yade-dev team mailing list archive

[Bruno Chareyre <bruno.chareyre@xxxxxxxxxxx>]

> Gee, we are having serious communication issues. Why did you
> re-introduce the isDynamic() blocks? I am sorry, but I am not (I
> hope;-) ) that stupid to remove them without reason.
I am so stupid that I introduced them without reason then... else why
was it removed without notice? I'm wondering.

I explained why they are usefull already. Trying again :
Say we have a rigid body represented with 10e6 clumped (or non-dynamic)
spheres or facets (typically
https://yade-dem.org/wiki/Screenshots_and_videos#ball_mill). Do we want
to iterate on all clump members to compute acceleration, block
individual dofs, and sum accelarations at the clump level, when we know
it is just wasted time? No of course we don't, it is so many useless
tests and operations.

In my local version, after fixing the FIXME (still waiting discussion on
this one btw), this loop is skipped, simply :
FOREACH(Clump::MemberMap::value_type mm,
static_cast<Clump*>(b->shape.get())->members){
 
>
> 1. computeAngAccell handles DOFs itself, there is no need to test
> isDynamic before calling it. Please do not use isDynamic, it is plane
> confusing; it does not have the granularity of blockedDOFs, and
> without looking at the source, it is not clear what it means. I will
> remove that function soon, to avoid this.
I see. For me isDynamic is exactly equivalent to !State::DOF_ALL, except
that it is shorter to write from body pointer - and it takes less of the
brain space I dedicate to syntax. There is no problem replacing it by
State::DOF_ALL everywhere. I'll do that.
Using finer granularity is still suboptimal, not a big problem since 99%
of the bodies I can imagine have either all or none DOFs blocked.
> Did I overlook something?
>
It seems. Or perhaps it's me. What do you think?
Please Vaclav, don't ask for better communication when all the feedback
I got on this question was oscillating between nothing and "why don't
you commit" (the latest was not a bad feedback though).
Do you agree if I replace isDynamic by !DOF_ALL and commit again (with
fixed FIXME)? 

Cheers.

Bruno