Re: [gnugo-devel]

[bump]

> It's still wrong in stephane_3_22.1b, because t is not initialised now !
> You forgot to assign its value after the if-else block :
> 
>    t = cosmic_importance;

Oops, the patch stephane_3_22.1b was correct but the link at
devel.html still pointed to stephane_3_22.1a. I've corrected
that.

The patch stephane_3_22.1a is different from the patch
originally posted. That should not inconvenience anyone
since the patch is definitely superceded.

Sorry about the confusion.

I've been running the regressions with the patch stephane_3_22.1b
but changing COSMIC_GNUGO to 0. Is this supposed to give the
same results as the standard CVS? It doesn't ...

./regress.sh . trevora.tst 
610 unexpected FAIL: Correct 'D2|F2', got 'F3'
./regress.sh . nngs1.tst 
39 unexpected FAIL: Correct 'R7', got 'S3'
./regress.sh . endgame.tst 
810 unexpected FAIL: Correct 'F5|F4', got 'L9'
./regress.sh . trevorb.tst 
440 unexpected FAIL: Correct 'E6', got 'K5'
./regress.sh . nngs.tst 
1280 unexpected PASS!
1770 unexpected FAIL: Correct 'M2|M3|M5|O1', got 'P1'
./regress.sh . strategy3.tst 
120 unexpected PASS!
./regress.sh . arend.tst 
33 unexpected PASS!
...


Dan