Inside Corners of Hollow Disappearing?

Prim-stingy me is trying to texture the inside of a hollowed megaprim. I'm getting the number of repeats about right thanks to other information on this forum but it looks like at the corners part of the texture disappears! It's not continuous from one wall to the next. Is this something that happens with all prims or just megas?

(15x15x5 megaprim hollowed 95% with texture horizontally repeated and offset to get 6 of the texture on each side)

ETA: Just to clarify, it's a window texture with solid wall on each side and at the corners the window glass from one repeat of the texture is right next to the glass of the next repeat with no intervening wall.

It's VERY difficult to get the repeats and offsets for textures on the inside of a hollowed prim to come out right, and the larger the prim is, the more obvious any miscalculation will be. Take a look at , in which you'll see the following....



the hollow parameter is a decimal equivalent of the hollow percentage. Her formulas will give you very nearly the perfect repeats and offsets, but you may have to do a tiny bit of manual tweaking. You may not have enough significant figures in the numerical fields of the texture page to get it perfect, even at that. Good luck with it.

BTW, remember also that if you use a 32-bit texture on both the inside and the outside of your prim, as I suspect you might be since you are making windows, you may have alpha sorting problems.

Thanks, Rolig, but that's not exactly the puzzle I'm puzzling out right now. I'm finding that at 95% hollow the texture doesn't wrap continuously around the inside corners; part of it disappears! Just now I was playing with it some more and, oddly, at a lower hollow percentage that doesn't seem to happen or at least not as much. Rather than continue fighting the whole hollow prim issue I'm going to go with separate, prim-limit-gobbling walls Ah, the mysteries of building in SL!

I'd be interested in knowing the parameters of the misbehaving mega, as I haven't see that issue yet...

you might also cheat, and use 2 megas that are half cut to force it and only lose one extra prim. regardless, post back because I'm working of several build related jiras and might as well include that one (especially if I can generalize it down to the accepted prim limits)

It doesn't appear to be a Mega issue after all. I'm seeing the same thing when I repeat the same texture settings using a conventional 10x10x5 prim. The images below are a 15x15x5 Mega I got as part of a large package. Is it OK if I state the creator's name here?

ETA: Hmmm...normal mapping doesn't do it. Seems to be a Planar issue. And please don't laugh at the ugly texture. It's just a stand-in for testing until I get the build finalized

First is the 15x15x5 hollowed to 95 percent with the texture settings shown:


Then I reduced the hollow to 83.5 percent and got rid of the offset adjustment:

I see what you're talking about now. Since you mention that it's only a problem with planar mapping, though, I have to ask why you want to use planar mapping for a setting like this anyway. I would have thought that normal mapping would do the job well. Frankly, I always find the math for planar mapping confusing, so I avoid it unless I need it for something like shingling a roof.

I don't remember exactly why I switched to planar from default. No doubt wine was involved. I may have been thinking plugging in repeats per meter would work better than repeats per surface and then the Group Chat got amusing and I forgot to change it back In any case I'm going the easier and apparently more effective route. Thanks for taking the time to look.

this is a bug.. 2 opposing corners do this while 2 corners are ok. In the 2 bad corners, One of the corner edges extends the texture all the way to the outside edge as if it were 100% instead of the 95% it is supposed to . This happens on a regular 10X10 prim as well.

Also not an alpha problem as the 24 bit version of a texture does the same thing..

mkay, I misinterpreted the problem, and yaeh, my original formula was meant to solve that, (it's a hidden edge problem, that one corner extends all the way to the outside face(s) regardles of not being cut at it's corner)

it's just a bit of nonsense from how the surface it calculated (essentially having a leading and trailing edge that's buried out of sight in the begin/end position)