3.
Decide which of the three options, Tiny, Small, or Big, you will use.
If memory is tight use the Tiny option first to get a Zoo extractor
quickly running on your system.  Then experiment with the other
options.  If memory is not tight, use the Big option.  Define any one
of the symbols TINY, SMALL, or BIG in the file `options.h' (or
alternatively define it on the command line that invokes your C
compiler).  Exactly one of these three symbols must be defined.

4.
Choose appropriate sizes for the input/output buffers and for the
size of the decompression stack in file `options.h'.  If memory
permits, make the buffers 8192 each.   (Bigger sizes, perhaps up to
16384, could be used but they have not been tested and could
conceivably cause overflow in certain bit-shift operations in file
`lzd.c').  A reasonable stack size is 2000, but in pathological cases
extraction of a file could require a larger stack.

If memory is tight, decrease the input/output buffers to 1024 each
each and the stack to 1000.

5.
Compile and link all the C files.  Two makefiles are supplied, one
suitable for **IX systems and the other for Microsoft C 3.0 under
MS-DOS.  Modify these as necessary for your system.  Also supplied is
a project file booz.prj for use with Turbo C 1.0 on MS-DOS.


                          MACHINE DEPENDENCE

Booz is relatively independent of machine architecture, except that
(a) the machine must be a 2's complement machine (all modern machines
are) and (b) `char' must be exactly 8 bits, `int' must be 16 bits or
larger, and `long' must be 32 bits or larger (though a `long' of a
little less than 32 bits may work).

Booz makes no assumptions about the filename syntax of the host
machine, except that Small and Big Booz assume that dot "." is used to
separate the extension "zoo" from the rest of the name of the
archive.  They will append ".zoo" to the archive name if it contains
no dot;  this fails if an archive name of the type "../junk" is
specified.  

If your system uses a different filename syntax, you may need to
change the code.  Also, if your system cannot accept some of the
characters legal in MS-DOS filenames, you may need to add a function
that will fix the filename before extraction.  A sample function
`fixfname()' is included in the file `oozext.c'.  It becomes activated
only if the preprocessor symbol FIXFNAME is defined.

				 NOTE

This program does not attempt to be case-insensitive.  Therefore you
will need to type names of files to be extracted in the correct
case.  Note, however, that some operating systems (possibly CP/M) may
fold command line arguments to uppercase.  If this happens, your
best bet is to extract all files from an archive instead of
specifying specific names.

			   REVISION HISTORY

Version 1.00
     Corresponded to just the Tiny booz 1.02

Version 1.01
     Included TINY, SMALL, BIG compilation options.  Had a bug in
     function needed() that sometimes cause file extraction to fail.

Version 1.02
     Fixed bug in function needed().  Added support for Turbo C 1.0.
     Revised this documentation and some comments in the source code.

                                -- Rahul Dhesi 1988/08/25