[patch][cla]dojo.hash doesn't set "trailingSpace " as the hash correctly

[miksago]

I've been working through writing better test cases for dojo.hash (#10003), and I've noticed that in the test case of setting a hash to a string with a trailingSpace that it isn't being set correctly.

dojo.hash('test ');
dojo.hash(); // should return "test "

Currently dojo.hash("test "); will set the hash to "test " instead of "test%20" which seems to be causing the test to fail in FF3.5

[miksago]

Proposed patch

[Adam Peller]

Standard practice is to comment out the assert until we get it working. You can put in a reference to this ticket.

[miksago]

The only reason the assert is in there was that it was in the original tests, but written incorrectly / inconsistently.

[Rob Retchless]

@miksago, this is actually behaving as expected. We decided to leave encoding up to the user to enable query-like hashes. If we did an explicit encodeURIComponent on every call, dojo.hash("param=blah&param2=blah2") would be encoded as "param%3Dblah%26param2%3Dblah2" - ugly. This is documented here: ​http://docs.dojocampus.org/dojo/hash#advanced

[miksago]

Okay, in that case, we'll need to note that dojo.hash() trims #'s from the start of urls, and spaces from the end of urls.