From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754335AbbGEIpJ (ORCPT ); Sun, 5 Jul 2015 04:45:09 -0400 Received: from mail-oi0-f50.google.com ([209.85.218.50]:35472 "EHLO mail-oi0-f50.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750978AbbGEIpB (ORCPT ); Sun, 5 Jul 2015 04:45:01 -0400 MIME-Version: 1.0 In-Reply-To: References: Date: Sun, 5 Jul 2015 08:06:33 +0200 Message-ID: Subject: Re: [PATCH 1/2] kconfig: also warn of unhandled characters in statements From: Ulf Magnusson To: Andreas Ruprecht Cc: Paul Bolle , Valentin Rothberg , rafael.j.wysocki@intel.com, "open list:KCONFIG" , Kernel Mailing List , hengelein Stefan , linux@dominikbrodowski.net, Michal Marek Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org A more exact commit title would be something like "warn for unhandled characters before the initial token on a line". On Fri, Jul 3, 2015 at 2:46 PM, Andreas Ruprecht wrote: > While commit 2e0d737fc76f ("kconfig: don't silently ignore unhandled > characters") introduced a warning for unsupported characters inside > parameters, it does not cover situations where a statement > has additional characters around it. This could say something like "...introduces a warning for unhandled characters after the initial token on a line, but does not cover unhandled characters before the initial token, which are ignored using a different mechanism." > > This change introduces a warning if superfluous characters are found > around statements. As the 'help' statement sometimes is written as > '---help---', the '-' character would now also be regarded as > unhandled, this change also adds a special rule for this case. Similarly above. > > Reported-by: Valentin Rothberg > Signed-off-by: Andreas Ruprecht > --- > scripts/kconfig/zconf.l | 11 +++++++++-- > 1 file changed, 9 insertions(+), 2 deletions(-) > > diff --git a/scripts/kconfig/zconf.l b/scripts/kconfig/zconf.l > index 200a3fe..84a5d05 100644 > --- a/scripts/kconfig/zconf.l > +++ b/scripts/kconfig/zconf.l > @@ -106,7 +106,15 @@ n [A-Za-z0-9_] > zconflval.string = text; > return T_WORD; > } > - . > + "---help---" { > + /* Support old syntax for help statement */ > + return T_HELP; > + } I haven't tried it out to see if there would be problems, but adding '-' to the n [A-Za-z0-9_] definition in zconf.l and adding '---help---' to zconf.gperf seems cleaner than special-casing it in zconf.l. If '---help---' is treated as a token, then it makes sense to include '-' among the token characters. (I also rambled a bit in http://www.spinics.net/lists/linux-kbuild/msg11393.html.) With that change, '-' could be removed from the parameter regex ({n}|[-/.])+ in zconf.l too. > + . { > + fprintf(stderr, > + "%s:%d:warning: ignoring unsupported character '%c'\n", > + zconf_curname(), zconf_lineno(), *yytext); > + } > \n { > BEGIN(INITIAL); > current_file->lineno++; > @@ -132,7 +140,6 @@ n [A-Za-z0-9_] > BEGIN(STRING); > } > \n BEGIN(INITIAL); current_file->lineno++; return T_EOL; > - --- /* ignore */ > ({n}|[-/.])+ { > const struct kconf_id *id = kconf_id_lookup(yytext, yyleng); > if (id && id->flags & TF_PARAM) { > -- > 1.9.1 > Cheers, Ulf