The un-isolate thing has been suggested a couple of times (just two reports, but I'm sure there are more: T58942, T68624). It seems to be the more useful behavior, but is also quite a bit harder to implement. Otherwise it would have been done by now I guess...
I'd rather keep this as a separate feature request and focus on making this patch work similar to visibility-isolation for now.

Locally I committed these steps separately. Will the commits stay intact when I submit the code for review in the end (when using arcanist)?

The render of the .blend file in your test-Renders.zip file look pretty much the same in Blender 2.80.75 and 2.83.2 for me.

I can reproduce the behavior in you screenshots. It is unexpected indeed.

I can agree that the Denoising Depth pass contains unexpected values here.

I set this as known issue, because it fixing it might require deeper changes to how Rigid Bodies are implemented in Blender and it is unlikely, that someone is fixing this in the next couple of month.

I'm also not able to reproduce the issue yet. It looks as if there is simply some missing update for you.

I can reproduce the issue as well. The faces that were disconnected before are suddenly connected again. I assume that this is a limitation of the current code as well, but don't know for sure.

The issue does not seem to be fixed. The behavior is still quite unexpected. However, I agree that this might be a limitation. Not sure if this is necessarily a limitation of the UI system though. The right click does not seem to trigger an undo (my breakpoint in BKE_undosys_step_undo_with_data_ex wasn't hit).

@Rohit (Rohit624), can you still reproduce the issue in the latest builds? In your test file selection seems to work as expected for me.

Please provide concise steps to reproduce the issue. Based on the current description I have to guess what I have to do to reproduce it, and I'll probably guess wrong.

@Arthur (PholaX), is this still an issue in the latest build? Unfortunately, I do not have the hardware to reproduce it myself.
If the issue still exists, can you make a short screen recording that shows what happens in Blender when you switch from Mouse to Tablet and back?

Seems to work fine here as well.

The instanced lights in the file seem to work fine here as well.
I close this for now.

Closing because there has not been a response for over a week.

I'm not able to reproduce the crash within 2 min on Ubuntu 19.10.

@Zoltán András Heiner (Andari) Can you make a screenshot of the new file (so we have something to compare against)?
It looks fine for me on Ubuntu.

I am also not able to reproduce the error in the latest build.

@Pavel Golubev (pavgolubev) Are you still able to reproduce the issue in the latest build? I'm asking again because it seems to be solved for @Ankit (ankitm) on macos.

Maybe @Brecht Van Lommel (brecht) knows if this is a known limitation of the Cycles shadow catcher.

I'm not able to reproduce the issue.

Is there a way to reproduce this without Unity? If not, can you screenshots and a concise step-by-step description on how to reproduce the issue?

Will close this for now. @Abdou (Ninjabdou), if you are still able to reproduce the issue in the latest build, you can reopen this report. Make sure to provide the additional information @Sybren A. Stüvel (sybren) asked for in the task description. Thanks.

Will close this for now. @Palo Piktor (PaulP), if you are still able to reproduce the issue in the latest build, you can reopen this report. Make sure to provide the additional information @Sybren A. Stüvel (sybren) asked for in the task description. Thanks.

• Use enum class
• Don't use auto

I can reproduce it. The issue is caused by these two lines in viewmove_apply:

CLAMP(vod->rv3d->camdx, -1.0f, 1.0f);
CLAMP(vod->rv3d->camdy, -1.0f, 1.0f);

Can you try to simplify the .blend file, please? Just remove everything that is not necessary to reproduce the issue.

I can document on how I profile with this tool next week: https://github.com/KDAB/hotspot.

I think it would be best if you could find a developer who worked on the Extra Objects addon. You can use "git blame" to figure out who worked on the files last. Looks like @Vladimir Spivak (cwolf3d) made some changes to it before.

I can see the issue, but I do not know if an overlapping mesh is supported.

Related work:
https://gist.github.com/JacquesLucke/65485586a86b2e1de6e3f3b92867947a
rBM5111, rBM5112

Closing because there has not been a response for over a week. If this is still an issue, feel free to reopen this report after you added the requested information to the initial report.

Can you provide examples for when this functionality is useful?
I'm wondering, does it make sense for this to be part of the Extra Objects addon?

Great, glad you were able to solve it.

Please try setting the Bending parameter to 1.0.

@Charlie Jolly (charlie) thanks for the info. Not sure if that is necessary here. Allowing powf(negative_number, integer) in the context of the color correction node does not seem to be correct.

@Philipp Oeser (lichtwerk), looks like you are right. D6696 seems to fix this as well.

@Campbell Barton (campbellbarton) I wonder if rB8f8e91987b53a6 should be reverted. To me, the issue described here sounds worse than what the commit was fixing. Both aren't that bad though.

Can you upload the .blend file? Even better, if you could remove everything from the file that is not necessary to see the bug.

The comment you are talking about was introduced in rB7f4479da4. Unfortunately, that was a large commit, so that's probably not very helpful...

I can confirm that the imported image is stretched to scene resolution. This worked the same in Blender 2.79. While I agree, that automatically changing the aspect ratio of images is not nice, this might not be a bug. @Richard Antalik (ISS), is this by design, a bug or a known issue?

Weird, I'm trying to copy you as closely as possible, but still cannot reproduce it.

Can confirm that it does not key the custom property called parent. I assume that this is because of a name collision, will check a bit more.

Seems to work for me. @Germano Cavalcante (mano-wii), am I doing something wrong?

I can confirm that this fixes the issue mentioned in the bug report, but I do not know the implications of this change.

By "immediatly afterwards" you mean while the error popup is still open? Then yes, I can reproduce it, otherwise not.
I do not get the python errors, but maybe that is because Blender crashes earlier for me.

This seems to work the same as in Blender 2.79. I can make it work by activating Stiffness in your file.