UV editor "Select Overlap" doesn't select identical faces #85508

New Issue

Luka Glisic · 2021-02-09T19:43:43+01:00

Luka Glisic commented

2021-02-09 19:43:43 +01:00

System Information
Operating system: Windows-10-10.0.18362-SP0 64 Bits
Graphics card: GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.51

Blender Version
Broken: version: 2.93.0 Alpha, branch: master, commit date: 2021-02-08 23:42, hash: eab9165c25

Short description of error
When in the UV editor "Select Overlap" doesn't work if the UV islands are exactly the same.

Exact steps for others to reproduce the error
*Open attached .blend file
*In the UV Editor, Select -> Select Overlap

Bug.blend

**System Information** Operating system: Windows-10-10.0.18362-SP0 64 Bits Graphics card: GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.51 **Blender Version** Broken: version: 2.93.0 Alpha, branch: master, commit date: 2021-02-08 23:42, hash: `eab9165c25` **Short description of error** When in the UV editor "Select Overlap" doesn't work if the UV islands are exactly the same. **Exact steps for others to reproduce the error** *Open attached .blend file *In the UV Editor, Select -> Select Overlap [Bug.blend](https://archive.blender.org/developer/F9753635/Bug.blend)

Luka Glisic commented

2021-02-09 19:43:43 +01:00

Added subscriber: @LG787

Falk David commented

2021-02-10 10:52:35 +01:00

Added subscriber: @filedescriptor

Falk David commented

2021-02-10 10:52:35 +01:00

Changed status from 'Needs Triage' to: 'Confirmed'

Falk David commented

2021-02-10 10:52:35 +01:00

I can confirm this on the latest 2.92.0 Beta, branch: master, commit date: 2021-02-10 00:16, hash: f617782fc1.

I can confirm this on the latest 2.92.0 Beta, branch: master, commit date: 2021-02-10 00:16, hash: `f617782fc1`.

Falk David commented

2021-02-10 11:13:29 +01:00

Added subscriber: @ideasman42

Falk David commented

2021-02-10 11:13:29 +01:00

It looks to me as though the code explicitly excludes the case of endpoint overlap:

bool result = (
          /* Don't use 'isect_tri_tri_v2' here
           * because it's important to ignore overlap at end-points. */
          isect_seg_seg_v2_point_ex(t1[0], t1[1], t2[0], t2[1], endpoint_bias, vi) == 1 ||
          isect_seg_seg_v2_point_ex(t1[0], t1[1], t2[1], t2[2], endpoint_bias, vi) == 1 ||
          isect_seg_seg_v2_point_ex(t1[0], t1[1], t2[2], t2[0], endpoint_bias, vi) == 1 ||
          isect_seg_seg_v2_point_ex(t1[1], t1[2], t2[0], t2[1], endpoint_bias, vi) == 1 ||
          isect_seg_seg_v2_point_ex(t1[1], t1[2], t2[1], t2[2], endpoint_bias, vi) == 1 ||
          isect_seg_seg_v2_point_ex(t1[1], t1[2], t2[2], t2[0], endpoint_bias, vi) == 1 ||
          isect_seg_seg_v2_point_ex(t1[2], t1[0], t2[0], t2[1], endpoint_bias, vi) == 1 ||
          isect_seg_seg_v2_point_ex(t1[2], t1[0], t2[1], t2[2], endpoint_bias, vi) == 1 ||
          isect_point_tri_v2(t1[0], t2[0], t2[1], t2[2]) != 0 ||
          isect_point_tri_v2(t2[0], t1[0], t1[1], t1[2]) != 0);

Which is weird because it is unclear why. I would expect the operator to handle this case as well... Maybe it would be good to add an option like "Include boundary" (not the best name) so the user can choose to select what's "inside" or additionally what's "exactly overlapping".
@ideasman42 what are your thoughts here?

It looks to me as though the code explicitly excludes the case of endpoint overlap: ``` bool result = ( /* Don't use 'isect_tri_tri_v2' here * because it's important to ignore overlap at end-points. */ isect_seg_seg_v2_point_ex(t1[0], t1[1], t2[0], t2[1], endpoint_bias, vi) == 1 || isect_seg_seg_v2_point_ex(t1[0], t1[1], t2[1], t2[2], endpoint_bias, vi) == 1 || isect_seg_seg_v2_point_ex(t1[0], t1[1], t2[2], t2[0], endpoint_bias, vi) == 1 || isect_seg_seg_v2_point_ex(t1[1], t1[2], t2[0], t2[1], endpoint_bias, vi) == 1 || isect_seg_seg_v2_point_ex(t1[1], t1[2], t2[1], t2[2], endpoint_bias, vi) == 1 || isect_seg_seg_v2_point_ex(t1[1], t1[2], t2[2], t2[0], endpoint_bias, vi) == 1 || isect_seg_seg_v2_point_ex(t1[2], t1[0], t2[0], t2[1], endpoint_bias, vi) == 1 || isect_seg_seg_v2_point_ex(t1[2], t1[0], t2[1], t2[2], endpoint_bias, vi) == 1 || isect_point_tri_v2(t1[0], t2[0], t2[1], t2[2]) != 0 || isect_point_tri_v2(t2[0], t1[0], t1[1], t1[2]) != 0); ``` Which is weird because it is unclear why. I would expect the operator to handle this case as well... Maybe it would be good to add an option like "Include boundary" (not the best name) so the user can choose to select what's "inside" or additionally what's "exactly overlapping". @ideasman42 what are your thoughts here?

Jesse Yurkovich commented

2021-02-10 11:27:00 +01:00

Added subscriber: @deadpin

Jesse Yurkovich commented

2021-02-10 11:27:00 +01:00

I was the author of this functionality. I think this can be addressed by having an additional, final, check to see if all the verts of the 2 tris are equal to each other but I have not tried it out yet (will have to account for vertex ordering differences).

Generally we have to ignore colinear edges and endpoints because verts/edges are shared between faces normally and we shouldn't detect an overlap in those cases.

I was the author of this functionality. I think this can be addressed by having an additional, final, check to see if all the verts of the 2 tris are equal to each other but I have not tried it out yet (will have to account for vertex ordering differences). Generally we have to ignore colinear edges and endpoints because verts/edges are shared between faces normally and we shouldn't detect an overlap in those cases.

Philipp Oeser commented

2021-02-10 11:30:12 +01:00

Added subscriber: @lichtwerk

Philipp Oeser commented

2021-02-10 11:30:12 +01:00

Adding Modeling as the high-level module

Falk David commented

2021-02-10 11:36:01 +01:00

@deadpin Right that would make sense. I would still add it as an option, so the user can choose to do the additional computation only if they want to.

blender-admin commented

2021-02-10 22:27:47 +01:00

This issue was referenced by 2d252b6d26

This issue was referenced by 2d252b6d26f90f81f2a2dc7a3031c407dc8a643c

Campbell Barton commented

2021-02-10 22:36:06 +01:00

Changed status from 'Confirmed' to: 'Resolved'

Campbell Barton closed this issue

2021-02-10 22:36:06 +01:00

Campbell Barton self-assigned this 2021-02-10 22:36:06 +01:00

Sign in to join this conversation.

No Label

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

UV editor "Select Overlap" doesn't select identical faces #85508