Penalty for all enemy pawns xrayed by our bishop.

[xoto10]

Penalty for all enemy pawns xrayed by our bishop.

STC:
LLR: 2.93 (-2.94,2.94) {-0.50,1.50}
Total: 159760 W: 30229 L: 29813 D: 99718
Ptnml(0-2): 2659, 18309, 37534, 18713, 2665
https://tests.stockfishchess.org/tests/view/5eb1d5032326444a3b6d33ce

LTC:
LLR: 2.93 (-2.94,2.94) {0.25,1.75}
Total: 26496 W: 3908 L: 3656 D: 18932
Ptnml(0-2): 192, 2512, 7610, 2720, 214
https://tests.stockfishchess.org/tests/view/5eb1e2dd2326444a3b6d33f9

Bench 5185517

[xoto10]

The code here is the same as tested, I had intended to replace the attacks_bb call with PseudoAttacks after useful suggestion from 31m059 but it appears that I only put the new code into a comment.
I will update the PR with whichever version is preferred.

[Vizvezdenec]

I think I see attacks_bb(s, 0) for the first time, how does it work?

[xoto10]

You can see it used at the start of the for loop. That 0 is just an empty bitboard, it returns all attacked squares with attacks stopped by marked squares in the bitboard. I give it Bitboard(0) so it returns the full diagonals the bishop is on.
I was originally thinking of fianchetto bishops being blocked by e.g. c6 and d5 (we have Bg2) but perhaps it is useful in closed positions when a bishop can't really move much.

[mstembera]

attacks_bb(s, 0) is slower than PseudoAttacks[BISHOP][s] because it uses magic bitboards.

[vondele]

@xoto10 yes, please use the PseudoAttacks version (and verify same bench after the changes).

[vondele]

Thanks!