[MLton] Long identifiers and def-use data

[Matthew Fluet]

>> How about reporting both the column of the (short) identifier and, when
>> present, the column of the long identifier?
>
> If you have the starting column of the short identifier, I think it's
> easy to scan the source backwards and forwards to find:
>
>  1. The start of the long identifier.
>  2. The end of the short identifier.
>  3. The end of the long identifier.
>
> Right?

I would think so, given that the regexp for the surrounding long id is 
nicely spelled out: (in ml-lex syntax):

alphanum=[A-Za-z'_0-9]*;
alphanumId=[A-Za-z]{alphanum};
sym=[-!%&$+/:<=>?@~`^|#*]|"\\";
symId={sym}+;
id={alphanumId}|{symId};
longid={id}("."{id})*;

> Assuming so, I'm all in favor of reporting the starting column of the
> short identifier.

Fine by me; my opinion wasn't particularly motivated, and if it makes the 
emacs mode that much better, I'm all for it.