Closed jonnor closed 9 years ago
Should either be done by blitting the filtered (fullsize) image to another canvas in the output phase, before saving to PNG. Or by resizing on the input side. The latter is more efficient, but could be problematic when we want to do cropping/area extraction, because we want that to be in input-referred coordinates...
Fixed in commit 422ce2009dc0e56077719b4ae1686abe09dbe1be
Must rescale the image to be the provided dimensions, like GEGL/imgflo graphs do