summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2008-09-02Added a few bugs to the TODO file.John Wiegley
2008-09-02Removed code that was appending text to the ends of variables, rather thanJohn Wiegley
using format string specifiers.
2008-09-02Implemented a safer method for parsing algebraic moves, since for some reasonJohn Wiegley
we weren't always seeing the full move when using looking-at. The way I do it now is to require some kind of whitespace after the SAN move text.
2008-09-02Updated TODO file.John Wiegley
2008-09-02If a chess message cannot be found, let the user know which one.John Wiegley
2008-09-02Don't highlight the opening paren for special chess-* forms.John Wiegley
2008-09-02Simplified the test runner's output, so it fits in 80 columns.John Wiegley
2008-09-02A COUNT of 0 now means "read to end of database", not -1, which Emacs wasJohn Wiegley
taking to be an option flag.
2008-09-02Corrected game indices in the debug output.John Wiegley
2008-09-02'make check' is now 0-based, like the rest of the database-related code. OnlyJohn Wiegley
in the display code (the top-level user interface) are database games presented using a 1-base system.
2008-09-01Highlight `chess-error' the same way as `error'.John Wiegley
2008-09-01The 'start' parameter passed in is now a 0-based index, like what is expectedJohn Wiegley
by chess-database.el.
2008-09-01chess-scid now accepts a 0-based index, but uses a 1-based index when talkingJohn Wiegley
to ChessDB.
2008-09-01Use the modulus of the read-count to govern when game counts are reported.John Wiegley
2008-09-01Report an error if we fail to find the beginning of a game loaded fromJohn Wiegley
ChessDB, or fail to see its end. Also, give up to 10 seconds to wait for the result to come from ChessDB, polling every second to look for new output.
2008-09-01Added a note about how to run the full test suite properly.John Wiegley
2008-09-01Erase the buffer before asking scid for a PGN position. What happens duringJohn Wiegley
running of the validation tests is that he buffer fills up before we can evaluate all the games.
2008-09-01When byte-compiling chess-pos, set cl-optimize-speed to 9. This increasesJohn Wiegley
the computation of chess-eco.fen by up to 50% on my system.
2008-09-01Off by one error in string clipping of tag values.John Wiegley
2008-09-01Only require 'cl when byte-compiling chess-ai. We don't yet use the CLJohn Wiegley
compatability functions at runtime.
2008-09-01Fontify chess-error and chess-with-current-buffer as keywords.John Wiegley
2008-09-01chess-auto.el is built properly now.John Wiegley
2008-09-01Simplified error message when an algebraic move fails to work.John Wiegley
2008-09-01Fix ics style12 regex.Mario Lang
2008-09-01Make game numbers displayed equal to indexes used in chess-database.Mario Lang
2008-09-01Remove the supposed failing game again since emacs-chess was actually right, ↵Mario Lang
the game violates the rules of chess.
2008-09-01Fix a "off-by-one" thinko, database indexes are zero based.Mario Lang
2008-09-01Add a failing test I found in TWIC.Mario Lang
2008-09-01Simplified some error text.John Wiegley
2008-09-01Added a TODO note: I found a test case for the failing promotions.John Wiegley
2008-09-01Removed the unnecessary chess-incr-index* function.John Wiegley
2008-09-01Removed historical note, since there is a better story in AUTHORS now.John Wiegley
2008-09-01Ignore the largedb files if a user has installed them.John Wiegley
2008-09-01Updated the TODO file.John Wiegley
2008-09-01Instead of using `read' to parse a PGN tag string, just clip off the leadingJohn Wiegley
and trailing quotes.
2008-09-01When a testing error occurs, mention the exact test that failed along withJohn Wiegley
the cause. Also, make it so that a failing test does not stop all the tests from running.
2008-09-01Report a clearer error in the case of an algebraic queening move with anJohn Wiegley
unspecified promotion.
2008-09-01If the COUNT is less than zero, run all the remaining tests.John Wiegley
2008-09-01Added a basic file of 400 historic games, which gets used as the baseline testJohn Wiegley
set if the large database is not present.
2008-09-01Several improvements to how tests are run.John Wiegley
First, if the user hasn't downloaded the monster test database, we just run a simple test of 400 historical games. Second, you can specify the variables START and COUNT on the command line, for running just a part of the huge database.
2008-09-01Allow the stress tests to be run across a specific range:John Wiegley
make START=2000 FINISH=5000 check If START is omitted, it begins at game 1. If FINISH is omitted, it ends at the last game in the test database.
2008-09-01Except for when making moves interactively (with chess-display-manual-move),John Wiegley
it is an error if a ply implies a promotion but not :promote keyword is passed to indication which piece to promote to.
2008-09-01Ignore the chess-test script.John Wiegley
2008-09-01Increased the verbosity and number of errors generated for incorrectly parsedJohn Wiegley
PGN games.
2008-09-01Wrote a testing function for emacs-chess, which computes the legality of pliesJohn Wiegley
for a database of known legal games.
2008-09-01Several fixes to the scid (aka ChessDB) interaction protocol.John Wiegley
2008-09-01Removed some trailing whitespace.John Wiegley
2008-09-01chess-test is now invoked with the name of the database to read.John Wiegley
2008-09-01Moved some code around to follow better Lisp style.John Wiegley
2008-08-30Quick title fix to the TODO file.John Wiegley