strange behaviour in mesh edditing

Sofia

Hello
I've been having an issue with edditing meshes in Spine, where sometime moving certain vertices causes other verticies to exibit very strange behaviours, those include but not limited to:

being stuck in their position and not be able to be edited at all
being able to move but in a very restricted manner that doesn't seem to be consistent with the actual image borders
straight up jumping to a different postion and being stuck in a state described by the second point (this is a more rare one, but I was able to catch it with my most recent asignment at work and capture it)

(all this applies to edditing mesh specifically, not moving meshes in setup mode)

This is an issue that's been very consistent for me over the years, I've learned to kinda deal with it and work around it, but I would still like to know why in the world this keeps happening. It happens with many different projects, on different versions, on different computers

I've searched the forum for something like this, and considering the frequency that I run into this issue with, I expected there would be a lot of discussions asking about this, but there is nothing exactly like it, and the most common advice I see in cases sort of like it is to edit with "deformed" unchecked, but I don't see any reason to do that if in the end all it's gonna do is change the image relative to the vertices and not the vertices relative to the image

Clearly I'm the one doing something wrong, maybe I'm exporting the images from Photoshop in a way that breaks their data somehow, or maybe I'm not aware of some big no-nos of mesh edditing. The point is that with any other issue I ever ran into I was able to figure out what the deal was on my own, through either experimenting or research, but this is the only thing that seems compleately unpredictable and incomprehensible, and I feel like I'm the only one dealing with it

I've attached the project, I took it from work to see if I could replicate this issue at home, and yes, following the steps in the video it happens every time

for-the-forum.zip

11kB

Spinebot

The issues you're experiencing with mesh editing in Spine, such as vertices being stuck or moving erratically, are not uncommon and have been reported by other users. One potential cause could be related to the mesh vertices being positioned incorrectly, such as being on the edge of the image or overlapping with other vertices, which can lead to unexpected behavior. Unchecking the "Deformed" option while editing can help visualize and correct these issues, as it allows you to see the original mesh setup without any deformations applied. Additionally, ensuring that your mesh topology follows the rules (e.g., no crossing edges) is crucial. If the problem persists, it might be beneficial to check for any constraints or weights that could be affecting the mesh behavior.

If these suggestions do not resolve the issue, it might be helpful to review the project setup or consider if there are any specific steps in your workflow that could be contributing to the problem. Sometimes, deleting problematic vertices and recreating them can also help in resolving such issues.

Sofia

Well, it's at least reassuring that it's been reported a lot, I was starting to think I was going insane

Thank you, good bot!

I'm just puzzled because it doesn't seem like I'm doing something too crazy with this particular mesh, and now that I think about it, for me it alsways happens with the simplelest of meshes with 5-6 vertices at most, whenever I do something more complex it NEVER happens

Nate

Not being able to move vertices because it would cross edges when Deformed is unchecked (which is expected) has been reported but no one has shown this issue happening in other cases.

We have reproduced the problem using your project, thank you! We'll work on a fix and hopefully it can go into 4.2.39.

In your case the deformed and undeformed mesh are identical, so the same vertex movement should be valid in both. Likely there's something wrong with the math that maps from deformed to undeformed (which are the actual UVs that are stored). If you move the upper vertex away from the edge, the problem goes away. It may be a numeric precision issue.

Sofia

Nate
Interesting...
I am truly fascinated by the fact that it's not something that's been shown before, bacause, again, I've been dealing with this thing for years

I keep thinking, what could I be doing that is so out of the ordinary that people haven't reported it on mass? My worst crime I can think of is that I grab an edge that has its two vertices slamed tightly against the opposite sides of the image, maybe moving it causes something to go haywire

In any case, thank you for replying, and I would love it if you could give me some insight on what the cause might be so that I could avoid it better in the future, because, unfortunately, my workplace is unlikely to upgrade to a newer version anytime soon, we're still working with 4.1

Edit: oh, I see that you DID provide insight, didn't see it before, thank you a lot

Nate

Sofia I am truly fascinated by the fact that it's not something that's been shown before, bacause, again, I've been dealing with this thing for years

Yeah, me too! 😆 It seems many users suffer in silence. Submit a report! We want to fix things! At least with crashes, users submit crash reports that give some clues. With behavior problems that don't crash, we probably don't even know it is happening, so the only way it'll get fixed is if we come across it ourselves or someone takes the time to report it. We test and have real world projects, but the legions of Spine users turn over every single rock. We appreciate you taking the time to explain the problem!

Sofia My worst crime I can think of is that I grab an edge that has its two vertices slamed tightly against the opposite sides of the image, maybe moving it causes something to go haywire

Doing this is OK, but can cause the edge to not move if the math is unable to squeak it past the tight area, since edges can never cross. Shaking your mouse around a little often ends up hitting just the right movement to get it past.

Sofia In any case, thank you for replying, and I would love it if you could give me some insight on what the cause might be so that I could avoid it better in the future, because, unfortunately, my workplace is unlikely to upgrade to a newer version anytime soon

It was a math issue, we've now fixed it in 4.2.39. It's hard to identify when it happens, but when it does moving a different vertex slightly may help.

Sofia we're still working with 4.1

You're doing what now!? Did you know 4.2 has physics? 😜 It's a great update, highly recommended!

Sofia

Nate You're doing what now!? Did you know 4.2 has physics? 😜 It's a great update, highly recommended!

Oh I have no doubts! But I'm usually not the one to call for a version upgrade at my jobs, I am but a humble ant worker

I actually never had the chance to play around with 4.2, might do that at my own time, for exploration's sake

And thank you again for being so responsive, working with Spine has been a great pleasure over the years, it really feels like you guys are pouring your hearts into making it the best it can be, keep up the good work!