LDraw.org Discussion Forums
Auto correcting (some) defects in the official library - Printable Version

+- LDraw.org Discussion Forums (https://forums.ldraw.org)
+-- Forum: General (https://forums.ldraw.org/forum-12.html)
+--- Forum: Official File Specifications/Standards (https://forums.ldraw.org/forum-32.html)
+--- Thread: Auto correcting (some) defects in the official library (/thread-14597.html)

Pages: 1 2

Re: Resulting patch v2 (was Re: Auto correcting..) - Roland Melkert - 2014-12-23

Thanks Chris.

Re: Resulting patch (was Re: Auto correcting..) - Magnus Forsberg - 2015-01-19

Hi Roland,

May I bring your attention to the torso 973p71 and p72. They share a common pattern subfile s\973p71b.dat included in the batch of fixed files.
Something has gone terribly wrong in the libfix process in LDCad. Don't know if it matters, but it might be interesting anyhow.
Is looks as if it tried to fix concave quad, but failed, and made it worse. The 'fixed' file still have the same number of concave quad, but now have the vertices in another order.
I'll start to fix it, since both torsos needs to become BFC compliant anyway.

   

Re: Resulting patch (was Re: Auto correcting..) - Roland Melkert - 2015-01-19

Quote:Is looks as if it tried to fix concave quad, but failed, and made it worse. The 'fixed' file still have the same number of concave quad, but now have the vertices in another order.
It thinks it's a twisted quad because the algorithm tests this by splitting the quad in two triangles which in case of a concave can end up having different windings. It then swaps some points and tests again, the first order where both triangles have the same winding wins but in case of a concave nothing is really fixed Smile

So what's really lacking here is a concave detection instead of a twisted quad detection.

Thanks for spotting/manually fixing this, please remove the libfix history line from this file as it didn't help in anyway.