A very strange one.
Every time I try to boolean join these 3 items the top sphere dissapears?
Help!

If you export the 3 pieces
then kill it
import it back
Boolean Union works fine :)

That sound like a work around rather than an explanation.
But, thanks, it will probably keep me going on the exercise.
I have had the same difficulty with 3 from scratch tries.

Where I was able to progress with the work around. Thanks.

Michael will give you an explanation ;)
maybe you have found a vicious bug ;)

A new Babel tower?

Hi Brian, Hi Pilou - definitely a bug in the booleans there, I will put it on my list.

Sometimes saving and reloading can make a difference because when a sphere is first created it has a special mark on it that labels it as a surface of revolution.

Then in the booleans the intersection calculation uses some special case stuff for surfaces of revolution, which in this case seems to have the bug in it.

When it is reloaded, that special "surface of revolution" mark is not there anymore, and it goes through the general purpose intersection processor which actually works better in this case.

- Michael

The strange thing is that I did a very similar exercise quite recently to put those hollows into a column without a problem?.
I thought I had done it exactly the same---well--------!

Hi Michael,

>>>definitely a bug in the booleans there, I will put it on my list.

May I ask what that list is for?


>>>Sometimes saving and reloading can make a difference because when a sphere is first created it has a special mark on it that labels it as a surface of revolution.

Saving / reloading to make simple boolean is really quite a strange way of working. But as simple example that will not always work:-


Boolean union made after save/ exit progam/ reload.(test.3dm)
I do see that on such a simple boolean that it does depend on where the (what I call) the split line is located on the sphere and how this interacts (the position of it) with the top of the cylinder.



I did start to create a chess piece, but have dropped it, as I am having a number of problems with further booleans on the head.

Hi Steve,

> Saving / reloading to make simple boolean is really quite a strange way of working.

It is, but this should only make a difference in a few particular instances, it isn't something that you should need to do on a regular basis or anything.

Please let me know if you run across this more frequently.

Of course it is something that I would like to fix up - in this case the bug is not in my own code but in the geometry library that I'm using.

- Michael

Hi Michael,

>>>It is, but this should only make a difference in a few particular instances, it isn't something that you should need to do on a regular basis or anything.

On such a simple boolean, I am more surprised than anything that it had not been reported during beta, had it?

>>>Please let me know if you run across this more frequently.

It would depend if I starting looking. Like on the simple example shown, I simply revolve the sphere. Various rotations will give various results. It is repeatable. But I mainly see such problems with 3d mesh.

>>>in this case the bug is not in my own code but in the geometry library that I'm using.

Does the library get updated with bug fixes?

Steve

Hi Steve,

> On such a simple boolean, I am more surprised than anything
> that it had not been reported during beta, had it?

Nope, not this particular one.

The particular arrangement of overlapping edges is what tends to trigger it - if you draw things in different views, the seam of the sphere will be aligned differently and won't run across the bug. For example if you drew the sphere in the top or 3D view it doesn't happen.

It's somewhat more natural to draw this in the 3D view since it is easier to target the center osnap at the top of the cylinder there, so that's probably why it wasn't reported during the beta.


>>Please let me know if you run across this more frequently.

> It would depend if I starting looking.

I mean let me know if you run across it while trying to build other models, I can repeat this example over here as well.

If this bug starts to get in your way such that it impedes your progress or forces you to do workarounds several times a day, then that would definitely notch it up in priority.

If it forces you to do a workaround once a month, then it is not as high of a priority.


> Does the library get updated with bug fixes?

It does, I should be getting an update pretty soon that I will incorporate into the v2 beta so that it can be tested to see if any updates have caused new problems or not.

- Michael

Hi Michael,

>>>Nope, not this particular one.

Not from beta testing, mmm, did you not see that problem yourself? Not that it is a problem for myself, as I construct differently, and in such a needed case would look for and check split lines, or as possibly correctly stated by yourself "seams".

>>>It's somewhat more natural to draw this in the 3D view since it is easier to target the center osnap at the top of the cylinder there, so that's probably why it wasn't reported during the beta.

I do find such as unpredictable, but, I cannot argue, or know how other users think or use MoI, but would expect such simple problem to be found by yourself. Do you not use MoI?

>>>I mean let me know if you run across it while trying to build other models, I can repeat this example over here as well.

Sorry, you say you can repeat this, but you need more info? I could cause this to cause me loads of problems, but I currently work around, as I do with any problems I find, but that does not infer I am happy to continue in such a direction.

Michael,
I do not expect your software to be without some problems/bugs, but what is reported is very simplistic and should not appear,... but, even the books I have on nurbs are not error free.

- Steve

Hi Steve,

> did you not see that problem yourself?

Nope! When I previously wrote "Nope, not this particular one." I meant I had not seen this particular bug before, regardless of seeing it myself or reported by others.


> Do you not use MoI?

Yup, I use it a lot, especially to help people, you can just look through the forum to find like 50 models that I have worked on to help people with.

But unfortunately due to the nature of geometry, I have not been able to manually test every single combination of every single geometric shape combined together, there is just too much variation.

I'm not sure exactly what you expect for me to have done, I mean do you expect for me to have manually drawn every single possible rotational alignment between a cylinder and a sphere in my own testing?


After seeing this, I did a quick test myself and was not able to run into any problem with like 15 quickly drawn cylinder and sphere combinations that I drew.

