jf barraud wrote:
Hi,lt
About lpe-knot, I don't see where the trouble is. Is it the "new" entry in the result viewer? I guess it's there because I changed the testcase/reference to test new features (better closed path support and groups).
The problem is it's rendered differently from the way it used to be rendered (see attached files). And as I, personally, have zero knowledge of this feature I can't judge which is right and which is wrong. Obviously I could assume the (pass) reference is right (as it should) and the current output is wrong, but I thought the current output seemed a little too "possible" to simply assume this.
In other words: either we have a regression in Inkscape's implementation of this effect, or we used to do it wrong and the reference needs to be updated.
As for the non crossing lines in the bottom right of the reference image, they are intentional: lpe-knot let you leave crossings unchanged if you want... (yeah, you generally don't want, but who knows?...)
Is this an answer or does the test fail indeed?
You tell me :)
