[REBOL] Parse failure (was newlines)
From: brett:codeconscious at: 9-Nov-2001 1:59
> the bug is a result of unexpected (for me) behaviour of PARSE:
>
> >> parse s: "0123456789" [8 skip p: (print p remove/part s 3) :s (print
> "OK") to end]
> 89
> == false
I narrowed down the problem similarly (your example is more concise). I
believe remove/part is the trigger for the problem. Most likely it leaves
Parse with an obsolete series reference to the input. The index of this
reference is beyond the tail of the input. I think that due to this, Parse
fails early - just after your first ")" and before the ":s".
So I think the real problem is that we do not have an opportunity early
enough to correct Parse's reference.
Which if confirmed, is a real issue because we lose some wonderul
manipulation opportunities that I thought we had!
Brett.
