Implement uniqueness (hash equality)

[jage]

Will reduce duplicates when using .uniq, to get even better result the user could first normalize, then uniq.

Close #123

[walro]

Couldn't help myself from checking how Ruby does it for the URI class. If I get it correctly they stuff all "components" into an array and piggyback on that array's hash. Feels kind of neat as we'd avoid the usage of digest.

[dentarg]

Link to the components in URI: https://github.com/ruby/ruby/blob/e6d1c72bec5c6544e9ae82501a6cdd7460222096/lib/uri/generic.rb#L44-L53

[jage]

Couldn't help myself from checking how Ruby does it for the URI class. If I get it correctly they stuff all "components" into an array and piggyback on that array's hash. Feels kind of neat as we'd avoid the usage of digest.

👍🏻

I’m thinking we don’t want “implicit cleaning” by using the components, thus. maybe .to_s.hash could be better in this case? Would remove the digest as well.

[walro]

I’m thinking we don’t want “implicit cleaning” by using the components, thus. maybe ‘.to_s.hash’ could be better in this case? Would remove the digest as well.

SGTM!

[walro]

LGTM

[jage]

So, should we handle this as a strict bugfix, i.e. patch release? I'm thinking this should be released as a patch, the reasoning is that if you write code with [url1, url2].uniq you expect the unique results which didn't work before. If you don't run #uniq nothing will change.

[walro]

Sounds reasonable, not sure what other unreleased stuff we have though, so it might be a minor release anyway (or even major) :)

[jage]

Sounds reasonable, not sure what other unreleased stuff we have though, so it might be a minor release anyway (or even major) :)

Hehe, yup. :)