Currently the situation is such that for moves like [N-B] a dummy mp is inserted when converted to a notation (mpafsyafW), and only pure mp steps can be merged with a following step. (So not mpo.) This would not cause the initial knight jump to be blocked by a hole on the W squares. But Knight moves intentionally defined as mpafsW to make them 'lame w.r.t. holes', and allow a single square separator betwen boards would then no longer work.
I have thought about changing all this to mpo. But that would then fail to do the merging in hand-defined mp legs for the purpose of defining partially lame moves on a board that had no holes. Now this might not be so much of an issue, because the only difference between merging and not merging is exemption of the p from a hop ban specified in the captureMatrix. Which again is important only when other p legs have to be exempted.
I could of course merge both mp and mpo legs. Or let the parser make a choice depending on whether the board contained 'isolated' holes to be ignored by leapers, or that the holes are used for splitting into multiple boards. That would be a cumersome decision to make automatically, though.
Currently the situation is such that for moves like [N-B] a dummy mp is inserted when converted to a notation (mpafsyafW), and only pure mp steps can be merged with a following step. (So not mpo.) This would not cause the initial knight jump to be blocked by a hole on the W squares. But Knight moves intentionally defined as mpafsW to make them 'lame w.r.t. holes', and allow a single square separator betwen boards would then no longer work.
I have thought about changing all this to mpo. But that would then fail to do the merging in hand-defined mp legs for the purpose of defining partially lame moves on a board that had no holes. Now this might not be so much of an issue, because the only difference between merging and not merging is exemption of the p from a hop ban specified in the captureMatrix. Which again is important only when other p legs have to be exempted.
I could of course merge both mp and mpo legs. Or let the parser make a choice depending on whether the board contained 'isolated' holes to be ignored by leapers, or that the holes are used for splitting into multiple boards. That would be a cumersome decision to make automatically, though.