Hello Michael,

I'm trying to boolean/union the left half sphere in the 'body' object


The result is quite strange.


The sphere was create by the sphere tool, then cuted by a line using the boolean merge tool.

Hope it's clear enough

It works fine here, did you bump the sphere?

Same here, no probs moi_v2_beta_Sep-21-2008.

---------

Works fine for me 21 sept 2008

Endeed reopening the file in a fresh moi session don't show the bug anymore.
(while a copy/paste into a new session was still 'buggy')

I'll try to find the complete procedure this evening and will let you know !

What about this procedure ?

Not at home for me! ( beta 21 sept 2008)
Maybe I have better fly tox :D
Maybe post the file with all objects before the crime?
PS I don't seen the 10.01 I come back in 5 minutes !
No even with the 10.01 all is fine for me!

Hi PaQ, thanks for the very clear repro steps!

After a couple of tries (seems sensitive to the position of the box) I could repeat it over here with the Sep-21 beta, and got it saved off to a file where I can reproduce it.

But it does not happen with v1 or with my current working build, I think that probably it is the same bug as that "broken fillet corner" problem that you posted previously, but I'm going to do some testing to verify that.

- Michael

Hi PaQ, I followed the procedure exactly and I could not replicate your problem, do you delete the left over pieces before the Boolean operation or just hide them ?

Just out of curiosity PaQ, what's the reason that you use Boolean merge in that situation, I tend to use Boolean Diff when splitting solids with a line.

---------

Happy that Michael was able to reproduce it, it's indeed a strange one, probably scaling and the box position like Michael said, is important too.

@Danny : the left part is deleted

>> Just out of curiosity PaQ, what's the reason that you use Boolean merge in that situation, I tend to use Boolean Diff when splitting solids with a line.

Because I didn't know that the boolean difference will work too :)

> Because I didn't know that the boolean difference will work too :)

They will both give the same result for this kind of case.

- Michael

Hi PaQ, turns out this was actually a separate bug from that fillet one, but it was also one that I happened to fix already for the next beta.

It was a regression in that geometry library that happened back a couple of betas ago, involving trimming sphere surfaces.

The way the bug happened, it actually did not cause a problem in the boolean calculation itself though, but in the tessellator. So that problem there is actually a messed up display mesh, not actually messed up geometry.

At any rate, it is all fixed up for the next beta.

I'm not surprised that it took until about now to work some of the kinks out of the geometry library update, that actually went a lot smoother than I had worried it might.

Thanks very much for gathering up the bug report!

- Michael

Thanks for the update !