[Bug-gnubg] Garbage output in double analysis

[Massimiliano Maini]

Hi all
I have a strange and annoying bug coming
up systematically (Win Vista and Xp, code snapshot 20080312).

Start gnubg (GUI), start a 7pt match
and double at your fisrt occasion: gnubg will take and you will  be
on
roll: before moving, inthe move list,
click on your double decision and you will see inthe analysis panel some
garbage in the L(g) percentage and in
the cubeless/cubeful equities (probably coming for the 1st bad percentage).

Clearing the analysis and redoing the
analysis shows the correct values.

It happens all the time (in some situations,
the take/pass decision taken by gnubg is even wrong once the
analysis is performed from scratch).

Anyone can reproduce ?

MaX.