Brandon Bocklund

@isuruf I'm looking into this a little more. It seems like the traditional solution is to use a weak reference to break the cycle. If my understanding is correct, it...

From the comments of the StackOverflow answer, it seems weak references would be viable. Can you help correct my misconception about which "still useful" objects could be garbage collected? My...

If I'm understanding right, the C++ function(s) need to be implemented here: https://github.com/symengine/symengine/blob/master/symengine/cwrapper.h Then those can be ccalled from Julia through the `deps.jl` that is generated. Will different relationals also...

Can we discuss what the behavior of Fireworks should be? My suggestion is that 1. Without parents, we keep the current implementation 2. With parents A. If Input set has...

@computron I think with the new changes in atomate v2, VIS objects will not be necessary for us as long as full string import paths are supported.

@computron Totally agree. @mkhorton I think testing the API on a set of example files or an example database built as part of the test suite is reasonable for builders....

Yes, it is not supported anymore

Isn't what @shyuep proposed is basically how it works now? Except fake vasp is a powerup (currently) vs. a environment variable (here). I see some validity in the idea of...

I agree with @shyuep. Further, I think that atomate calculations should be entirely driven by PMG input sets (or any object that implements the API). To me, it seems rare...

@shyuep Yes, it should also be easy to start from the MPRelaxSet, make a minor modification and give that to atomate without creating an entirely new class. In my experience,...