Design of Assembly crashes repeatedly

[TrippyLighting]

I am helping a user in this thread to identify odd behavior in the assembly (jumping joints for lack of better words.)

Currently when I try to update a linked assembly in the overall assembly Fusion crashes repeatedly and I've come to a hard stop.

I cannot share the design publicly but with an AD employee.

I've submitted several CERs in respect to this,. This is the last one. CER_142856985

[ahreum.ryu]

Hi @TrippyLighting, 

I really appreciate your support! 

I'm going to report it ~ thank you for your help again. 

@frc68cad

To report this issue to development team, could you share your file to me? sorry if we request many times. 

Opps I forgot to leave my email address : ahreum.ryu@autodesk.com  I look forward to your reply. 

Many thanks, 

Ahreum.

[TrippyLighting]

Hi @ahreum.ryu please have your developers share their email with me =, so I can invite them to the project. @frc68cad and his team of student might not have experiemced this problem yet, but I have. I have only updated one file and know exactly what I changed.

Also, I've seen several threads lately with the same topic where users reported crashes when attempting to update assemblies, so this I believe is not just an isolated incidence. I just pinged you from another on of these threads.

[frc68cad]

Ahrem, I have sent you an email.

Thanks for your help!

Paul

[frc68cad]

TrippyLighting,

Yes, we've been plagued by crashing during updates.  Sometimes updating individual components (sub assemblies) in the browser tree worked better than trying to update the entire design.  Other times, we couldn't get things to update, and had to delete the assembly and re-insert it.

Thanks again,

Paul

[ahreum.ryu]

Hi @frc68cad, @TrippyLighting, 

I found the same issue and it was reported it(FUS-29935) and they are working on it. 

I've posted the comments and share your file to development team. if I receive any response from them, I will get back to you. 

Sorry for the inconvenience.

Many Thanks,

Ahreum. 

[MichaelT_123]

Do not worry Mate,

we share your pain in our joints with you ... but seriously ...

The forward kinematics feature is a significantelement fused into whole F360 environment. It is not easy for sure and there are significant challenges there.

I have had similar "events" with more complex then presented in demos joint scenarios.  An important thing  I learned, that particularly Motion Links have had to be updated not by editing them in place, but bringing them into a battle front of the design. It is from the stand point of a design process, quite natural (even in a real hard world).

Here is a plead for TF360 to systematise this or similar approach ... of course after a thoughtful deliberation.

In Pain We Trust

MichaelT

[TrippyLighting]

@MichaelT_123

I appreciate the sympathy but I fail to see how what you posted is relevant to this thread.

The assembly we are talking about in this case has no motion links or forward kinematics in it, nor does it need any.

[MichaelT_123]

Hi,

"Assemble" ... the pull down menu in F360 is the section dealing with "forward kinametics" . Sorry for the confusion ... I did not want use your time ...

Cheers

MichaelT

P.S.

No replay needed.

[frc68cad]

Hello,

More info that might help identify what is causing the crashing...

A few days ago, I decided to start going backwards in version to find one that does not crash.  Fortunately it was only one version prior.  So if you can compare v50 with v51... 50 works, 51 always crashes.

The current version 55, now has all updated subassemblies.

Thanks for your help,

Paul

[ahreum.ryu]

Hi @frc68cad, 

I'm really sorry you are experiencing this issue. I searched your CER report again and I'm going to ask them again. I will be back as soon as possible.  

Many thanks, 

[ahreum.ryu]

Hi @frc68cad, 

Sorry to have kept you waiting. I received the reply from the development team.(FUS-30687)

They said that the date set is actually "corrupt". 

It contains a bogus OccurrencePath. The first entry in the path points at a valid ComponentInstance. But the second entry points at a different meta type. Hence, we crash when trying to dereference the second entry as a ComponentInstance.

This is a classic corruption that happens when a file is uploaded/imported as a new version in an existing lineage. For example, Fusion lets you upload a nut.f3d as a new version into the lineage of a screw.f3d. Since the files are so different, the entity ids will not match.

Anyone referencing this file, and looking for entity id X will expect a particular meta type there, but could find a totally different type. We have fixed these kinds of issues now, but older files that are already corrupt cannot be repaired. The best thing is for the user to go back to a previous version of the file.

I'm sorry to say that we couldn't repair your file. if you need any further help, please let me know. 

Many thanks, 

[frc68cad]

Thank you for the update.  It is good to know these issues are fixed going forward.

Even though we had some difficulties, I still consider overall our move to Fusion 360 to be a success.