You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
l1/r1 are equivalent to l0/r0 except the straight segments get converted to cubic segments by B.union. They draw and fill correctly, so I think it's a problem with (sample :: _ -> _ -> Crossings) rather than a problem with B.union.
One problem is that filter (liftA2 (&&) (>=0) (<=1)) in trailCrossings from Diagrams.TwoD.Path should be filter (liftA2 (&&) (>=(-eps)) (<=(1+eps)), but there is still another problem because given
crossingYIsMinus1 now produces the expected result, but crossingYIsPlus1 still does not.
edit: The param is also compared to 0 or 1 in signFromDerivAt. Loosening those comparisons fixes the example, but might possibly break other things. It requires further consideration.
bacchanalia
pushed a commit
to bacchanalia/diagrams-lib
that referenced
this issue
Apr 8, 2019
This is an attempted fix for issue diagrams#337. While it gets rid of some
instances where test falsely returns a zero value, it is possible it may
introduce some instances where test falsely returns a non-zero value.
l1/r1 are equivalent to l0/r0 except the straight segments get converted to cubic segments by B.union. They draw and fill correctly, so I think it's a problem with (sample :: _ -> _ -> Crossings) rather than a problem with B.union.
The text was updated successfully, but these errors were encountered: