new Bmesh boolean is deleting some of edges data, (Creases and Bweigh) #47129

New Issue

Adrian Rutkowski · 2016-01-06T20:43:15+01:00

Adrian Rutkowski commented

2016-01-06 20:43:15 +01:00

linux 64

Blender Version
7e6d096

new Bmesh boolean is deleting some of edges data, (Creases and BWeight), even though it keeps others. Its especially problematic for Bevel Weight since it's often set before applying boolean.

Old(left) vs new(right) behavior

linux 64 **Blender Version** 7e6d096 new Bmesh boolean is deleting some of edges data, (Creases and BWeight), even though it keeps others. Its especially problematic for Bevel Weight since it's often set before applying boolean. Old(left) vs new(right) behavior ![Screenshot from 2016-01-06 20-29-04.png](https://archive.blender.org/developer/F271708/Screenshot_from_2016-01-06_20-29-04.png)

Adrian Rutkowski commented

2016-01-06 20:43:15 +01:00

Changed status to: 'Open'

Adrian Rutkowski commented

2016-01-06 20:43:15 +01:00

Added subscriber: @niewinny

blender-admin commented

2016-01-06 20:43:15 +01:00

#64497 was marked as duplicate of this issue

Campbell Barton was assigned by Sergey Sharybin

2016-01-07 10:08:27 +01:00

Sergey Sharybin commented

2016-01-07 10:08:27 +01:00

Added subscriber: @Sergey

Campbell Barton commented

2016-01-08 15:06:05 +01:00

Could you attach an example file? - I just tested this and it worked.

Note that this may not be considered a bug - if mesh that the boolean modifier is on doesn't have the custom-data to begin with (booleans also wont add vertex colors, UV's etc if the mesh doesn't already have them).

Could you attach an example file? - I just tested this and it worked. Note that this may not be considered a bug - if mesh that the boolean modifier is on doesn't have the custom-data to begin with (booleans also wont add vertex colors, UV's etc if the mesh doesn't already have them).

Adrian Rutkowski commented

2016-01-08 21:17:52 +01:00

Hey Campbell,

It looks like I provided a little missliding info. It do not work only if applied from below other modifier(this is important, bevel modifier for me).

so here are the screencast showing it,so You can replicate it.
first the new one.

bug1.webm
(or link https://photos.google.com/album/AF1QipPnGwacsMNP7Vdor44b4b2DgHUJaMGhVcjgFZK6/photo/AF1QipNk-JS3NNqrSLRGV6t4Oq3RreuXosMWQHGuVhql)

and the old one

bug2.webm

(or link https://photos.google.com/album/AF1QipPnGwacsMNP7Vdor44b4b2DgHUJaMGhVcjgFZK6/photo/AF1QipPywQWz5C1l8elQYKCQhylAohViyj-6EEkQFHfK)

as You can see the bevel weight(but not only) is not preserved anymore. It would be nice to have it but keep it mind that I have no idea how new bmesh mod works, it is still superior to old one though even with no edge layers save))

I used new buildbot cd6c6ee aswell, same thing linux 64

Hey Campbell, It looks like I provided a little missliding info. It do not work only if applied from below other modifier(this is important, bevel modifier for me). so here are the screencast showing it,so You can replicate it. first the new one. [bug1.webm](https://archive.blender.org/developer/F272052/bug1.webm) (or link https://photos.google.com/album/AF1QipPnGwacsMNP7Vdor44b4b2DgHUJaMGhVcjgFZK6/photo/AF1QipNk-JS3NNqrSLRGV6t4Oq3RreuXosMWQHGuVhql) and the old one [bug2.webm](https://archive.blender.org/developer/F272053/bug2.webm) (or link https://photos.google.com/album/AF1QipPnGwacsMNP7Vdor44b4b2DgHUJaMGhVcjgFZK6/photo/AF1QipPywQWz5C1l8elQYKCQhylAohViyj-6EEkQFHfK) as You can see the bevel weight(but not only) is not preserved anymore. It would be nice to have it but keep it mind that I have no idea how new bmesh mod works, it is still superior to old one though even with no edge layers save)) I used new buildbot cd6c6ee aswell, same thing linux 64

