Re: EBNF conflict avoidance

"SLK Parsers" <slk12@earthlink.net>
19 Dec 2002 12:43:39 -0500

          From comp.compilers

Related articles
EBNF conflict avoidance slk12@earthlink.net (SLK Parsers) (2002-11-20)
Re: EBNF conflict avoidance clint@0lsen.net (Clint Olsen) (2002-11-24)
Re: EBNF conflict avoidance slk12@earthlink.net (SLK Parsers) (2002-12-11)
Re: EBNF conflict avoidance clint@0lsen.net (Clint Olsen) (2002-12-13)
Re: EBNF conflict avoidance slk12@earthlink.net (SLK Parsers) (2002-12-19)
| List of all articles for this month |
From: "SLK Parsers" <slk12@earthlink.net>
Newsgroups: comp.compilers
Date: 19 Dec 2002 12:43:39 -0500
Organization: Parsers Inc.
References: 02-11-118 02-11-134 02-12-073 02-12-081
Keywords: parse
Posted-Date: 19 Dec 2002 12:43:39 EST

> So, I would actually approach it from the standpoint of providing an
> init-block and then and action block in the event the rule is matched.


The roughly equivalent EBNF is


list -> _action_initialize { item _action_item } _action_finalize


which in BNF is


list :
        _action_initialize item_01_*


item_01_* :
        item _action_item item_01_*
        _action_finalize _epsilon_


So the init action really does not need special handling by the
parser-generator. Its placement is natural, whereas the finalize action
needs to be handled differently from the item action. In many cases, no
action is needed for the null production. The idea of having two separate
actions is to fully reflect the structure of the underlying BNF grammar. See
http://parsers.org/slk/example.html for a simple EBNF expression grammar
with actions.


http://parsers.org/slk


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.