Support for GNGNS Strings

[bmurray]

I also rolled in a change to allow SNR values to be Floats. I have a receiver that always reports SNR with a float value.

[coveralls]

Coverage remained the same at 100.0% when pulling c7d459b on bmurray:master into a32116e on adrianmo:master.

[icholy]

Considering that the mode values are contextual, I don't see the use-case for the IsMode method.

[icholy]

@bmurray I'm ready to merge it if you revert back to the original PR (and the tests are passing).

[icholy]

@adrianmo the change to a float64 will be a breaking change. We'll have to bump the major version :/

[bmurray]

Thinking more about it, the *GSV values should be Int64, as that's what's in the spec. The library shouldn't compromise for broken chips.

I've also removed the IsMode from the GNGNS string. Test's are all passing on my end.

[icholy]

@bmurray the CI is failing because everything exported needs to be commented. I'm not the owner, so I can't merge until it's passing.

[bmurray]

Hm, they're commented now, and it runs against current Go. I don't know what those other errors are related to; it seems like Travis-CI is having issues pulling golint ?

[icholy]

@bmurray yeah, this is related to #41

[bmurray]

golang/lint#421

Looks like 1.11, and 1.10 are the only ones officially supported for lint. Should just drop testing against 1.7 and 1.8; 1.9 still works for now, so that's a bit of a tossup.

[icholy]

Thanks for the contribution!