Work with file paths, not URI substrings, in custom URI scheme

Using …_get_uri() gives a string that includes the anchor, which isn't a valid entry path in the archive. …_get_path() gets just the path (but with a leading slash).

Rather than special case the widget code to handle the starting slash, the archive and doc now ignore the leading slash in paths. This allows …_get_path() to be used anywhere in the code base without additional text manipulation each time.

Fixes #9 (closed) (Note: We don't need to do anything about the anchor, because Webkit does it for us once we load the stream correctly)