GitLab

integrated Pieces functions into Board, independently spotted and fixed one bug in testmove where target 'o' changed to '*' a while back, there was one 'o' left unchanged

Gui: split FONTFAM into two font families MAINFFAM and GRIDFFAM, added $gridfont; g.s: set GRIDFFAM to a nice fixed width font family (dejavu sans mono)

split out ansi colour codes from Piece into new AnsiColour; added backgrounds too, and find() function, used it in Board and Piece; in Piece->as_string, added optional fg and bg colour override and got that working..

minor changes: removed unused gamestate{RESULT} and {PREVBOARD}, used (x,y) tuple assignments in 3 functions for clarity

renamed pushUnicode/popUnicode (and underlying stack) to pushFancy/popFancy cos it enables Unicode and **COLOUR** together

centralised all esc-code knowledge in Piece.. Board uses Piece::black and Piece::whitebg (new in Piece, was in Board) now

ok, changed Piece's %unicode and %piece2colour from my our, removed 2nd copy from graphicalsnowman, accessed %Piece::unicode and ..piece2colour.. saved approx 40 lines

initial work to extract common piece2colour and unicode hashes from Piece.pm and graphicalsnowman.. made both identical in both files..

hmm.. need utf8 now to enable the ansi corner and side codes, after switching to \N unicode char names in Piece.pm.. not quite sure why

preparatory changes (splitting out colours and unicode sequences) before realising that Piece.pm and graphicalsnowman have common (colour and unicode) lookup tables..

split out the colour path logic (redid as a nice class) and modified graphical snowman; will use the colour path stuff in snowman soon..