		       Release Notes CBUFF 0.10
		Anjo Anjewierden (anjo@swi.psy.uva.nl)
		           December 14, 1993

IMPORTANT
=========

CBUFF now assumes that C char's are unsigned.  In GCC the flag
-funsigned-char must be given and in Borland the flag -K.  The respective
Makefile's take this change into account.

The "culatex" utility may be BROKEN.


BUG FIXES
=========

This intermediate release fixes a rather nasty bug: some ambiguous moves
would not be detected correctly.  It turns out that this bug involves rooks
on a1.

In "culatex" both Black and White players would have the same name.

In "cusort" -white 0 (etc.) would not work correctly.

The utility "cumerge" is excluded.  (It will reappear once merging
of databases is implemented.)


PERFORMANCE IMPROVEMENTS
========================

Reading and writing databases is much faster.  Most of the changes
are in "cbase.c", see for example the functions readGameFileCBase()
in that file.


CUSORT
======

There is a new option called -place (without an argument).  This sorts
on the place where the game was played.  The -place and -source options are
mutually exclusive (-place uses the "places.def" file to distinguish
between places that have a space in them --e.g. "New York", which was
called "New Amsterdam" until Peter Stuyvesant sold it for a dollar--).

If no sort options are given to "cusort" then it will use the "default
sort order", the following two lines are equivalent:

    % cusort database
    % cusort -white 8 -black 8 -place -year database


CUPGN
=====

"cupgn" now tries to confirm as much as possible to the proposed PGN
standard written by Stephen Edwards.  Edwards has suggested to write player
names as "Robert Fischer", this is of course completely hopeless for
ChessBase files (I have mailed him some comments on the issue; in
particular that sorting is impossible if the surname is not first).  The
aim is that "cupgn" can generate files suitable for reading by the free
program "pgnread" by Keith Fuller.

The default of "cupgn" is now to load the "ascii.sym and then the "pgn.sym"
symbol files.  So you can do:

    % cupgn database > database.pgn

To get a PGN database roughly as the standard proposes.

New options are:

-simple
    This produces output similar to the original version of "cupgn".
    The default is to follow the proposed PGN standard.
-complete
    This produces headers for all fields in the game (including
    ELO and ECO if these are known).

-event <event>
    This forces <event> to be the event for the entire database.
-place <place>
    This forces <place> to be the place for the entire database.
-date <date>
    This forces <date> to be the date for the entire database.

A more complicated example is:

   % cupgn -event "PCA interzonal" -place "Groningen, NED" \
        -date "1993.12.19" pca

would be used for the first round of the given tournament.


VERSION MANAGEMENT
==================

The version numbers are now:

CBUFF library	0.9.2

Utility	        Version    DOS limitations (approximately)
----------------------------------------------------------
cuclean		1.2.0      60000 games / database
culatex		1.2.0	   60000 games / database
culist		1.2.0	   60000 games / database
cunames		1.1.0	   60000 games / database
cupgn		1.2.0	   60000 games / database
cusort		1.2.0	   13000 games (see Memory available statistics)
custats		1.2.0	   60000 games / database
cuverify	1.2.0	   60000 games / database


