colorizer's error shouldn't suppress output of original command

[mutanabbi]

trafficstars

Any errors, exceptions or assertion faults into a colorizer-wrapper shouldn't suppress the output of an original command. For example see #5. There is a mount-colorizer assertion fault is described there. It caused an exception throwin so an original command output was lost. But actually the command was OK. So I think this is a good idea to catch any exceptions and to print an original command output if some colorizer faults. This remark relates for any kind of wrappers, I beleive.