topic Re: Bizarre sketch bug in Fusion Support Forum

Bizarre sketch bug

urnotryan — Fri, 08 Jan 2021 16:54:48 GMT

Super annoying bug here that is actually making Fusion 360 unusable for me right now.

When working in my sketch, I cannot edit the dimensions of a sketch object once a constraint has been added to a feature - even though the dimension should not be constrained at all.

Please watch the screencast to clarify. You will see me try to change the dimensions on a few ellipses and rectangles. You will see that the dimensions of shapes that have a constraint to some geometry can no be edited (even though they should be). You will see me freely change dims on shapes that are unconstrained, but once I constrain them to some other geomety, the dims are locked, with a bunch of bogus errors. Oddly enough, after adding a geometric constraint, you will notice I can set a dimension once, but then I cannot edit it (even back to the original dimensions!

You can imagine this makes it near impossible to make any edits.

This seems to be a straight up bug. I haven't made any changes to my system that would explain why I can't edit files that were working fine previously. Not sure if there was an update over the last few days that could have caused it.

Re: Bizarre sketch bug

g-andresen — Fri, 08 Jan 2021 16:10:05 GMT

Hi,

screencast missing.

Please try again.

günther

Re: Bizarre sketch bug

jeff_strater — Fri, 08 Jan 2021 16:48:10 GMT

also, add the model. Most likely, even if the screencast were added, we'd need the model to further investigate

Re: Bizarre sketch bug

urnotryan — Fri, 08 Jan 2021 16:55:40 GMT

Added. I stripped it down to just the sketch so it should be able to find exactly what I'm talking about.

Re: Bizarre sketch bug

urnotryan — Fri, 08 Jan 2021 16:56:11 GMT

weird, I had the link pasted in the box. Oh well, added now.

Re: Bizarre sketch bug

urnotryan — Sat, 09 Jan 2021 19:29:53 GMT

This is such a specific problem with such a basic issue... I was hoping this was a shoe in for getting some help here. Still hoping my only option is rebuilding my files from scratch.

Re: Bizarre sketch bug

jeff_strater — Sat, 09 Jan 2021 20:30:50 GMT

you still have not provided the screencast. The model is there, so thanks for that, but without knowing what you are doing, it's hard to offer help.

Re: Bizarre sketch bug

urnotryan — Sun, 10 Jan 2021 05:01:12 GMT

Not sure what to tell you... I've added the screencast twice... it's showing up before I click post.... trying one more time here...

Re: Bizarre sketch bug

urnotryan — Sun, 10 Jan 2021 05:02:56 GMT

Re: Bizarre sketch bug

urnotryan — Sun, 10 Jan 2021 05:03:58 GMT

If it doesn't work this time I give up.

Re: Bizarre sketch bug

jeff_strater — Tue, 12 Jan 2021 00:59:17 GMT

Apologize for the slow response. I took the rest of Sunday off, and then got caught up in Monday tasks. Thanks for the screencast, @urnotryan - it helps a lot. Yes, there seems to be a solver bug here. Something about being coincident to that particular geometry is causing an error, even with a very simple geometry like a rectangle. I was able to reduce the sketch a bit and still reproduce it. We will investigate further.

Nice to see another fan of the political gabfest, as well. In fact, I was probably listening to that episode Sunday instead of investigating this...

Re: Bizarre sketch bug

jeff_strater — Tue, 12 Jan 2021 01:22:02 GMT

more info. I created bug FUS-77020 for this. But, in the process of trying to investigate this, I found that the error is really telling us something useful. When you get these solve errors, you can click on the listed constraint, and it will get selected. It is supposed to auto-zoom into that area, but as you can see below, this is another bug - it zooms WAY out instead. But, if you hunt around a bit, you can find the offending constraint, and if you delete it, things behave much better. I don't understand why this particular parallel constraint is a problem, but deleting it does seem to cure the issue:

Re: Bizarre sketch bug

urnotryan — Tue, 12 Jan 2021 12:30:34 GMT

I suppose it is good to confirm this as a bug and that there was not some simple solution I was missing, but this has been popping up pretty frequently (or some variation of it claiming there are errors due to constraints or geometry that have no reason to cause conflict), and going through the errors to delete constraints (some of which are necessary and I need to replace afterward) is really killer to workflow.

Is there a way to track progress for the resolution of a specific bug? I have about reached my maximum level of frustration with Fusion for now, and I think I'm going to switch over to some projects in different software for a bit, but it would be nice to monitor it so I can have some assurance that my sketches will stop locking up and making changes a huge hassle.

Thanks again for the help.

Re: Bizarre sketch bug

TheCADWhisperer — Tue, 12 Jan 2021 12:55:03 GMT

@urnotryan wrote:

I suppose it is good to confirm this as a bug and that there was not some simple solution I was missing, but this has been popping up pretty frequently

it would be nice to monitor it so I can have some assurance that my sketches will stop locking up and making changes a huge hassle.

I will make a significant wager that you will continue to see issues using this sketching technique.

You have unnecessary duplicated geometry in your sketch.

Pattern features, not sketch elements.

I could very significantly simplify this sketch.

Re: Bizarre sketch bug

urnotryan — Wed, 13 Jan 2021 13:35:38 GMT

I agree with what you are saying in a general sense, but in this application having a base sketch with key relations all in the same place optimizes workflow and makes edits much easier.

Even if you think the sketch can be simplified, I think you would agree that this is hardly an incredibly complex sketch. I do not think it is unreasonable to expect Fusion to handle this without random errors popping up for no reason.

Re: Bizarre sketch bug

TheCADWhisperer — Wed, 13 Jan 2021 13:46:46 GMT

@urnotryan wrote:
I think you would agree that this is hardly an ~~incredibly~~ complex sketch.

I disagree.

In all of the parametric modelers I use - that is not the best technique to use.

Based on my experience there is a much easier way - both for the designer and for the software.

It would take more effort for a complete explanation than I am motivated to explain (I am lazy), but I doubt I am wrong in this case.