Bastien Montagne commented

2016-01-15 14:38:46 +01:00

Added subscriber: @mont29

Campbell Barton removed their assignment 2016-08-03 06:35:14 +02:00

Campbell Barton commented

2016-08-03 06:35:14 +02:00

Added subscriber: @ideasman42

Cirno commented

2018-03-07 22:20:09 +01:00

Added subscriber: @Lambdadelta

Cirno commented

2018-03-07 22:20:09 +01:00

Bevel weight still not preserved after bmesh boolean. This is especially annoying after removing carve solver, which previously can be used instead

Ray molenkamp commented

2018-03-07 22:40:47 +01:00

Added subscriber: @LazyDodo

Ray molenkamp commented

2018-03-07 22:40:47 +01:00

please attach a sample file (.blend file, not a video) showing the issue

marking as incomplete until additional information is supplied.

please attach a sample file (.blend file, not a video) showing the issue marking as incomplete until additional information is supplied.

Cirno commented

2018-03-07 22:50:24 +01:00

Bmesh boolean bug.blend

If you apply boolean then bevel weight disappear. I tested it on blender-2.79-737a5ef-win64-vc14

[Bmesh boolean bug.blend](https://archive.blender.org/developer/F2430400/Bmesh_boolean_bug.blend) If you apply boolean then bevel weight disappear. I tested it on blender-2.79-737a5ef-win64-vc14

Ray molenkamp commented

2018-03-08 02:30:14 +01:00

can confirm the weights disappear

Adrian Rutkowski commented

2018-03-08 02:36:12 +01:00

since that was not mentioned,
creases disappear as well

since that was not mentioned, creases disappear as well

Brecht Van Lommel commented

2019-05-11 21:23:47 +02:00

Added subscriber: @WitoldJaworski

Adrian Rutkowski commented

2019-08-28 16:05:06 +02:00

since there is something happening here

that was fixed at some point
as long as all models (active one and the one we boolean with ) got edge data enabled

it will no longer clear edges

since there is something happening here that was fixed at some point as long as all models (active one and the one we boolean with ) got edge data enabled ![blender_2019-08-28_16-03-36.png](https://archive.blender.org/developer/F7707801/blender_2019-08-28_16-03-36.png) it will no longer clear edges

Philipp Oeser commented

2019-11-29 14:36:34 +01:00

Added subscriber: @lichtwerk

Philipp Oeser commented

2019-11-29 14:36:34 +01:00

Not sure if we can close this?
(confirmed to work with storing the data on all participating meshes...)

Not sure if we can close this? (confirmed to work with storing the data on all participating meshes...)

Anthony Edlin commented

2019-11-30 17:54:44 +01:00

Added subscriber: @AnthonyEdlin

Anthony Edlin commented

2019-11-30 17:54:44 +01:00

Although it works as long as you check store data on the boolean mesh target, it seems like the system should be able to handle combining smartly.

I look into it a little bit and notice boolean code uses BM_mesh_bm_from_me first on boolean target mesh and then second on it's own mesh. This is used to merge the two meshes into one bm before intersecting. But if you look in BM_mesh_bm_from_me you can see that the function BM_mesh_cd_flag_apply that adds the crease and weight layers is only called if it is the first mesh used. This means that those layers only exist if the first mesh (in this case the boolean target) has those layers. I added code to ensure those layers exist also if the second mesh has those layers. Here is patch: #47129.diff

It seemed to fix the issue for me in my limited testing. I am totally new to bmesh code though, so don't really know what I'm doing.

Although it works as long as you check store data on the boolean mesh target, it seems like the system should be able to handle combining smartly. I look into it a little bit and notice boolean code uses BM_mesh_bm_from_me first on boolean target mesh and then second on it's own mesh. This is used to merge the two meshes into one bm before intersecting. But if you look in BM_mesh_bm_from_me you can see that the function BM_mesh_cd_flag_apply that adds the crease and weight layers is only called if it is the first mesh used. This means that those layers only exist if the first mesh (in this case the boolean target) has those layers. I added code to ensure those layers exist also if the second mesh has those layers. Here is patch: [#47129.diff](https://archive.blender.org/developer/F8181715/T47129.diff) It seemed to fix the issue for me in my limited testing. I am totally new to bmesh code though, so don't really know what I'm doing.

