0 and 1 token clones not detected

[malinkallen]

I'm running SourcererCC on some really simple test data, among others a couple of empty files and 2 instances a file containing only one (identical) token. I have set
MIN_TOKENS=0
and
MAX_TOKENS=2000000000
in sourcerer-cc.properties.
Clones with two tokens or more are detected, but not the ones with 0 or 1 token. Is this inherent in the algorithm, a feature of the clone detector or may it be a bug? Attached is my blocks.file, obtained following the README instructions. Irrelevant lines are removed. (".txt" needed to be added before GitHub would let me upload the file.)
blocks.file.txt