[solved] (Dangerous) behavior from the sketcher

[chrisb]

If you alway make your sketches fully constrained, there is nothing dangerous. If you don't, there is always some increased potential for dangerous things.

[openBrain]

That's right, but I felt like I was far out of the window and told my customers that it was reliable. it is enough for me to switch off the circle with e points. i change the title

I don't know what's the context of your business and who are your customers, but feeding them with development version is IMO a harmful choice.

[thomas-neemann]

That's right, but I felt like I was far out of the window and told my customers that it was reliable. it is enough for me to switch off the circle with e points. i change the title

I don't know what's the context of your business and who are your customers, but feeding them with development version is IMO a harmful choice.

my customers have version 0.19 stable, but that is not yet "1.0"

at 0.19 there is the same problem

[openBrain]

OK I looked at the video (several times) and now I think I understand. Probably there was some wrong terms used in the OP.
Actually the problem is : "3-point circle geometry does not correctly handle automatic constraining at creation time".
In the video, no automatic constraint is proposed with axis (while they are highlighted) - should be point-on-line -, and point-on-line is proposed with root point - should be coincident -.
Did I get it right?

[thomas-neemann]

If you alway make your sketches fully constrained, there is nothing dangerous. If you don't, there is always some increased potential for dangerous things.

I did that afterwards, but that didn't remove the error. you don't even notice it because you don't know the size of the step reference.

[thomas-neemann]

OK I looked at the video (several times) and now I think I understand. Probably there was some wrong terms used in the OP.
Actually the problem is : "3-point circle geometry does not correctly handle automatic constraining at creation time".
In the video, no automatic constraint is proposed with axis (while they are highlighted) - should be point-on-line -, and point-on-line is proposed with root point - should be coincident -.
Did I get it right?

i work with a translator. I guess we don't have any more misunderstandings now.

[openBrain]

i work with a translator. I guess we don't have any more misunderstandings now.

Ok. IMO it's always good to add a notice on bottom of a post that is automatically translated.
Now we are clear, there is indeed a bug (at least on your video, didn't try to reproduce yet).
On the technical side, this is just an inconvenience. But on the user side this is really a significant malfunction as one would probably not pay attention to that and then get weird results.
As soon as one will confirm, please somebody create a ticket.

[chrisb]

I had it already mentioned in the German forum: It works on macOS correctly.

OS: macOS 10.16
Word size of FreeCAD: 64-bit
Version: 0.20.25311 (Git)
Build type: Release
Branch: master
Hash: c5524bb58a7126fa2242d6137eb82ee9d54e82d4
Python version: 3.9.6
Qt version: 5.12.9
Coin version: 4.0.0
OCC version: 7.5.2
Locale: C/Default (C)

[thomas-neemann]

I had it already mentioned in the German forum: It works on macOS correctly.

do you (or someone else) know whether freecad works under mac os in the virtualbox

[Syres]

I'm really struggling to see how the auto constraints can work in one OS and not in the other two, I've found one set of inconsistent coding between the 3 point arc and 3 point circle and my Linux compiled build now allows coincident constraints to be picked from vertices as expected in the 3 point circle but more burrowing is required to fix everything.