physics Physics problem EXIT:: SOLVED!

[MIVIMEX]

Hello! The problem solved!

 

Hello! Help please I have a problem. I'm trying to upload through the Firestorm the roof with an exit to the top, but I can not do physics. I tried a simplified cubic physics and the same model. But the passage is not formed. Applied rotation and scale. Prim / Convex hull don't work. I press the analyze and without, it does not help. It will be the roof of the moving object, so I think the analyze needs to be done.
I will be grateful for any help! Thank you!

P.S.  When uploading with cubic physics, I do not even see the outlines of these cubes. Two months ago I uploaded another model and every cube could be seen.

Edited September 18, 2018 by MIVIMEX

[Rolig Loon]

Make your roof thicker?

[MIVIMEX]

3 hours ago, Rolig Loon said:

Make your roof thicker?

I do not know why the physics of the passage was not visible on this horizontal surface. it is still displayed as a solid object, although it could be passed through it. but when I uploaded the walls the doorways were visible. and you can go through them. and their physics is displayed correctly.

by the way nice new profile picture!

Edited September 18, 2018 by MIVIMEX

[Kyrah Abattoir]

It would be nice if you posted what the solution turned out to be.

[ChinRey]

17 hours ago, MIVIMEX said:

I do not know why the physics of the passage was not visible on this horizontal surface. it is still displayed as a solid object, although it could be passed through it.

That's a known bug. The physics display the viewer generates is not relaible for very thin objects set to prim physics.

[ChinRey]

2 hours ago, Kyrah Abattoir said:

It would be nice if you posted what the solution turned out to be.

Obvious answer: Do not use the opensim enabled version of Firestorm to upload with analyzed physics because ti doesn't actually have the analyze function.

[Beq Janus]

While the problem itself is almost certainly the thin walls, the physics view may have added confusion here (sowwy). That said the OP is not clear on this as they state it did not work with either analysed or non-analysed. It may be related to using the OS version as @ChinRey suggests, perhaps the OP can explain. I will explain why physics view in this case might have added to the confusion if not checked thoroughly.

If the mesh had not been analysed, then the physics would not allow an avatar to pass through, the physics display, however, should have shown this clearly (and looked something like the third image posted by the OP (even though that was showing "forced convex hull" so no surprise it was filled. Expanding the mesh to >0.5 would have worked fine and displayed correctly (in FS) in physics view. This is because I fixed the bug that caused it not to.

Unfortunately, in in fixing that bug I opened up a separate use case where the analysed meshes do not show correctly under 0.5m. In this situation, however, the mesh itself is passable so most people don't notice the bug because they don't need the physics view to test something that works!!

TL;DR There's a bug in my physics view in FS. I'm busy working out how to fix it at the moment. 

The essence of the problem is that in the code there is a function that tells the viewer what type of physics is being used. It is in that function that the code looks at the user selected shape (convex hull/prim/none) as well as the dimensions of the object, to see if it is going to be forced to be convex due to thinness. There is no data available inside that function to allow it to distinguish whether the object in question has a mesh or hull physics and so, by making it behave right for one case, it went wrong for the other. ?

I am working on providing the additional information into that function now so that it can make the right decision. Hopefully, I'll get it into the next release RL permitting. Not sure whether this bug applied to this circumstance, but it would have confused the debugging potentially.

I raised a Jira against myself to track this https://jira.phoenixviewer.com/browse/FIRE-23053

 

 

[MIVIMEX]

11 hours ago, Kyrah Abattoir said:

It would be nice if you posted what the solution turned out to be.

@Kyrah Abattoir

when I uploaded the first model, I could not pass through the passage, the physics was solid, then I uploaded more and more models, trying various versions of physics, etc. waiting for when in physics the passage will be visible. and now quite desperate, I asked a question on the forum and was already going to go out, and decided to try to go through and I got it! although the passage in physics was never seen. so I do not know what really caused the problem and its solution ...

[MIVIMEX]

9 hours ago, ChinRey said:

Obvious answer: Do not use the opensim enabled version of Firestorm to upload with analyzed physics because ti doesn't actually have the analyze function.

@ChinRey

Oh, I didnt knew that! Can I use official LL Opensim enabled viewer for this?

Edited September 19, 2018 by MIVIMEX

[Beq Janus]

2 minutes ago, MIVIMEX said:

@ChinRey

Oh, I didnt knew that! Can I use basic LL Opensim enabled viewer for this?

You should only use the OpenSim viewer if you go to opensim regularly. It is different to the Havok build because there are libraries and functions that are not licensed to use outside of SL. If you are only ever uploading in SL then using the OpenSim build of Firestorm is never the right choice.

[MIVIMEX]

4 minutes ago, Beq Janus said:

You should only use the OpenSim viewer if you go to opensim regularly. It is different to the Havok build because there are libraries and functions that are not licensed to use outside of SL. If you are only ever uploading in SL then using the OpenSim build of Firestorm is never the right choice.

@Beq Janus

so can I use a firestorm without an os to upload a mesh with analyze?

[Beq Janus]

11 minutes ago, MIVIMEX said:

@Kyrah Abattoir

when I uploaded the first model, I could not pass through the passage, the physics was solid, then I uploaded more and more models, trying various versions of physics, etc. waiting for when in physics the passage will be visible. and now quite desperate, I asked a question on the forum and was already going to go out, and decided to try to go through and I got it! although the passage in physics was never seen. so I do not know what really caused the problem and its solution ...

This sounds exactly like you tripped over my bug then. Basically, you had uploaded with non-analysed and then found it not working (too thin) then, in desperation, uploaded other and did not test but just observed, the analysed one was probably fine all along, but it was not until you tried it that you fond that it worked because the bug had misled you.

On the plus side I appear to have fixed the issue, and after a bit more testing I'll submit the change so hopefully it will appear in the next release.

In the following video, I have applied the fix. The frame on the right (yellow) is analysed and thus will always work no matter how thin, the frame on the left is non-analysed and will only allow you to pass when >0.5 thick. 

https://i.gyazo.com/7eb9ee6314b8c0e7ec711a8a89bbefd3.mp4

 

 

 

[Beq Janus]

9 minutes ago, MIVIMEX said:

@Beq Janus

so can I use a firestorm without an os to upload a mesh with analyze?

Correct, you should use the Firestorm for SecondLoifeonlky download. from http://www.firestormviewer.org/windows/

Any version of the Linden Lab viewer will also work of course.