[Bug-gnubg] Python scripts

[Massimiliano . Maini]

Hi all,
I was playing with the exportBad() function declared in gnubg.py : I
modified
it slightly in order to classify the errors into chequer, cube,
chequer-and-cube
and resign categories (appending the category at the end of the file name),
and
put my "own" version into a separate file (max.py) as exportErrors() (text
export).

I tried it on 4 analyzed matches (2 real, 2 just silly) and I noticed
something
strange : it works fine on matches ztm1 (2 chequer errors), ztm2 (2 chequer
errors) and ztm4.sgf (1 chequer-and-cube), but something is broken for
ztm3.sgf.

The files created do corresponds to the "errors" in the match (chequer,
cube,
cube, cube, chequer), but only the first export is correct, the others are
"empty" (only the board is printed out, no analysis, and the position
doesn't
event match).
Even stranger, running the script on ztm3 triggers some evaluation : gnubg
plays a move !!

In the attached archive there's the max.py script, the 4 sgf files and the
various output produced by the export (text). Can somebody take a look and
tell me if the problem is in the script or ...


MaX.

(See attached file: pbpy.zip)

pbpy.zip
Description: Zip archive