Indigo 1.1.4

suvakas · Post by **suvakas** » Thu Jun 12, 2008 6:20 am

Here's the obj
(renders fine in 1.0.9)

Post by **OnoSendai** » Thu Jun 12, 2008 6:30 am

Yup, UV bug has been squashed. *crosses fingers*

fishbowl · Post by **fishbowl** » Thu Jun 12, 2008 12:22 pm

Seems to exit if an object has no UV set on exporting from Blendigo. "UV set with index 0 was exposed, but the UV set data was not provided." Is this Blendigo or Indigo causing that error?

cpfresh · Post by **cpfresh** » Thu Jun 12, 2008 2:13 pm

probably blendigo. i found a similar error in our skindigo. the solution was a manual edit of the igs file to remove any extraneous uv set entries.

fishbowl · Post by **fishbowl** » Thu Jun 12, 2008 2:50 pm

Thanks, that did the trick.

This block:

Code: Select all

<expose_uv_set>
<index>0</index>
<name>uv</name>
</expose_uv_set>

...seems to cause the error.

BbB · Post by **BbB** » Thu Jun 12, 2008 7:24 pm

That's one of the two problems I was having. But with something like 500+ objects in the scene, hand-editing is no option. Hopefully the next Blendigo will fix it.

suvakas · Post by **suvakas** » Thu Jun 12, 2008 7:36 pm

I get the same kind of errors from Max too. Now who is going to fix it for Max?

BbB · Post by **BbB** » Thu Jun 12, 2008 7:43 pm

You?

EDIT: Suv, you also getting the "normal is not unit length" red flag? And the sudden, unexplained crashes with no messages?

Ono: Remind me what the point is of making Indigo more intolerant of sloppy meshes? Isn't that the opposite of making your app more robust? I don't care either way if the exporters are rewritten so that they fix the problem. But as a tech-idiot, I'm just curious...

Post by **OnoSendai** » Thu Jun 12, 2008 7:49 pm

BbB:
Well, for one thing, it helps exporter writers to catch issues in their code.
An example is non unit-length normals.
If you have a normal with length != 1, the normal is probably pointing in completely the wrong direction, for example due to using incorrect transformations on it.
It's better to fix the problem than to sweep it under the rug

BbB · Post by **BbB** » Thu Jun 12, 2008 7:51 pm

But who cares if it renders fine?

The scene I'm cooking now looks great in 1.0.9 but won't get passed 1.1.2 and upwards because of these. It's probably me being stupid but I just don't see the benefit... Unless this unit-length thing is crucial to displacement...

Post by **OnoSendai** » Thu Jun 12, 2008 7:53 pm

But it won't render correctly if the normals are pointing in the wrong direction

BbB · Post by **BbB** » Thu Jun 12, 2008 7:55 pm

Well, it depends on the material you're using, right? Glass and emitters won't, but pretty much everything else will, will it not?

Post by **OnoSendai** » Thu Jun 12, 2008 7:56 pm

No, if the normals are wrong, and the shading normals are used (normal_shading = true), every material will render incorrectly.

BbB · Post by **BbB** » Thu Jun 12, 2008 7:59 pm

Alright boss. Hope we get a new Blendigo soon then. And I promise I'll be more disciplined in applying transformations.

I read somewhere that Chema had abolished the need for correct normals in his latest Fry code. Know anything about this?

Post by **OnoSendai** » Thu Jun 12, 2008 8:28 pm

nope.

And I don't think you'll have to 'apply transformations' anymore.