From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753211Ab1LUJzp (ORCPT ); Wed, 21 Dec 2011 04:55:45 -0500 Received: from mx1.redhat.com ([209.132.183.28]:51468 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753161Ab1LUJzn (ORCPT ); Wed, 21 Dec 2011 04:55:43 -0500 Date: Wed, 21 Dec 2011 10:55:20 +0100 From: Jiri Olsa To: Arnaldo Carvalho de Melo Cc: a.p.zijlstra@chello.nl, mingo@elte.hu, paulus@samba.org, cjashfor@linux.vnet.ibm.com, linux-kernel@vger.kernel.org Subject: Re: [PATCH 1/3] perf, tool: Add parser generator for events parsing Message-ID: <20111221095520.GA1726@m.redhat.com> References: <1323963039-7602-1-git-send-email-jolsa@redhat.com> <1323963039-7602-2-git-send-email-jolsa@redhat.com> <20111220173721.GB22107@infradead.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20111220173721.GB22107@infradead.org> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Dec 20, 2011 at 03:37:22PM -0200, Arnaldo Carvalho de Melo wrote: > Em Thu, Dec 15, 2011 at 04:30:37PM +0100, Jiri Olsa escreveu: > > Changing event parsing to use flex/bison parse generator. > > The event syntax stays as it is. > > Seems really cool, using the right tool for the task, comments: > > . Can we avoid more globals? Like that __event_list one. ok, looks like there's a way to pass an parameter to the parse function, I'll chec and make the change if possible > . All those ABORT_ON do what? die() like stuff? /me googles YYABORT... > > ---- > YYABORT > Return immediately with value 1 (to report failure). > ---- > > Ok, so it gets propagated back and hopefully the we die() just > back at main() in tools/perf/perf.c, right? right, it'll make parse_events_parse return error > > . make help | perf > > Then running one of those targets and trying to build perf still > works? I guess so, but better check that. > > Having it added to some new "make perf-test" that does that + > 'perf test' too would get you brownie points. Leave it for later/someone > else if you're busy/not looking for more work, of course 8-) > > . We have some tests to parse events in 'perf test', right? From the top > of my head it was even ya that did that, guess those are ok/improved > (still need to look at the other patches in the backlog). right ;) thanks, jirka