[Fixes #12] Fixed `+` encoding to `%2B`

[korotovsky]

@trowski Probably we have to exclude [ and ] from this regexp as well

[trowski]

Yes, in fact I think !$&'()[]*+,:;=/% all need to be removed from that regex. Does this sound correct?

[korotovsky]

No, we have to remove [, ], +, =, /, %, & and '.

[nexor]

Сould we just use rawurlencode function instead of this method ?

[trowski]

@nexor The function is using rawurlencode() to replace characters, but some characters should not be encoded, particularly those that may already be an encoded character.

@korotovsky Why only those characters? Based on the URI RFC, the only excluded characters should be alphanumeric plus -, _, ., !, ~, *, ', (, and ) (see section 2.3).

[korotovsky]

@trowski Probably we should follow URI RFC, but my suggestion was based on Google Chrome 47 behavior.

[korotovsky]

As I see this commit is already in master, so I close this PR.

[trowski]

I excluded other characters from the URI RFC as well except [ and ]. Should that be fine or is there a reason to encode them as well?

[korotovsky]

Google Chrome encodes it :) RFC says that we should not... I'm confused.

[trowski]

Actually the RFC says [ and ] should be encoded. I thought Google Chrome was not encoding them. Probably means it should be encoded here as well.

[korotovsky]

Then we have to encode them. I misunderstood you.