possible testOverlap bug

[quinton-ashley]

Congrats on releasing v1! 🎉🙌

But one of my users reported an error with planck beta 16 that seems to be an issue with v1.0.1 as well. Can you guys look into it for me? 🙏🏻

Here are the messages they sent me:

---

hey @quinton_ashley could you please tell me what this error might mean. I keep getting this error inconsistently from time to time when I play my game. Like everything in the game seems to be working fine but then this pops up randomly and I can't play the game anymore... It wasn't happening until about a few hours ago. I have been trying to figure out what I've been doing wrong for hours but I can't

I would greatly appreciate it if you could help me

---

You're not doing anything wrong, it seems to be a bug in planck. I'll need to report it to planck's developers. What do you think triggers the error? Can you DM and send the code of the project you're working on?

---

Hi, here's the link to my project:
https://github.com/dhyan-rai/cs30-major-project/tree/main/rough-sketch

all of the code is in sketch.js and I have used a few other libraries as well. I'm sorry if it's a bit messy since I have recently begun coding and I don't quite know all the conventions yet... This is for my school project and I have been working on it for a few weeks and only a few days back did I see the error popping up. As to what might be causing it, I don't quite know. Nothing I have done in my code seems to be standing out as something that could cause an issue and besides, I don't understand the error all that well.

The error pops up very randomly... sometimes when I run the sketch and sometimes in the middle of running it

[zOadT]

Thank you for creating the issue! This seems to be related to #269 (both throw because .getVertex return null in https://github.com/piqnt/planck.js/blob/master/src/collision/Distance.ts#L175 or https://github.com/piqnt/planck.js/blob/master/src/collision/Distance.ts#L178 probably)

Unfortunately I wasn't able to reproduce either (and I won't be able to look into them the next 3 weeks) Could you maybe check if the issue was already present in beta-15? It could be that I introduced some numerical issue with #267

[quinton-ashley]

hmm okay I'll ask them if they can try it with planck beta-15