search

pdf-association / pdf-issues

Industry-based resolutions for issues and errata reported against any PDF-related specification
https://pdf-issues.pdfa.org/
67 stars 2 forks source link

PRC ill-definition of face_number in 8.8.9.6.1 #429

Open mvrhel opened 6 months ago

mvrhel commented 6 months ago

The compressed tessellation object includes compression of the face normals. The specification states "face_number is derived from the maximum value in triangle_face_array" but does not tell how it is derived, nor what the use of this value is. I suspect that one looks through the triangle_face_array values and picks the max and adds one to that to get the number of faces. Then this value is used for the number of elements in the is_face_planar array. But this is just conjecture and needs clarification in the spec. In a similar manner, the is_point_color_on_face array size, is_multiple_line_attribute_on_face array size, and face_has_texture array size are not well defined in the same object in, but I am guessing they also use the derived face_number. Again, clarification on this and how the face_number is derived rather than just saying it is derived would be good.

t3dr0 commented 4 days ago

Alternate reference location is section

7.8.9.7.1 Description of the Data Written to the File

Note face_number is not present in the following Table 174. Looks like it is not part of the file data structure? I speculate this may be a local loop variable going from 0 to len(triangle_face_array)-1. Summary: Needs more investigation.