tiny pattern tuning (endgame and patterns2)

[alain]

Remove CF3 redondant and broken pattern in endgame.db => 1 PASS
Fix some symetry in endgame.db and patterns2.db

nngs:308        PASS L3 [L3|M19]
Total nodes: 1691662016 3342344 12622102
Total uncertainty: 47.59
1 PASS

[alain]

bogus display of the patch in trac! There are 2 files, the last one is not shown :(

[nando]

Replying to alain:

bogus display of the patch in trac! There are 2 files, the last one is not shown :(

Maybe you better do a diff against CVS (cvs diff -u). I guess, the display is bogus, because of the missing "Index: ..." lines.

[alain]

Replying to nando:

Replying to alain:

bogus display of the patch in trac! There are 2 files, the last one is not shown :(

Maybe you better do a diff against CVS (cvs diff -u). I guess, the display is bogus, because of the missing "Index: ..." lines.

Yes, that's it. :) everything ok now.

[alain]

diff -u6 for easier check of the patch

[nando]

Glad it worked. Now a few comments :

• pattern commentars usually follow the Pattern XXX line and, if it only fixed the symmetry, then it says just that (keeping the previous setting isn't very helpful)
• an exception to the above rule: pattern removal notices usually precede the Pattern XXX line
• the SenteXXX pattern class is by convention about moves which are either sente or reverse-sente. Personally, I have no problem with RevSenteXXX names (doesn't change a thing), but there might be (possibly good) reasons to keep it the way it is. Unless you have a more specific reason to introduce a new naming convention, maybe it's better to stick to the old one...

[gunnar]

Stylistic fixes.

[gunnar]

The attached patch al_3_7.10b makes the pattern comments follow the usual style. I also reverted the name change RevSente13. Just changing the name of one of the reverse sente patterns is meaningless and I don't see much gain in doing it consistently either. The comment for that class of patterns does say that it applies for both sente and reverse sente patterns.