FreeCAD uses ifcopenshell 0.5.0 dev for import of ifc. It seams I have an ifc which has en entry for which ifcopenshell does not respond a usable shape. The ids are 78, 182 and 4504. If you are an Architect you will for sure like the way ifcopenshell represends the stair. Because I can not attache files I post the link to FreeCAD Forum entry.
Kind regards bernd
Yeah no wonder, the github repository was lagging a good 25 commits behind. I updated the repo so that ought to take care of the missing UNP profiles.
If you guys have any suggestions to keep the repo's in sync a little bit better, I'm all up for that. Complicating situation is that not all commits from the svn repo apply cleanly onto the git repo, because of some of the changes for the more elaborate python bindings.
thank you very much Thomas for attracing it that fast :-)
May I send just one issue before disscussing the work flow of updating and bug tracking? Have you been thinking of supporting IfcReinforcingBar? See http://forum.freecadweb.org/viewtopic.php?f=23&t=5526&p=50853#p50855 for an example file.
Regarding update and bug tracking ...
Update: Since I do not know any C++ I do not know how I can help with this.
bugtracking: How about using the issue tracker on github?
kind regards bernd
You're right, the representation is defined as a IfcSweptDiskSolid, for which IfcOpenShell currently does not provide support. Since it is basically a special case of IfcSurfaceCurveSweptAreaSolid, which is supported, implementing this should be fairly trivial. I'll see if I can find some time for that somewhere this week.
I'm fine with both github or the sourceforge bug tracker: https://sourceforge.net/p/ifcopenshell/bugs/new/ you can add files there as well.
Yorik van Havre
Couldn't resist to join the discussion ;)
Yeah, I would also vote for using the issues tracker... (You need to enable it on the github repo if you want it). IMHO this would allow us to track issues related to the github version separately, which would be cleaner, and easier to manage for you Thomas. And I think it would also make it easier for other people to contribute with solution and fixes (I think there is a way to submit a merge request as a solution for a certain bug)
I see no easy solution for the svn/git problem... Probably the best path would be to immediately commit the same change to both repositories, when possible. One way to do that might be to import the svn repo as a branch in the git repo (with git-svn) but since many files are different it might not be of much use either. Other is to work with patches, bu I suppose that it is what you do at the moment...
Thanks for weighing in on this. Ok, I enabled the github issue tracker. As for the merges, I pretty much gave up on something automated. What I have now is a git-svn copy of the svn repo and use 'git format-patch' and 'git am' respectively to swap things around. It's good enough although it completely trips over windows vs linux line ending differences :(
Would be really cool if you could adopt this. Since I'm engineer the reinforcement stuff is as important for me as windows are for you guys. ;-)
Regarding bug tracking. I'll use the new enabled issue tracker for further problems. There might be some issues in the future since we now are able to do heavy use of ifcopenshell 0.5.0dev inside FreeCAD.
Again thanks Thomas for responding so fast regarding all the issues we had.
Kind regards Bernd
I know, it's not that I discriminate against engineers ;) It's just that up until now I've seen most of these elements being exported as faceted geometry. Looks like Allplan is a little bit more advanced than some of the other exporters.
Your file should import correctly now. I didn't have a good model with IfcSweptDiskSolids yet. Is it ok if I add your model to IfcOpenShell's set of public test models? Thanks.
YEAH, ifcopenshell does it. http://forum.freecadweb.org/viewtopic.php?f=23&t=5526&p=51002#p51002
Many Thanks Thomas. Sure you can add the modell to the public test models. I'd be proud to help ifcopenshell. It is an simple model. If you'd like a more sophisticated one I could help too ?
Kind regards bernd
Thanks. Actually the simpler (more isolated) the better, easier to spot regressions and modelling failures :).
The file I IfcReinforcingBar.ifc you took to your examples has an problem in the File Header which is not liked by solibri viewer.
line 4 should be
It was my misstake when I erased some internal informations manually.
Kind Regards Bernd
I see, thanks for that. IfcOpenShell doesn't look at the header. Actually the author and organization should be between parenthesis, as it is a list.
OK lets test the new work flow. https://github.com/aothms/IfcOpenShell/issues/7