ETA with leading zeros

[oliverpool]

This is related to #65: the time string is making some "jumps" when it passes from 1m 10s to 1m 9s.

This PR adds a leading 0 to all "trailing" units:
1m 9s becomes 1m 09s, but 9s stays as it is

[arturi]

Thanks! Do you think we should do the same with speed?

[oliverpool]

Thanks! Do you think we should do the same with speed?

Regarding the upload speed "string jumps", the package prettier-bytes could be used instead of pretty-bytes.

(prettier-bytes) augment pretty-bytes to make the output a little more readable (and a little less precise)

[arturi]

Regarding the upload speed "string jumps", the package prettier-bytes could be used instead of pretty-bytes.

Probably yes, sorry I missed your comment on that. I tried using it and had to revert because of IE10 support — the other package used something that was incompatible. Will look into this again.

[oliverpool]

I tried using it and had to revert because of IE10 support

The 2 codes don't seem to different from one another: prettier-bytes, pretty-bytes

[arturi]

It’s the Number.isNaN here, it’s ES6: http://caniuse.com/#search=isnan

[oliverpool]

Number.isFinite from pretty-bytes is also unsupported by IE11: http://caniuse.com/#search=isFinite