just-path returns different results for clj and cljs

[Serhii Klymenko]

When url contains encoded symbols just-path works in different ways on java and javascript.
qwe/asd%20zxc -> qwe/asd%20zxc on java,
-> qwe/asd zxc on js.

It seem goog.Uri.getPath automatically decodes uri. But java.net.URI/getRawPath keeps it encoded.

[Dominic Monroe]

I believe there was a reason to get it with encoding in. The result from goog.Uri should likely be URI encoded afterwards to match, but I'm not certain.

Does this casue #155 to behave differently for clj & cljs?

[Serhii Klymenko]

yes, when I use regex [#".+" :par] :par is decoded in cljs, but it's encoded in clj. It is not big problem for me, but I just noticed difference in behavior.

[Eugene Pakhomov]

I just got bitten by goog.Uri.getPath:

(new goog.Uri('http://bidi.bidi/%25_%5C')).getPath()
=> "/%25_\"

I would expect either "/%25_%5C" or "%_\", but not something in between.

[Eugene Pakhomov]

Created google/closure-library#1083