Fix rcb balance issue #928
Conversation
correct issue reported in comment at lammps#911
In src/rcb.cpp:460 there is an if (smaller > largest). now if we have one particle you will see that lo[] = hi[] and because of this smaller == largest == 0 for all values of dim. This causes this particular part of the code to never be run. In particular the memcpy inside this if is never executed. This causes an unitialized memory access in line 472. Additionally, dim is initialized with -1 and thus the accesses in 484 and 485 are problematic. Additionally, valuehalf_select is never initialized either. closes lammps#888
akohlmey
force-pushed
the
fix-rcb-balance-issue
branch
from
4bdc2b2
to
d10a470
Compare
|
@sjplimp here is the "lost commit" rebased to the current master and another doc update commit, that may have been lost in the process as well. FYI, the branch is "open", so you can pull it and push changes to it. |
|
changed the corner case with shrink-wrap size = 0.0 to reset it to original box |
|
@sjplimp it is a bit difficult to see what are the real final changes that are going to be applied to master with a pull request that was created such a long time ago. however, i did a manual test merge in my local repository and compared that to master and it looks ok there. so i am handing this back to you for merging. |
|
@akohlmey I made my changes to all 3 files based on the current master files, so it should be fine. |
Purpose
implement changes suggested in issue #888