Fix capacity overflow in offset path node #3308
Conversation
|
I don't love lowering of the accuracy, is that a necessary change? |
…m/ashishmohapatra240/Graphite into fix/offset-path-capacity-overflow
|
i initially relaxed the accuracy from 1e-3 to 1e-2 thinking it would reduce recursion depth in fit_to_bezpath, reverted to 1e-3. |
|
Hey, @TrueDoctor Let me know if this works! |
|
simplified with |
| let is_degenerate = | ||
| (cubic_bez.p1 - start).hypot() < MAX_ABSOLUTE_DIFFERENCE && (cubic_bez.p2 - start).hypot() < MAX_ABSOLUTE_DIFFERENCE && (cubic_bez.p3 - start).hypot() < MAX_ABSOLUTE_DIFFERENCE; |
There was a problem hiding this comment.
Could you add a quick comment about when this is true and why we return None in those cases? Also could you use start.distance(cubic_bez.p1) instead to make this more readable? If we introduced a squared constant, we could also use start.distance_squared(cubic_bez.p1) which would avoid the cost of computing the square root
There was a problem hiding this comment.
added comment explaining degenerate curves and switched to distance_squared(), please check.
2 participants
Fixes #3288
Prevents infinite recursion in
fit_to_bezpathby:Graphite.mp4