I recently added some new capabilities to farray slicing, but unit tests run about 20s slower on commit 19cbb53f than commit af162061. I traced the difference down to the Sudoku solver. The _soln2str method has to slice all 729 points in the solution for every puzzle.
Bottom line: some optimization of __getitem__ would be beneficial, especially around the _iterdim function.
I recently added some new capabilities to
farray
slicing, but unit tests run about 20s slower on commit19cbb53f
than commitaf162061
. I traced the difference down to the Sudoku solver. The_soln2str
method has to slice all 729 points in the solution for every puzzle.Bottom line: some optimization of
__getitem__
would be beneficial, especially around the_iterdim
function.