Re: Issue 1584 in lilypond: removing parser DIGIT rule

[David Kastrup]

address@hidden writes:

> Comment #2 on issue 1584 by address@hidden: removing parser DIGIT rule
> http://code.google.com/p/lilypond/issues/detail?id=1584
>
>> What is puzzling, however, is that the unpatched Lilypond works: the
>> same problem
>> (namely being unable to parse 0\mm) should occur since 0 is clearly
>> a DIGIT
>
> While reading input where 0\mm is legal, the lexer is in
> start-condition INITIAL, for which DIGIT is not a match condition.
> (The start-conditions are strangely named and strangely used, probably
> for historical reasons, but it is maddening to try to follow.)

My personal take on this would be that moving towards less reliance of
the lexer on starting conditions (in particular quirky ones like this)
would be desirable, even if it renders some previously valid input
relying on such fine distinctions invalid, as long as there are easily
understandable replacements.

-- 
David Kastrup