Landon Curt Noll (chongo was here) /\../\
http://www.isthe.com/chongo/index.html

@(#) $Revision: 13.1 $
@(#) $Id: README,v 13.1 2006/08/14 03:16:33 chongo Exp $
@(#) $Source: /usr/local/src/cmd/hash/RCS/README,v $

Lanndon Curt Noll (chongo was here) /\oo/\
http://www.isthe.com/chongo/index.html

Share and enjoy! :-)

This is an implementation of the Secure Hash Standard-1 (SHS1)
as well as implementation of the RSA Data Security, Inc.
		MD5 Message-Digest Algorithm

The digests produced from strings (-s string), files or stdin are
identical to the original programs.  The command line and output
interface are upward compatible as well.  Users of the original shs
program may replace it with this version has their existing use
and digests will be preserved.

To build:

    <<Edit the Makefile as needed>>
    make all
    make test

To install:

    make install

This implementation has been improved over the original posting in
a number of ways:

  * performance

    On some platforms, the characters per second processed have been
    increased by a factor of almost 3.5 times the original postings.

  * inode digests

    This flag allows one to include the filename and various inode
    information in the digest.  This option is useful in detecting
    file tampering where the contents of the file was later restored.

  * version string

    The -v flag prints the version.

  * test suites

    The standard test suite (-x) been changed.  The initial string
    for shs now says 'shs test suite results' instead of 'SHA test
    suite results'.  While it is true that it we testing the
    algorithm the program is called shs.  The initial string for
    md5 now says 'md5 test suite results' instead of 'MD5 test
    suite results'.  These changes also helps distinguish this
    version from much older versions which did not have the -v
    (print version) flag.

    The original test file "foo" which contained the string "abc" (with
    no newline) was renamed file1.

    To avoid filename conflicts, the shs and md5 verify.data files
    were renamed shs.data and md5.data respectively.

    A new file, "file2" was added to the standard test suite.  This
    file is larger than an internal read and thus helps test file
    processing.  The contents of this file are two highly munged
    copies of the original README.  The munging to prevent people
    from using this file as documentation.

    The test suite will first look in the current directory for
    "file1".  If it is not found, it will look in ${DESTDIR}
    (usually /usr/local/lib/{shs,md5}) as defined in the makefile.
    It is assumes that "file2" is in the same location as "file1".

    With the exception of the initial title line and the addition
    of file2, the md5.data and shs.data file is identical to the
    respective originals.

  * prepend data

    By using '-p string' one may prepend any set of data with a
    string.  The digest produced is equivalent to digest that is
    produced with the string pretended to the data.  By use of the
    '-P file' interface, one may prepend using up to the 32k of a
    file.  This interface allows one to prepend using binary data.

  * performance test suite

    The -t performance test has been changed to compute user cpu
    time.  Older tests looked at wall clock time.  Variations in
    system load produced variations in the reported performance.

    The test length was increased to 16 megabytes.  One may change
    the number of megabytes processed by changing the value of
    TEST_MEG in the file shsdrvr.c or md5drvr.c.

    On some systems with certain virtual memory characteristics,
    you may need to run the performance test 3 times, ignoring the
    first result.

  * quiet mode

    When -q is given, only digests are printed.

  * C-type hex values.

    When -c is given, digests are prepended by 0x so that they
    may be directly used by programs such as calc:

		ftp://ftp.uu.net/pub/calc

    that want C-style hex values.

  * multiple digests

    When -m num is given, num multiple digests are computed in parallel.
    The prepend data (if any) and the input data are divided so that
    each num-th byte is digested by a separate hash.  The result is
    a hash value that is num times as long as a single digest.

    The -d is a historic (pre version 3) shorthand for -m 2.

    When -C is given, the digests are cancatinated together to form
    one large hex value.

  * use by other programs

    External programs may use the hash and I/O modules without all of
    the other structure.

	To use SHS grab:    shs.c,  shs.h and if I/O is needed shsio.c.
	To use SHS1 grab:  shs1.c, shs1.h and if I/O is needed shs1io.c.
	To use MD5 grab:    md5.c,  md5.h and if I/I is needed md5io.c.

    One will need to declare:

	char *program;		/* program = argv[0] */
	int debug = 0;		/* 1 => enable debugging */

See shs1drvr.c (and shsdrvr.c) and md5drvr.c for version information.  See
the man pages shs.1 and md5.1 for other details.  Also see README.shs and
README.md5.  Happy hashing!