Integer overflow
dlrobertson opened this issue · comments
A bit contrived, but the following will not yield the expected output when built from master
bitShiftL(x, 32) # yields the value of x
bitShiftL(2147483647, 32) # yields 2147483647
bitShiftL(5, 31) # yields the same
I see. I haven't created the package, just too maintenances at a time the previous maintainer "disappeared".
My current inclination would be to either just document the behavior, or additionally also give a warning, but not change it. What's your use case? How did you find the problem?
No use case, I was more or less playing around. I originally was trying to find out what the difference was between bitops and the bitwise functions in base. I was figuring out if a logical or arithmetic shift was used when I thought I'd test what happened when I did something that should be undefined.
bitwise will always return NA in these cases.
Awesome documentation additions!