New bug

[edowning3]

Hello,

I've discovered a potential bug with parallel.

Details:
  * The version number: 20161222

  * The bugid: Less than 1 byte free

  * The command line being run: $ time parallel --memfree 50G --retries 10 -a 
commands.txt 2> parallel_stderr.txt > parallel_stdout.txt

    * The specific commands in "commands.txt" are calls to a Python 3 script 
which uses the SHAP library to explain Keras deep learning models. 
https://github.com/slundberg/shap

  * The files being read (put the files on a webserver if they are big)

    * Unfortunately I cannot release the data files or code as they're for a 
private project I'm working on. If need-be, I can try to replicate this error 
using SHAP on an open dataset using a toy Python 3 script I can release.

Please let me know what you think.

Cheers,
Evan