If only `ParserMode.MENTION` was set as a parsing mode, the initial check for `input.endsWith(">")` would be bypassed and the input would begin to be parsed like a mention. If the input was less than three characters long then the `substring` calls would give a very ambiguous error along the lines of `begin 2, end X-1, length X`.
This commit fixes this issue by adding a check to make sure the input is an actual user mention. If the input isn't a mention then an exception will be set stating that the sender's input was not a user mention.
In addition, this commit also removes the `modes.size() == 1` check in the mention parser. I'm not sure why this was there to begin with, but I don't think that the input parsing check should be bypassed because only one parsing mode is specified. I settled on a check in the `UserParser` constructor to make sure that the parsing modes list isn't empty in order to ensure that at least one parser mode handles the input.
a2796bee04