uci doesn't like dashes in section names

[perbu]

Hi.

I noticed that the parser is a bit more relaxed than the uci parser in OpenWRT. Would you be interested in a patch that would fail is we see something other than a-z, A-Z, 0-9 or "_"?

@relet made me aware of this.

[perbu]

(this would naturally be only relevant for the v2 branch)

[perbu]

There is an issue on this; #8, which we might address at the same time.

[dmke]

As hinted in #8, go-uci ought to try its best to mimic the behaviour of uci itself (i.e. if uci parses "wifi-device", so should go-uci). Keep that in mind when rectifying the parser :)