Jeroen Bakker commented

2020-01-28 09:36:19 +01:00

Added subscriber: @Jeroen-Bakker

Jeroen Bakker commented

2020-01-28 09:36:19 +01:00

Patch available but needs testing.

Michael Weisheim commented

2020-03-06 12:44:09 +01:00

Added subscriber: @MichaelWeisheim

Howard Trickey commented

2020-09-01 02:12:28 +02:00

Added subscriber: @howardt

Howard Trickey commented

2020-09-01 02:12:28 +02:00

Changed status from 'Confirmed' to: 'Resolved'

Howard Trickey closed this issue

2020-09-01 02:12:28 +02:00

Howard Trickey self-assigned this 2020-09-01 02:12:28 +02:00

Howard Trickey commented

2020-09-01 02:12:28 +02:00

I believe this is fixed if one uses the Exact mode of new boolean. 9e09b5c418

I believe this is fixed if one uses the Exact mode of new boolean. 9e09b5c418

Adrian Rutkowski commented

2020-09-01 20:11:16 +02:00

this was fixed back in 2.81 for old solver as well (option in geometry data tab needs to be enabled)

It can be closed

this was fixed back in 2.81 for old solver as well (option in geometry data tab needs to be enabled) It can be closed

Anthony Edlin commented

2020-09-02 22:08:16 +02:00

As said previously, it always worked if you checked store data on the boolean target. It's been that way since, well I just tested 2.79 bmesh boolean and it's the same. So it always worked if you remember to check it on the boolean target even if you don't have any need to have it checked on the boolean target.

I just tried on a build from f2d26409e8 with the new Exact method and it also seems to NOT preserve the data if you don't have it checked on BOTH the original mesh AND the boolean target.

It seemed like it wouldn't be a bad idea for it to preserve the data in the applied mesh if EITHER the original mesh OR the boolean target had store data checked, which is what the diff I proposed tried to do.

As said previously, it always worked if you checked store data on the boolean target. It's been that way since, well I just tested 2.79 bmesh boolean and it's the same. So it always worked if you remember to check it on the boolean target even if you don't have any need to have it checked on the boolean target. I just tried on a build from f2d26409e8 with the new Exact method and it also seems to NOT preserve the data if you don't have it checked on BOTH the original mesh AND the boolean target. It seemed like it wouldn't be a bad idea for it to preserve the data in the applied mesh if EITHER the original mesh OR the boolean target had store data checked, which is what the diff I proposed tried to do.

Anthony Edlin commented

2020-09-02 22:30:16 +02:00

Since I tested in 2.79 I also notice that the Carve solver does preserve the data even if the boolean target does NOT have store data checked. So it's not like it was always this way, just the bmesh boolean has always been this way.

Philipp Oeser commented

2020-12-22 13:04:38 +01:00

In #47129#1006153, @howardt wrote:
I believe this is fixed if one uses the Exact mode of new boolean. 9e09b5c418

like @AnthonyEdlin said, this is still not "fixed" in the way many users expect this [we just had #83932 reported], but yes, 'old' way of storing the data works.

> In #47129#1006153, @howardt wrote: > I believe this is fixed if one uses the Exact mode of new boolean. 9e09b5c418 like @AnthonyEdlin said, this is still not "fixed" in the way many users expect this [we just had #83932 reported], but yes, 'old' way of storing the data works.

Anthony Edlin commented

2021-03-05 09:10:15 +01:00

Note that now that the new "Exact" mode doesn't go through bmesh it preserves the data. Updated patch to compile on recent version in case anyone still wants bmesh boolean to also work.boolean_customdata_save.diff

Note that now that the new "Exact" mode doesn't go through bmesh it preserves the data. Updated patch to compile on recent version in case anyone still wants bmesh boolean to also work.[boolean_customdata_save.diff](https://archive.blender.org/developer/F9869058/boolean_customdata_save.diff)

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

new Bmesh boolean is deleting some of edges data, (Creases and Bweigh) #47129