Ok, I have just tested it a little bit. I dont see any regressions on speed calculation. All is ok from that side.
I was talking about one IG, but I did say nothing about merging, sorry for misunderstanding. So, it is ok for me now. ______________________________ Anton Gladkyy 2010/7/15 Bruno Chareyre <[email protected]> > > Suppose sphere and facet in contact, where the facet rotates along axis >> parallel to the contact normal, so that the contact circumscribes a >> circle around the facet's rotation center -- will that work? >> >> > Yes, of course. Why not? > > So, I revert and resume working with ScGeom as usual? >>> >>> >> Yes, please -- unless you finally say what is the benefit of that >> change. >> >> >> > There must have been a misunderstanding at a point. I thought merging > ScGeom and Dem3Dof was clearly an objective for all of us. Anton asked for > that once again a few days ago. We discussed that in many mails, and for > hours in Prague. That is why I didn't elaborate the reason for merging nor > discussed it in mails before commit, I thought the point was clear. I also > saw that GenericSphereContact was planned for removal at a point : how would > it be removed without merging classes? > > Do you remember this discussion on precomputing different things in a > unique geometry class, based on flag saying if the functors are incremental > or total? You said it would be like emulating inheritance, very true, so I > thought using inheritance was smarter than emulating it. > In my vision, ScGeom would precompute displacement for incremental laws in > contact point paradigm, while Dem3Dof would precompute total strains in > continuum-discrete paradigm (not fully consistent, since we have in fact 4 > possible combinations with point/continuum vs. incremental/total, but only > two combinations are in Yade atm). > > Actually, I will not try to proove a real need for merging. I did it only > because I felt it was high in the list of TODOs and I was the assignee (like > e.g., don't ask me to explain why there is a need to change signs, and why > me the assignee). From a productivity point of view, having two different > families of geometry classes and functors in the code is really not a > problem. I'll be happy with the previous situation. > > I need to revert carefully : the same commit contains independant > improvments and cleanings in ScGeom (bad mistake). I'll see if I can do it > today, it will be soon anyway. > > Bruno > > > _______________________________________________ > Mailing list: > https://launchpad.net/~yade-dev<https://launchpad.net/%7Eyade-dev> > Post to : [email protected] > Unsubscribe : > https://launchpad.net/~yade-dev<https://launchpad.net/%7Eyade-dev> > More help : https://help.launchpad.net/ListHelp >
_______________________________________________ Mailing list: https://launchpad.net/~yade-dev Post to : [email protected] Unsubscribe : https://launchpad.net/~yade-dev More help : https://help.launchpad.net/ListHelp
