From mboxrd@z Thu Jan 1 00:00:00 1970 From: Alexander Holler Subject: Re: [PATCH v2] parser: add kludges for "param-problem" and "redirect" Date: Mon, 06 Apr 2015 11:01:36 +0200 Message-ID: <55224B70.1070309@ahsoftware.de> References: <551FC211.6000907@ahsoftware.de> <1428145986-15421-1-git-send-email-holler@ahsoftware.de> <20150404115550.GA5832@salvia> <20150405113214.GA23433@acer.localdomain> <20150405121104.GD23433@acer.localdomain> <552187FB.60904@ahsoftware.de> <20150406015128.GA20515@acer.localdomain> <55224776.4040108@ahsoftware.de> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Cc: Pablo Neira Ayuso , netfilter-devel@vger.kernel.org, Arturo Borrero Gonzalez , Eric Leblond To: Patrick McHardy Return-path: Received: from h1446028.stratoserver.net ([85.214.92.142]:45419 "EHLO mail.ahsoftware.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752009AbbDFJBn (ORCPT ); Mon, 6 Apr 2015 05:01:43 -0400 Received: from wandq.ahsoftware (p4FC37568.dip0.t-ipconnect.de [79.195.117.104]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by mail.ahsoftware.de (Postfix) with ESMTPSA id 9E76D2C9C1BF for ; Mon, 6 Apr 2015 11:01:41 +0200 (CEST) In-Reply-To: <55224776.4040108@ahsoftware.de> Sender: netfilter-devel-owner@vger.kernel.org List-ID: Am 06.04.2015 um 10:44 schrieb Alexander Holler: > Am 06.04.2015 um 03:51 schrieb Patrick McHardy: >> On 05.04, Alexander Holler wrote: >>> Am 05.04.2015 um 14:11 schrieb Patrick McHardy: >>>> On 05.04, Patrick McHardy wrote: >>> >>>>> Basically this involves splitting the expression types into lhs >>>>> (non-const) >>>>> and rhs (const) parts. Keywords on the RHS side can be caught using an >>>>> error statement and deferred to resolution during runtime. >>> >>> Sounds like trial and error. ;) >> >> The approach is, the patch isn't, it changes the grammar to have >> these kinds of errors in a defined state. The patch I sent >> however is, but I'm quite sure i understand the implications. > > Just to mention it, there is still the possibility to define and use > keywords for all the icmp type names. Preferable with names as used in icmp.h and icmpv6.h. As these are defines in C-headers, there is very high probability that these names are unique, even across a large number of different sets of type or other names.