Le 7 nov. 04, à 09:18, Robin Lee Powell a écrit :
Lojban has a reverse polish (i.e. postfix math) mode built in.
The ABNF form of this rule is:
rp-expression = rp-operand rp-operand operator
rp-operand = operand / rp-expression
which does exactly what you'd expect.
I can't figure out a way to convert this into a version without left recursion. I've got:
rp-expression = operand rp-expression-tail rp-expression-tail = rp-expression operator rp-expression-tail / ()
(where () is the empty production), but I'm not at all certain it's equivalent, and the parse trees it produces look very wrong (i.e. not left recursive at all).
Help?
What about :
rp-expression = operand operand rp-tail rp-tail = operand rp-tail / operator
?
-- Raphaël
Attachment:
PGP.sig
Description: =?ISO-8859-1?Q?Ceci_est_une_signature_=E9lectronique_PGP?=