Re: [Bug-gnubg] feature request: tuple2string move converter python function (inverse parsemove)

[Guido Flohr]

On Thu, 2012-10-25 at 17:40 +0200, bagrat lazaryan wrote:
> i'm afraid i rushed a little in offering my solution. it seems it isn't as
> trivial a task as it appeared to be at first. as for now, my solution isn't
> handling correctly the doubles. (it gave me "13/9 13/5 9/5" on (13, 9, 13,
> 9, 9, 5, 9, 5) instead of "13/5(2)".) i didn't succeed in coming up with a
> quick fix so i will also have to put it off till the weekend or, given that
> indeed most of the functionality required should be already present, wait
> for your solution. anyways, if you need any help on this will be glad to
> contribute.

I solved that problem with regular expressions resp. substitutions, see
http://git.savannah.gnu.org/cgit/gibbon.git/tree/src/gibbon-position.c
currently after line 1112 in function gibbon_position_format_move().

My approach depends on the point pairs involved in a move on being
sorted, so that movements with the rearmost checkers appear first.

Then I convert the integer pairs into a string, for example something
like "a/c c/e f/h h/j" (for "24/22 22/20 19/17 17/15").  The rest is
done by the glib regex engine, for example

    s/\/([a-z]) \1\//\//g

compresses the above string into "a/e f/j".  Inserting the '*' for blot
hits or grouping identical movements with "(2)", "(3)", or "(4)" works
essentially the same, only it's more complicated.

Maybe buggy, maybe inefficient, but I still like my function. ;)

Guido
-- 
Империя ООД                         | Imperia OOD
ул. „Княз-Борис-I“ № 86, София 1000 | ul. "Knyaz-Boris-I" № 86, Sofia
http://www.imperia.bg/