XPRXMODEM v34.3

A small bug crept into the code, which went undetected, until XPRXmodem
was tested with other versions of XModem.  (Thanks to Scott Velez for
helping to confirm this.)  It has now been fixed.  Following is the
"readme" file from the previous release.  Note the file size is now
12,352 bytes, and I've adapted xcrc.h (from XPRZmodem v2.10) for
XPRXmodem, since it's a bit faster.  Also, MYREVISION is now 36 (libsup.c),
so "version xprxmodem.library" should now correctly, return 34.36.
(I forgot to bump this in v34.2.)

                                        My humble apologies. -- Anthon

                      - - - - - - - - - - - - - - - - - -

                                XPRXMODEM v34.2

Following is an independent (ie by me) update to xprxmodem.library (v34.1),
by Marc Boucher.

The only change is to crc checking, in which, the old method of computing
crc polynomials on the fly, has been replaced with a table of pre-computed
polynomials and the appropriate code (by Marc G. Mendel), reducing the time
between receiving the block & acking it.  Thus, transfers are sped up.
(The really observant will notice the source code was extracted from the
Unsit1.5 archive that I recently uploaded to Unix.source).

Here's some data rates (reported by VLT's update window) for a null modem
setup between two Amigas running VLTjr & xprxmodem (1K blocks with CRC, of
course):

            2400 baud    19200 baud(*)   |   file size
                                         |
     34.1      201           654         |    11,820
                                         |
     34.2      212           796         |    12,332
    _____________________________________|
                                         |
     Gains      5%           20%         |

Diffs to dlink.c are included for those interested in the code.  Also,
I've bumped up the revision number to v34.2.  The library was compiled and
linked with Manx Aztec C v5.2.

                                   -- Anthon Pang (a665@mindlink.bc.ca)

(*) You'll have to look elsewhere for an explanation of why it's
    not 8 times faster...