Hi Pilou,

> 2 same primitive cylinder Bolean union have only the internal common
> faces dispears but not the common section :)

Yup, that's normal - notice that I did not list cylinder surface combining as one of the special cases that are currently handled ?

It could be possible for that particular case to happen in the future, it would require some modifications to some pretty complex areas of the geometry library though.

In the meantime usually it is possible to solve such things yourself if you want to eliminate internal edges by creating one larger cylinder and trimming it to the proper boundaries.

Also if your 2 cylinders were made up of pieces cut out from one single large cylinder, they will combine in that case since it then is handled by the "same underlying surface" special case. So for example if you draw a tall cylinder, and then cut it in half with a horizontal line to make 2 cylinders that way, if you then boolean union those pieces together it will reform into one single cylinder piece. This can be handy sometimes, because it can sometimes be convenient to carve off a chunk of a model using boolean difference, then do some work on the individual pieces and then when you union it back you will then get surface combining - that's done in the "Crown of clubs" tutorial video for example which is why the final unioned piece does not have extra edges at the bottom of each club where they connect to the bottom part that was carved off from the whole initial revolve.

- Michael

< then cut it in half with a horizontal line to make 2 cylinders that way, if you then boolean union those pieces together it will reform into one single cylinder piece.
It's that who is very pertubating! :)
2 copy cylinders don't union + disapear, 2 cylinders from a cutting one yes! :)

Hi Pilou,

> 2 copy cylinders don't union + disapear, 2 cylinders from a cutting one yes! :)

Yes - it's because there is special case handling used for the situation of "2 surfaces that are each trimmed surfaces using the exact same underlying surface" - that case is more straightforward for the code to detect that the surfaces can be combined together.

For that case and for planes there are mechanisms in place in the booleans to try and reform the unioned result with larger glued together surfaces. For other kinds of cases like the 2 copied cylinders - that is a different case and would require different handling in order for it to work, which is certainly possible but not implemented currently in the geometry kernel.

- Michael

Michael,
Thanks for the help. I think I've got it now.
Mark