Return-Path: Received: from mail-ee0-f50.google.com ([74.125.83.50]:55317 "EHLO mail-ee0-f50.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758547Ab3BZUkw (ORCPT ); Tue, 26 Feb 2013 15:40:52 -0500 Received: by mail-ee0-f50.google.com with SMTP id e51so2808596eek.9 for ; Tue, 26 Feb 2013 12:40:51 -0800 (PST) Date: Tue, 26 Feb 2013 22:40:40 +0200 From: Ilya Dryomov To: Eric Sandeen Cc: kreijack@inwind.it, Goffredo Baroncelli , linux-btrfs@vger.kernel.org Subject: Re: [PATCH 17/17 V2] btrfs-progs: replace strtok_r with strsep Message-ID: <20130226204040.GA1993@zambezi.lan> References: <1361832890-40921-1-git-send-email-sandeen@redhat.com> <1361832890-40921-18-git-send-email-sandeen@redhat.com> <512D0325.5050308@gmail.com> <512D190E.2050002@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <512D190E.2050002@redhat.com> Sender: linux-btrfs-owner@vger.kernel.org List-ID: On Tue, Feb 26, 2013 at 02:20:30PM -0600, Eric Sandeen wrote: > The coverity runs had a false positive complaining that save_ptr > is uninitialized in the call to strtok_r. > > We could initialize it, but Zach points out that just using > strsep is a lot simpler if there's only one delimiter, > so just switch to that. > > Signed-off-by: Eric Sandeen > --- > > V2: Remove accidentally-added debug printfs, thanks Geoffredo! > > diff --git a/cmds-balance.c b/cmds-balance.c > index b671e1d..cfbb8eb 100644 > --- a/cmds-balance.c > +++ b/cmds-balance.c > @@ -67,11 +67,8 @@ static int parse_one_profile(const char *profile, u64 *flags) > static int parse_profiles(char *profiles, u64 *flags) > { > char *this_char; > - char *save_ptr; > > - for (this_char = strtok_r(profiles, "|", &save_ptr); > - this_char != NULL; > - this_char = strtok_r(NULL, "|", &save_ptr)) { > + while ((this_char = strsep(&profiles, "|"))) { > if (parse_one_profile(this_char, flags)) > return 1; > } > @@ -136,14 +133,11 @@ static int parse_filters(char *filters, struct btrfs_balance_args *args) > { > char *this_char; > char *value; > - char *save_ptr; > > if (!filters) > return 0; > > - for (this_char = strtok_r(filters, ",", &save_ptr); > - this_char != NULL; > - this_char = strtok_r(NULL, ",", &save_ptr)) { > + while ((this_char = strsep(&filters , ","))) { ^^^ whitespace One of the differences between strtok() and strsep() is that the former allows multiple delimiters between two tokens. With strsep(), this btrfs balance -dfoo1=bar1,,,foo2=bar2 fails with error, whereas with strtok() it passes. I don't have a strong opinion here (this has been loosely modeled on the way mount(8) handles -o options), but might it be better to just initialize save_ptr? (And yes, I know that strsep() is better ;)) Thanks, Ilya