Trailing whitespace not stripped when parsing SVN auth file

There's a bug in the parsing of the SVN auth file. Whitespace isn't stripped from the last token token on the line, so if you have an entry that looks like

[/]
@my-group = rw 

with a space after the 'rw' on the line, it will not be matched properly when working out the authz.

see CWDSUP-618