Uri.removeFragment() corrupts URIs

[nex3]

If I run Uri.parse("foo:bar#baz").removeFragment().toString(), I get "foo://:0bar?", where I'd just expect "foo:bar". This is particularly problematic for working with data: URIs, which aren't at all structured like typical hierarchical URIs and whose fragments must be ignored.

[lrhn]

Good catch! Fixed.