Merging hamish2014/assembly2 into v0.17, help wanted
Forum rules
Be nice to others! Respect the FreeCAD code of conduct!
Be nice to others! Respect the FreeCAD code of conduct!
Re: Merging hamish2014/assembly2 into v0.17, help wanted
Hi triplus and looo, I was missing those forum links. Before I run this thread further off-topic let's continue the Python3 and CI/CD/Jenkins/AWS discussion right there: https://forum.freecadweb.org/viewtopic. ... 62#p164962.
Thank you.
Thank you.
Re: Merging hamish2014/assembly2 into v0.17, help wanted
(extract taken from https://forum.freecadweb.org/viewtopic. ... 41#p171541)NormandC wrote:[...]IMHO that would be duplicating the work of the already existing official Assembly workbench. It is not in a working state at this time, but once the massive changes in the PartDesign workbench are finalized and 0.17 is released, I believe the plan is to resume work on it (provided there are developers with time to dedicate). [...]CADennis wrote:I wanted to debug a bit into the native code to see if it's worth porting parts of assembly2 from python to C++
Re: Merging hamish2014/assembly2 into v0.17, help wanted
Years have passed and a few more will before there will be Assembly 1. And work on Assembly 2 could just as well be what will make Assembly 1 happen. Therefore whoever wants to work on FreeCAD Assembly capabilities go ahead and don't wait (anymore). In my opinion PartDesign NEXT work is about PartDesign workbench and not about Assembly workbench. Therefore whatever Assembly FreeCAD capability exist at any given point in time i doubt the plan is to make it exclusive to PartDesign workbench. And if that is not the plan whatever work is done on FreeCAD Assembly capabilities that can always be adapted to whatever PartDesign NEXT will evolve to.
Re: Merging hamish2014/assembly2 into v0.17, help wanted
My point was that porting assembly2 code even partly to C++ would mean that it would need to be compiled, and therefore it would no longer be available as an easy to add external module. So merging its C++ code to freecad master would be necessary; at this point, we come into the duplicating work I was talking about.
Not to mention that there would be a lot of work to do in Assembly2 to support the new stuff brought by PartDesign. It may require more work than for resuming Assembly1. ickby already mentioned that big compromises and hacks were necessary to make assembly2 work, and it shows in its limitations. It is not a good solution in the long run.
I think a more realistic immediate proposal would be to port Assembly2 to v0.17, see if it can benefit from the new stuff without spending too much energy, but keep it to python. But if someone is willing to dedicate much more time to assembly coding, I think it would be more beneficial to FreeCAD to do it in the "official" Assembly workbench which aims to do much more than what Assembly2 does at this time.
Not to mention that there would be a lot of work to do in Assembly2 to support the new stuff brought by PartDesign. It may require more work than for resuming Assembly1. ickby already mentioned that big compromises and hacks were necessary to make assembly2 work, and it shows in its limitations. It is not a good solution in the long run.
I think a more realistic immediate proposal would be to port Assembly2 to v0.17, see if it can benefit from the new stuff without spending too much energy, but keep it to python. But if someone is willing to dedicate much more time to assembly coding, I think it would be more beneficial to FreeCAD to do it in the "official" Assembly workbench which aims to do much more than what Assembly2 does at this time.
Re: Merging hamish2014/assembly2 into v0.17, help wanted
I already aligned Assembly2 to FC0.17 Part Design New Body featureNormandC wrote: I think a more realistic immediate proposal would be to port Assembly2 to v0.17, see if it can benefit from the new stuff without spending too much energy, but keep it to python.
https://forum.freecadweb.org/viewtopic. ... 42#p144714
and added some other features that I often use in FC0.17
https://forum.freecadweb.org/viewtopic. ... 30#p147482
https://forum.freecadweb.org/viewtopic. ... 40#p159898
Moreover the latest release has a numpy version that fixed all weirdness that where appearing only under Windows...
https://forum.freecadweb.org/viewtopic. ... 60#p159964
Is there any other issue related to FC0.17 that I missed?
Re: Merging hamish2014/assembly2 into v0.17, help wanted
All i am saying is we need to start thinking on how to attract at least one developer working on Assembly workbench effort directly and to start doing that ASAP. And that developer doesn't necessarily need to care all that much about PartDesign NEXT ATM.
Re: Merging hamish2014/assembly2 into v0.17, help wanted
I can't say, as I don't use Assembly2 at all. Sorry, I wasn't aware of your recent work. But DeepSOIC has some proposed changes (App.Container or something to that effect) that may cause issues.easyw-fc wrote:Is there any other issue related to FC0.17 that I missed?
@ triplus
I'm all for the first part of your reply, but as for the second, I disagree. A good knowledge of PartDesign Next inner workings will be essential.
Re: Merging hamish2014/assembly2 into v0.17, help wanted
I perceive Assembly effort as a stand alone effort. Therefore i can't advice willing developers to wait anymore. Start, discuss, propose, code ... now.NormandC wrote:A good knowledge of PartDesign Next inner workings will be essential.
Re: Merging hamish2014/assembly2 into v0.17, help wanted
Yes, as someone who has waited as long as you to see the mythical Assembly workbench, I agree 100%... but even if a stand-alone effort, without PD Next knowledge (I'm not talking about its development, just awareness and understanding of it), it can only be utterly counter-productive.triplus wrote:Start, discuss, propose, code ... now.
Re: Merging hamish2014/assembly2 into v0.17, help wanted
Assembly 2 shouldn't be affected at all by App.Container because A2 is acting on just the result of the geometry... but anyway any new feature will have to be aligned/tested to all WBs in general...NormandC wrote: But DeepSOIC has some proposed changes (App.Container or something to that effect) that may cause issues.
Then would be nice to spend some time on it to get a deeper knowledge of its way on.NormandC wrote: I can't say, as I don't use Assembly2 at all.