neomonkeus
commented
8 years ago Did some actual testing. Yep, looks like the original implementation was completely broken. If the update_tangent method had been actually called for either branch, it would have failed with an AttributeError because ExtraVectorsFlags doesn't exist in the model. The reason being that the nif.xml submodule version (0.7.1.0) was never updated to point to the correct updated version (0.7.1.1).
Before this can be merged I need to add an additional commit to update the submodule so that the versioning is correct
Memory dump on object with implementation as is
Checking data
<class 'pyffi.formats.nif.NiTriStripsData'> instance at 0x0A6C1940
* unknown_int : 0
* num_vertices : 6
* keep_flags : 0
* compress_flags : 0
* has_vertices : True
* vertices :
<class 'pyffi.object_models.xml.array.Array'> instance at 0x0A67FB88
0: [ -84.730 63.487 220.734 ]
1: [ -84.730 105.453 220.777 ]
2: [ -84.730 40.911 220.734 ]
3: [ -84.730 80.577 220.734 ]
4: [ -84.730 105.453 220.777 ]
5: [ -84.730 63.487 220.734 ]
* num_uv_sets : 0
* bs_num_uv_sets : 4096
* skyrim_material : SKY_HAV_MAT_LIGHT_WOOD
* has_normals : True
* normals :
<class 'pyffi.object_models.xml.array.Array'> instance at 0x0A67F9A8
0: [ 1.000 -0.000 0.000 ]
1: [ 1.000 -0.000 0.000 ]
2: [ 1.000 -0.000 0.000 ]
3: [ 1.000 0.000 -0.000 ]
4: [ 1.000 0.000 -0.000 ]
5: [ 1.000 0.000 -0.000 ]
* tangents :
<class 'pyffi.object_models.xml.array.Array'> instance at 0x0A67FC28
0: [ -0.000 -1.000 -0.000 ]
1: [ -0.000 -1.000 -0.000 ]
2: [ -0.000 -1.000 -0.000 ]
3: [ 0.000 -1.000 -0.000 ]
4: [ 0.000 -1.000 -0.000 ]
5: [ 0.000 -1.000 -0.000 ]
* bitangents :
<class 'pyffi.object_models.xml.array.Array'> instance at 0x0A67FC78
0: [ -0.000 0.000 1.000 ]
1: [ -0.000 0.000 1.000 ]
2: [ -0.000 0.000 1.000 ]
3: [ 0.000 -0.000 1.000 ]
4: [ 0.000 -0.000 1.000 ]
5: [ 0.000 -0.000 1.000 ]The implementation gets away with this because it is just an assignment, but because the attribute isn't a modelled attribute it won't get written to memory. I have been able to run the scenario on the edge mesh and can see it has no affect, even with the current implementation without the updated nif.xml
Dexesttp
Scope
As from Pull request #15 , the
Num UV Setsflag from theNiGeometryDataobject switched from a bitflags type to an enum, and changes were made accordingly to the PyFFI file format interface.This flag handles the
TangentsandBitangentsmembers existence in the object, and could take a value &-able with 16 to show that the members should exist, and any other otherwise.Issue
A problem was found where an
UV Setwas added and thetangents/bitangentswere automatically recalculated. However, the data was never saved to the file.Solution
This change handled the case in which UV sets didn't exist by setting the new enum to 0. However, it didn't handle the case when an UV value was added.
The pull request solve this case by setting the flag to 16[
Tangents_Bitangents] when an UV_set is added. This way, the tangents and bitangents can be computed and stored in the object, and it will be written in the file down the line.Extra information
The structure and expected data of the
UV Setsmember was described here : http://niftools.sourceforge.net/forum/viewtopic.php?f=10&t=321As PyFFI seems to only use the first UV set, it may be an incomplete feature that might need fixing later.