The difference is that I tend to draw these in the 3D view, which as I mentioned arranges the seams in such a way that the problem does not happen. You have to pick between 2 different views to make it happen, like pick the center point in the top view and the radius point in the front view.

If you go to the 3D view and stay in that one view and pick both the center and radius point of the sphere there it won't happen.

I suspect that is why it hasn't been run across before, because that type of workflow is probably more common - I know it is for me anyway.

If it was a really common problem I would have expected to hear about it more frequently.


> Sorry, you say you can repeat this, but you need more info?

Yup, I would definitely like to know if you run across it more often.

In fact, I'm a bit confused - have you run into the problem at all on your own, or just when you experimented with the specific example?

Definitely the frequency and impact of a bug is an important factor for me, especially in a situation where I need to decide whether to mess with a stable release or not and possibly destabilize it.

During a beta release period it is a lot easier for me to slip in bug fixes more rapidly.... I think I've already explained this before.


If you are expecting to get an immediate service release for a released product on every single discovered bug, I'm sorry that is just not feasible.

If it causes data loss or a crash, or has a major impact on workflow or regular use, then I would place a much higher priority on distributing a fix for it.

So yes, the information about frequency is definitely an important factor for me to know about.

If you find yourself wasting time by being forced to work around this particular problem in the future, please let me know and it will help me to assign a higher priority to distributing a fix for it.


If you are unhappy with the quality of the product or with my release policy, then I can certainly refund your money if you would prefer to switch to a different product.

- Michael

The curious thing is that now I can't remake the bug! All works fine for me! ???
Is the first file change?

Hi Pilou - remember that if you load the saved file the bug doesn't happen anymore, it is only when you draw the objects brand new in the particular viewports.

- Michael

I have been trying to repeat my problem, starting with empty scenes.

No way I can get the problem to happen again!

A Correction!
If I try to build the exercise in the Right pane I can repeat the problem, but not if I work in the Front pane?!!

Hi Michael,

>>>But unfortunately due to the nature of geometry, I have not been able to manually test every single combination of every single geometric shape combined together, there is just too much variation.

I thought the beta testers would of done most combinations.

>>>I'm not sure exactly what you expect for me to have done, I mean do you expect for me to have manually drawn every single possible rotational alignment between a cylinder and a sphere in my own testing?

Possibly 4, one in each view.

>>>The difference is that I tend to draw these in the 3D view

It is not always easy to use the 3d view for some construction.

>>>In fact, I'm a bit confused - have you run into the problem at all on your own, or just when you experimented with the specific example?

Yup, on the chess piece. I was having a number of problems attempting to boolean out sweep solids.

>>>During a beta release period it is a lot easier for me to slip in bug fixes more rapidly.... I think I've already explained this before.

But you say this is a problem with the library and not in your own code, so how could you fix this anyway?

>>>If you are unhappy with the quality of the product or with my release policy, then I can certainly refund your money if you would prefer to switch to a different product.

So I have to be completely happy and not question, or take a refund and go away?

-Steve

> I thought the beta testers would of done most combinations.

Nope, that turned out not to be the case.


> Possibly 4, one in each view.

Nope, the way I test is more like drawing simple models through many steps, not just repeating one step over and over again from different views.

The method that I use tends to be more helpful for the "big picture" because it lets me test and experience the overall workflow of MoI.

It is funny that you would think I have never used or tested MoI, if you only knew the huge amount of time I have spent doing just that...

If I had an unlimited amount of time, I would be able to test each step of every model in every viewport like you want me to. But I'm limited to only 24 hours in a day.


> Yup, on the chess piece. I was having a number of problems
> attempting to boolean out sweep solids.

Why do you think these problems you found were the same as this bug?

You say you had a sweep instead of a sphere and a cylinder? And did your problem disappear when you saved and reloaded the model?

If you can post your model it would help me to determine if you hit this bug or some other one.


> But you say this is a problem with the library and not in your own
> code, so how could you fix this anyway?

I have the source code for the library and I can make fixes myself in certain circumstances (and have done so). But the more complicated a piece is (like fillets and booleans), the more difficult it is for me to do this instead of the library's authors.

Sometimes it is also possible for me to solve a problem by incorporating a workaround more directly into the code that calls the library instead of making you do it as an individual step yourself.


> So I have to be completely happy and not question, or take a refund and go away?

I welcome questions, and especially feedback and bug reports that help to improve the product. What I don't welcome is repeating the same answer about my release schedule 50 times.

It just seems like you are quite unhappy with the product (or is it all software in general?) - if that is the case I can give you a refund instead of prolonging your bad experience with it.

- Michael

>>>What I don't welcome is repeating the same answer about my release schedule 50 times.

You have repeated this 50 times where?

>>>It just seems like you are quite unhappy with the product (or is it all software in general?)

No, just your attitude towards bug reports, and your indication that none will be fixed.

Of all the authors of software that I know up to date
Michael is the most helpfull and fast responding I have
seen and worked together with !
Sorry I can not share your experience!
Moi 3d is one of the most powerfull applications out there
and a real joy to use !
There is nothing like bug free software. ; )
I am looking foreward to having the privileg to
be able to use and test forthcoming betas when ever they come....
I wished that other software were so stable and predictable
as Moi 3d !
I´ve actually seen quite a few friends that i gave the heads up on the
software that are very happy with it !
A happy user
Many thanks
Peter