One of these I have corroborated (the Edwards TB ep code). Another I currently have doubts about, as the offsets in the structs in Rybka and Crafty do not match. This might because we haven't found the right Crafty version yet, or because this Rybka version deleted some of the struct members. Or it might be that Rajlich managed to accidentally make a typo somewhere in the pawn_hash clearance, just as Bob did -- coincidences do happen.it has transpired that there is evidence that pre-Fruit Rybkas contained significant portions of Crafty,
I think Zach has found more evidence of re-use of Crafty code, but it is not public yet. Once the extent of any copying is understood, the weightiness of the copied elements must be explored. It seems that most of it (currently) involves a "Crafty chassis" for more mundane things, while search/eval seem to have been replaced, at least to some degree. I'm not sure (say) whether CCT6 would have accepted Rybka if the extent of Crafty overlap had been known -- that's probably just a hypothetical inquiry. Recall also that this was merely a "private" version of Rybka, so having Crafty parts might not be so derelict for some purposes. All of this is somewhat tangential to the Fruit/Rybka story per se, though it obviously has some circumstantial impact.
