[java vs js] StringTools.urlEncode/urlDecode behaves differently

[exFalso]

In the java backend StringTools.urlEncode encodes characters like ~ ( ) and ', whereas the js backend doesn't.
StringTools.urlDecode also works differently for non-ASCII characters, in particular java replaces them with U+FFFD, and javascript throws an exception.
Our workaround was to override the java std StringTools with code taken from https://github.com/mozilla/rhino/ which replicates the js behaviour.
Check out https://github.com/exFalso/haxe-qc/ for a QuickCheck-based test for these functions alongside Std.parseX functions.

[waneck]

Okay, I've taken a look at the implementation for rhino, but unfortunately the licenses aren't compatible, so I can't directly use it.

I've folowed some StackOverflow questions like https://stackoverflow.com/questions/14321873/java-url-encoding-urlencoder-vs-uri and https://stackoverflow.com/questions/607176/java-equivalent-to-javascripts-encodeuricomponent-that-produces-identical-outpu , and implemented a post-process step that should deal with inconsistencies.

Please let me know if you have more concrete examples of what fails on Java - or try to run your own tests and report back what you find!

[Simn]

I have disabled the test for now because it fails on too many targets to deal with it for the RC.