@ -1,3 +1,6 @@
- looks like we have a race in tiled threadcache? see
https://github.com/jcupitt/libvips/issues/347
- convert float mask to int for int convolution has rounding problems
The note is not visible to the blocked user.