[developers] token mapping determinism

[Stephan Oepen]

hi glenn, thanks for bringing up this question!

> can either result in
>
> a----b----a
>
> or
>
> a----b----b

i would expect the first of the above as the ‘correct’ result.  i
am not sure this is written down anywhere, but i recall peter
and me discussing cases like the above.  to resolve, i think
we landed on an expectation of left-to-right matching (though
i also vaguely recall there may be cases that require one to
stipulate additional ordering constraints).

i must confess, i am not sure i understand the second part
of your message.  can there be cases where an application
of a rule creates a new matching target for the same rule to
the left of the original match?

best wishes, oe