From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: weis
Received: (from weis@localhost) by pauillac.inria.fr (8.7.6/8.7.3) id QAA11958 for caml-redistribution; Wed, 2 Feb 2000 16:03:34 +0100 (MET)
Received: from nez-perce.inria.fr (nez-perce.inria.fr [192.93.2.78]) by pauillac.inria.fr (8.7.6/8.7.3) with ESMTP id CAA06620 for <caml-list@pauillac.inria.fr>; Tue, 1 Feb 2000 02:51:59 +0100 (MET)
Received: from www.nextsolution.co.jp (www.nextsolution.co.jp [202.33.212.2])
	by nez-perce.inria.fr (8.8.7/8.8.7) with ESMTP id CAA23417
	for <caml-list@inria.fr>; Tue, 1 Feb 2000 02:51:45 +0100 (MET)
Received: from sparc1.nextsolution.co.jp (sparc1 [202.235.80.1])
	by www.nextsolution.co.jp (8.9.3/3.7W) with SMTP id KAA12530
	for <caml-list@inria.fr>; Tue, 1 Feb 2000 10:48:10 +0900 (JST)
Received: from sparc1.nextsolution.co.jp by sparc1.nextsolution.co.jp (SMI-8.6/SMI-SVR4)
	id KAA02299; Tue, 1 Feb 2000 10:51:02 +0900
From: "Frank A. Christoph" <christo@nextsolution.co.jp>
To: "CAML List" <caml-list@inria.fr>
Subject: Stream parsers
Date: Tue, 1 Feb 2000 10:53:30 +0900
Message-ID: <000001bf6c57$23c30080$0150ebca@nextsolution.co.jp>
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3 (Normal)
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0
Importance: Normal
X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2919.6600
Sender: weis

I have a problem with reporting error locations in stream parsers. Using the
count at the beginning of a parser match, I can get the token number of the
first element that matched a clause. But when the parser expression looks
like this (using the new syntax):

  parser bp [< e = p; 'T ?? err bp >] -> ...

you have to bend over backwards to figure out what token number the error
could have occurred on, since p could have parsed an arbitrary number of
tokens already. Is there any easy way to work around this?

Would it be possible to change the semantics of the expression on the rhs of
?? so that it is a function which gets passed the token count after having
matched the previous stream component (p, in my example)? Or, if you want to
maintain retain the current semantics, add a new form with three ?'s...

--fac