Align rotation adds some offset when snapping #42362

New Issue

gandalf3 · 2014-10-23T09:14:19+02:00

gandalf3 commented

2014-10-23 09:14:19 +02:00

System Information
Archlinux

Blender Version
Broken: 2.72

Short description of error
Transform snapping with align rotation enabled consistently adds a small gap between the transformed object and the target object.

Exact steps for others to reproduce the error

Open rotation_snap_test.blend
Press G to grab the spike object (should be active by default)
Move the cursor over the slanted face of the other object and press LMB to confirm (note that the object doesn't snap to the surface under the cursor, it goes a bit up and to the left. This seems to be another bug?)
Note that the bottom of the spike object is a little below the surface of the target object. This distance seems to vary based on the rotations of the two objects.

**System Information** Archlinux **Blender Version** Broken: 2.72 **Short description of error** Transform snapping with align rotation enabled consistently adds a small gap between the transformed object and the target object. **Exact steps for others to reproduce the error** 1. Open [rotation_snap_test.blend](https://archive.blender.org/developer/F119350/rotation_snap_test.blend) 2. Press G to grab the spike object (should be active by default) 3. Move the cursor over the slanted face of the other object and press LMB to confirm (note that the object doesn't snap to the surface under the cursor, it goes a bit up and to the left. This seems to be another bug?) 4. Note that the bottom of the spike object is a little below the surface of the target object. This distance seems to vary based on the rotations of the two objects.

gandalf3 commented

2014-10-23 09:14:19 +02:00

Changed status to: 'Open'

gandalf3 commented

2014-10-23 09:14:19 +02:00

Added subscriber: @gandalf3

Bastien Montagne commented

2014-10-26 13:02:40 +01:00

Added subscriber: @mont29

Bastien Montagne commented

2014-10-26 13:02:40 +01:00

yeah, seems to be something wrong here… :/

Bastien Montagne changed title from ~~Align rotation adds imprecision when snapping~~ to Align rotation adds some offset when snapping

2014-10-26 13:03:28 +01:00

Bastien Montagne self-assigned this 2014-10-26 13:03:28 +01:00

Bastien Montagne commented

2014-10-30 13:11:22 +01:00

Added subscriber: @ideasman42

Bastien Montagne commented

2014-10-30 13:11:22 +01:00

OK, so basically, issue is that the object is rotated (to align along hit face's normal) around its center, while it should be rotated around its hitting vertex (i.e. the vertex that is used to determine the snapping position).

Now, our transform code in general, and snapping one specifically, is quite horrible, but having a look at it I have a feeling this won't be that easy to achieve… So Campbell, question is, based on your knowledge of that area, do you think this is worth spending time to fix in current system, or better make it a TODO for when we redesign that code? Would rather avoid spending more time on it to find that there is no good solution. ;)

OK, so basically, issue is that the object is rotated (to align along hit face's normal) around its center, while it should be rotated around its hitting vertex (i.e. the vertex that is used to determine the snapping position). Now, our transform code in general, and snapping one specifically, is quite horrible, but having a look at it I have a feeling this won't be that easy to achieve… So Campbell, question is, based on your knowledge of that area, do you think this is worth spending time to fix in current system, or better make it a TODO for when we redesign that code? Would rather avoid spending more time on it to find that there is no good solution. ;)

Campbell Barton commented

2014-11-10 15:52:32 +01:00

@mont29, afraid I cant give a good answer off-hand,
there is some real bad stuff going on here, this is one area I didn't touch much (and hasn't been touched since Martin wrote it)

Suggest...

Check ~2.49 see if this is a regression.
Investigate the cause...
... if it can be fixed without lager, long term changes - go ahead! otherwise maybe do as part of larger reworking.

Note, after you snap the object a few times to the same surface, it gets closer and closer to the faces surface (another bug?)

@mont29, afraid I cant give a good answer off-hand, there is some real bad stuff going on here, this is one area I didn't touch much (and hasn't been touched since Martin wrote it) Suggest... - Check ~2.49 see if this is a regression. - Investigate the cause... ... if it can be fixed without lager, long term changes - go ahead! otherwise maybe do as part of larger reworking. Note, after you snap the object a few times to the same surface, it gets closer and closer to the faces surface (another bug?)

marc dion commented

2014-11-10 16:10:22 +01:00

Added subscriber: @MarcClintDion

marc dion commented

2014-11-10 16:10:22 +01:00

The offset seems to be influenced by the object dimensions. I've applied Scale after re-sizing but the results seem the same either way.

The longer the object becomes on the z-axis, the more the widget-ama-bob's offest along the y-axis becomes.

The offset seems to be influenced by the object dimensions. I've applied Scale after re-sizing but the results seem the same either way. The longer the object becomes on the z-axis, the more the widget-ama-bob's offest along the y-axis becomes. ![final.png](https://archive.blender.org/developer/F123100/final.png)

Bastien Montagne commented

2014-11-10 16:26:09 +01:00

Ok, checked 2.49, bug is still present, so no regression here.

The fact that object is OK after second transform-snap on same surface is expected, since on second snapping there is no rotation, so everything works as expected. As said above, cause of the bug is that object is rotated around it’s center, when it should be rotated around it's snapping point.

Reading the code, I could not think of an easy way to do this currently, we'd need to pass upward hitting point, imho this should be kept as TODO then, that whole transform area needs refactor anyway.

Ok, checked 2.49, bug is still present, so no regression here. The fact that object is OK after second transform-snap on same surface is expected, since on second snapping there is no rotation, so everything works as expected. As said above, cause of the bug is that object is rotated around it’s center, when it should be rotated around it's snapping point. Reading the code, I could not think of an easy way to do this currently, we'd need to pass upward hitting point, imho this should be kept as TODO then, that whole transform area needs refactor anyway.

Bastien Montagne commented

2014-11-20 16:15:01 +01:00

Changed status from 'Open' to: 'Archived'

Bastien Montagne closed this issue

2014-11-20 16:15:01 +01:00

Bastien Montagne commented

2014-11-20 16:15:01 +01:00

Thanks for the report, but for now this is not easily fixable, so closing as TODO.

Thanks for the report, but for now this is not easily fixable, so closing [as TODO](http://wiki.blender.org/index.php/Dev:2.5/Source/Development/Todo/Tools#Transform).

Germano Cavalcante commented

2022-04-18 20:09:36 +02:00

Added subscriber: @mano-wii

Germano Cavalcante commented

2022-04-18 20:09:36 +02:00

Changed status from 'Archived' to: 'Resolved'

Germano Cavalcante commented

2022-04-18 20:09:36 +02:00

This seems to have already been resolved.
Maybe in 5ecc6a824a
or ddf99214dc

This seems to have already been resolved. Maybe in 5ecc6a824a or ddf99214dc

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

Align rotation adds some offset when snapping #42362