After reviewing your previous post a little more closely, it seems everything is working as it's supposed to (as far as I can tell).
Depending on how you insert the edge (i.e. which options you chose on the split edge operator, and edge on mesh which operator is applied), the sample count will change by different amounts. It's very case-specific. Example: Given a grid comprised of a single 4-sided polygon, it will have 4 samples - one for each corner of the polygon. If you apply the "Split Edge with split Control" operator to one of the four edges, you'll increase the sample count by 1 sample as only a new vertex will be inserted, and that vertex needs to have it's own sample. Therefore, a total of 5 samples. If you select two edges as input to the "split edge with split control" operator, then you'll add 4 samples to the mesh for a total of 8 samples. Why? Because you're inserting a new edge to cut the polygon in half into two 4-sided polygons. Each 4-sided polygon has 4 samples, one for each corner of the polygon. Now if you deactivate the "connect" parameter in the split edge operator, you'll notice the sample count drops to 6 because only 2 new vertices will have been inserted into the mesh, not a new edge. If you repeat the procedure with an interior edge of a more complex mesh, you'll see similar case-specific additions/subtractions of samples. I don't know what your test mesh looks like, but based what I'm seeing in your pictures, I'm guessing you applied the split edge operator to a boundary edge? or didn’t activate the 'connect' option? Furthermore, I wouldn't go by that old thread you dug up from Michele and Kim. That pre-dates a lot of the architecture we're using now. Some changes were made under the hood to the modeling when ICE was introduced as well as many bugs fixed. They were likely using XSI v6.x, which had a ton of problems across the board. In that thread it is mentioned you must specify your operator as a topology operator. However, if you're using the self-installing plugin architecture, there is no way to make such declarations. It's implied. I think their conversation revolved around the older .spdl based plugins which required a lot of the parameters and attributes be declared ahead of time in a fixed format. The current architecture doesn't require that. Matt Date: Sun, 13 Nov 2016 17:40:56 +0300 From: Andrew Prostrelov <[email protected]> Subject: Re: Preserve UV methods after PolygonMesh.Set() To: [email protected] Ok i tried all over again. Tried different type of clusters and different type of port connection configurations. Clusters still do not updates components amount even if this clusters were 'always complete'. If you add new components (verts,edges, polys etc) via PolygonMesh.Set() they do not appear in clusters attached to CustomOperator. I'am also find this thread: https://groups.google.com/d/msg/xsi_list/y5bY5FAHOzk/qDCOcVHY9o8J Michele:" ... Complete clusters need to be applied to the object after the operators have been connected, for some unknown reasons that's a requirement...." I'm not sure what does Michele wanted to say. How should i connect additional port to CustomOperator if we already perform CustomOperator.connect() function ? So if XSI should automaticly update clusters connected to CustomOperator, it does not. It updates already existed components indices, but it doesn't get new components to clusters automatically. ------ Softimage Mailing List. To unsubscribe, send a mail to [email protected] with "unsubscribe" in the subject, and reply to confirm.
