Jump to content

arton Rotaru

  • Content Count

  • Joined

  • Last visited

Community Reputation

989 Excellent

About arton Rotaru

  • Rank
    Cost Effective Resident

Recent Profile Visitors

1,563 profile views
  1. You're welcome! It might be worth mentioning, that there is yet another cause for meshes changing there physics to convex hull like behavior. When the mesh has a non-analyzed (a triangle based) physics mesh, and something is linked to it that is using a TARGET_OMEGA function (a door for example). This will make the whole built acting convex hull like when the target omega function is called. It will revert back to normal prim type physics when it's selected as well. This can be worked around by decomposing (analyze) the physics mesh during upload. (Or by unlinking the door, or by avoiding
  2. You are not alone with this. Physics Setting for mesh object reverting on rez
  3. You can create an emissive mask for a diffuse texture. The mask will go into the alpha channel of the texture, just like what we do for transparency. Emissive is not what we can set as glow in the build floater though. Emissive is what we know as Fullbright. White in the alpha channel is 100 % Fullbright, black is 0%. And all the greys inbetween make the rest the brightness. So you just paint the areas which should have the "glow" more or less white, and everything else just plain black. Copy that layer into an alpha channel, save as 32 bit Targa (.tga). In-world you set the Alpha mode of
  4. Actually, the importer in Second Life does support this. The question is, if Blenders collada exporter does support this still. At least it used to work in old Blender versions.
  5. Actually, the issue with material names was the space in the default name. Not the # sign. This has been fixed a couple of years ago. I know because it was me who reported it 5 years ago. (Well, actually Whirly was so kind to file the jira for me, I just gave her some test files). What fixes broken UVs for me is indeed triangulating the mesh before exporting it. Changing the material names does nothing for me in this regard. I would recommend triangulating meshes anyway. I certainly don't want a third party triangulator mess with a model I spent a couple of hours on to create it. With
  6. To avoid crazy UVs I usually have to triangulate the mesh before I export it as dae. I use the Turn to Poly modifier with Limit Polygon Size to 3, and Keep Polygons Convex checked. You can also try to add a Turn To Mesh modifier, on top of the stack, instead. I also keep the Triangulate option checked in the export settings. And as Tonk mentioned, Reset XForms should be applied anyway.
  7. The SL mesh importer sorts Materials (Faces) alphabetically by their names.
  8. I suspected that. Then you need to follow a specific naming convention for your LoD meshes. Otherwise the uploader mixes up the meshes, and the tire will be assigned to the rim and vice versa. SL naming convention. But, I would recommend to upload such a wheel as one mesh only. If you want to animate the wheel on a vehicle, it's much easier when all the spinning parts are just one mesh. Also make sure you apply Rotation & Scale (Reset XForms, Freeze Trasformations etc.) to your LoD meshes as well.
  9. Is this one object, or is it multiple objects?
  10. Sounds familiar to me. This happens always with meshes which have a triangle-based physics shape, and have linked something to it that use TARGET_OMEGA to rotate/spin something smoothly. Which most doors do actually. There are 2 options to fix this. 1) Uploading the mesh with a decomposed physics mesh instead of a triangle-based. Or 2) unlinking the rotating parts from the house. Now the bad news. There are also some rare occasions where meshes behave like this, without anything linked, or script functions attached.
  11. #4 wouldn't help, because I did already.
  12. Yeah, these changes will be in the next LMR RC viewer. I guess it will be released next week. Perhaps there will be some more changes in this next release, IDK. The code Henri already pulled in his latest release is from 3 days ago. Judging from the images, it looks too dim again. I'll wait for the next LMR viewer to check it out. At least there is progress. We went from too dim, to too strong and, now something inbetween. Looks promising that we will be getting there.
  13. I feel pretty confident that the Devs at Linden Lab will work it out.
  • Create New...