Fillet crashes
-
- Posts: 13
- Joined: Wed Oct 16, 2024 3:00 pm
Fillet crashes
We have a small demonstration VB6 project that illustrates the behavior we have found with the latest 7.2 library when trying to apply fillets between two united intersecting bodies (a thin box intersecting a cone). The curious behavior is that the fillets work very nicely on two of the four edges, and crash the program (in IDE and EXE) on the other two. I have attached a ZIP for the small VB6 demo project. You can select between successful and unsuccessful behavior in the Fillet sub (in modFillet) by setting blnTest to True or False. Any thoughts are appreciated. Thanks!
- Attachments
-
- modFillet.zip
- (2.67 KiB) Downloaded 769 times
-
- Posts: 13
- Joined: Wed Oct 16, 2024 3:00 pm
Re: Fillet crashes
I see that the demo project has been downloaded several times. Can anyone confirm that you are seeing the same error that I described? Thanks.
-
- Posts: 86
- Joined: Tue Jun 18, 2024 6:12 am
Re: Fillet crashes
Hello.
We have entered the problem into the system. It is in a queue. It might take some time.
Thank you for reporting.
Regards
We have entered the problem into the system. It is in a queue. It might take some time.
Thank you for reporting.
Regards
-
- Posts: 13
- Joined: Wed Oct 16, 2024 3:00 pm
Re: Fillet crashes
Thanks. I appreciate the attention. Please let me know if you need clarification or additional details.
-
- Posts: 86
- Joined: Tue Jun 18, 2024 6:12 am
Re: Fillet crashes
We have started the work. I expect this one to be tricky as the crash happens inside of OCCT code.
-
- Posts: 86
- Joined: Tue Jun 18, 2024 6:12 am
Re: Fillet crashes
The test seems to be popular. Here is a C++ version of it.
Change bDump to true in debugger or code to see the edges in 3D.-
- Posts: 86
- Joined: Tue Jun 18, 2024 6:12 am
Re: Fillet crashes
This is an OCCT bug. This case of geometric configuration is not handled inside the algorithm, and it is a pretty old bug, The case may have never been handled.
We have submitted a clear proof in the OCCT bug reporting system. See the issue 0033877 at tracker.dev.opencascade.org.
We have to wait for their response now.
Thank you for your patience
We have submitted a clear proof in the OCCT bug reporting system. See the issue 0033877 at tracker.dev.opencascade.org.
We have to wait for their response now.
Thank you for your patience