MtI 2.x beta 2

[CTZn]

I figured out that if I freeze transforms pivot position doesn't change.

Freeze Transforms is baking everything into the transformation matrix (including pivots position) while transformations are all set to zero, so you don't see a change.

MtI is not changing the Maya behaviour in this regard, can you clarify the issue you noticed please SreckoM ? I've read many times that freezing transforms was the Evil, maybe you found out why

[SreckoM]

1. Create cube.
2. Reposition pivot.
3. Make several instances of cube.
4. Export to Indigo. You'll see that pivot of cube is back to original position. And that instances are not in positions that are in viewport.

That is what I discovered.

[EDIT] Everything is OK when rendering with Maya Software, Mental Ray or Vray.

[CTZn]

Thank you, helpfull bug report.

[Lemo]

Good work with the matrix neo, now my scaled instances problem is gone

The scale is okay but there is still that issue with hidden instances I'm affraid
In the following test scene with two spheres, when I hide only the "master" instance, I don't have any sphere at rendering.
And when I hide only the "slave" instance, I have two spheres at rendering.
I also had that pivot problem SreckoM describes at first with the cubes under the spheres.

The new "one click away" installer is great

Nice flowers SreckoM

[CTZn]

Nice Lemo.

I took the red pill indeed, and now I'll follow these rabbits !

I'd like to raise a few questions regarding MtI orientation... I'd rather fix first.

[CTZn]

By the way I just realised my mistake... I forgot two lines wich do activate way more junk output than one could expect. I never meant to expose this intimate and delicate... blah

[CTZn]

And... triple post, you can't beat me banzaï... oops.

Updated for the matter. I'm also changing the way files are named, there might have happened stuff with web browsers and caches leading into confusion. Like using the same file name for consecutive test versions.

MayaToIndigo 2.x beta 2 "Tamed Edition" test 5

minor fix

- menus twitbits
- removed awful debug output
- Initialize on Scene > Open if the scene is using Indigo to render
- still camera now to use <keyframe> (meh otherwise its upside down:P)

[ior]

Maya crash when I open scene file made with MtI_2.x_beta_2 in MtI_2.x_beta_2_5 installed version.

I reinstalled the older version and it didn't crash anymore.

here is the crash dumpdata:

dumpdata.zip

(52.47 KiB) Downloaded 262 times

[CTZn]

Someone else bumped into this ?

Did this happen with all scenes created with the older MtI versions or with one special scene ? Did this happened if you closed the two MtI windows before opening the scene ?

ior, the dump data is not relevant to MtI; however I have an idea on a plausible cause. Since I can not seem to reproduce the crash, you will debug this yourself The pay off might be an inmediate fix.

1) reinstall beta 2 test 5
2) find the file mti_CHK_self.mel that was just installed (in 'user/Documents/Maya/scripts'), open it in any editor
3) into the block global proc mti_init() (around line 219), find:

Code: Select all

	mti_newSceneJob;
	mti_sceneOpenedJob;

Comment the two lines out (write "//" (two slashes) before the command) and save the file.
4) Restart Maya (!) and let me know.

Normally the MtI windows present should just close, and Maya should not crash

- - - -

I have fixed the two issues Lemo and SreckoM submitted (pivots reset on export, wrong identification of visible instances) and I finally did a decent UI for textures, the former was some rigid auto layout. I'm at tinkering around a burn (now "convert to file") bug and will update then.

edit: in fact I'm rewriting the texture pipeline, it deserved this.

[ior]

It continues to crash even with the oldest scene file created with MtI_JAN-2010.


output window says:

Code: Select all

TsampleManager::setLastBatch() failed on
shading group swatchShadingGroup:
Postcondition violated, fLastBatchCache is NULL.

and script editor says:

Code: Select all

// Error: line 0: The indigoshader "glass" has no 'mat' attribute. //
// Error: line 0: Node "glass" has no attribute "mat". //

And it repeats the sentence for every material in the scene and for each attribute,
were "glass" is the material name.
and 'mat' can be 'c', 'bem', 'em', 't', 'mp', 'mior', 'mcb', 'msss', 'mssss', 'mgs', 'st', 'tl', 'ge', 'oc'.


Can this be a plugin conflict?

[SreckoM]

Happens over here too. Can not open older scenes for some reason.

[CTZn]

Did you try the edition of mti_CHK_self.mel as posted above ?

It seems like some conflict preventing a correct initialisation, yes. ScriptJobs are candidate for an uncontrolled execution order, hence the suggested manip.

Do you have the indigoPlugin set to auto-load in the plugin manager ? Try to load it before the older scene.

I'll finish with textures, then if the issue is not something that was fixed meanwhile I'll be on it at 100%. I tried to provoke it with no chance.

Update in about 48h.

[ior]

It is ok now, we have to load the plugin after the installation for it to work properly.
I thought that it was already loaded. My mistake. Thanks.

[CTZn]

Thank you for reporting, ior. This means the mistake is mine tho

Known issue with displacement (since b2t4 - Qr matrices support):
you must currently freeze the scale of displaced meshes, or beware that normals in Indigo will be scaled with it. We are looking into a fix.

I'm at checking MtI edits for the next hours.

[CTZn]

Update

Hot one:

• fixed bug in obj export procedure
  fixed bug in instances pipeline
  fixed motion blur
  splitted objects and camera sub frames attributes
  rewrote texture pipeline (5x less code)
  texture UI
  'merge all edges' option (mesh)
  misc fixtures

Tell me if that scene load issue still happens, I'll try fix it asap.

edit: re-updated:
last minute fixes

- fixed frenetic initializations just introduced
- ui fixtwixes
- more annotations (camera, textures)
- set "merge all edges" to be the default for meshes. Prevents totally displaced meshes from exploding.