I discovered this when trying to track down occasional spurious values in new Enzo fields I was working on in a different fork. I do not know if this was the cause of the problem, nor do I know if this affected anything elsewhere. Only discovered this when running with AddressSanitizer, which caused a crash the first time this line was hit.
Good catch! This likely isn't the cause of your problem, but it is undefined behavior.
@bwoshea OK to just merge this or should we wait for a test run?
It's probably fine, but I'll run the tests anyway. I'm doing so right now. Sorry for the delay on this!
Predictably, the test suite passes with flying colors. I'm going to approve and merge.