From mboxrd@z Thu Jan 1 00:00:00 1970 From: yamada.masahiro at socionext.com Date: Fri, 2 Nov 2018 07:55:14 +0000 Subject: [Buildroot] [PATCH] merge_config.sh: Fix merging buildroot config files In-Reply-To: <20181102021259.GA8966@smart-ThinkPad-T410> References: <20181031205608.GA4061@x230> <20181101010953.20996-1-afshin.nasser@gmail.com> <20181101062448.GC8261@x230> <20181101105513.GB4937@smart-ThinkPad-T410> <7ba255c0-e72e-7e02-d972-f8f082ce5214@mind.be> <20181101161918.GA30470@x230> <20181102021259.GA8966@smart-ThinkPad-T410> Message-ID: <5b88530949cb4aaf93436bcbc36f9fbf@SOC-EX01V.e01.socionext.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Hi. > -----Original Message----- > From: Nasser [mailto:afshin.nasser at gmail.com] > Sent: Friday, November 02, 2018 11:13 AM > To: Petr Vorel > Cc: Arnout Vandecappelle ; buildroot at buildroot.org; > Matthew Weber ; Angelo Compagnucci > ; Yamada, Masahiro/?? ?? > > Subject: Re: [PATCH] merge_config.sh: Fix merging buildroot config files > > Hi Petr, Arnout, > On Thu, Nov 01, 2018 at 05:19:19PM +0100, Petr Vorel wrote: > > Hi Arnout, > > > > > > Removing unset will cause the following make command work wrong. Note > > > > that the value of this environment variable is checked [1] and used. > The > > > > result of removing the unset line is that we will have double BR2_ > > > > prefixes in the final .config file as well as lots of warnings when > > > > checking if all specified config values have been taken. > > > > > We have no prefix, the BR2_ is mere convention. So when calling > merge_config, > > > we should set CONFIG_ to empty, not to BR2_. > > > > > So indeed, the unset should be removed. > > > > +1. I overlooked that regexp works with empty $CONFIG_ as well. > > Thanks for explanation. > As we discussed earlier [1], setting CONFIG_ to empty, makes most comments > to > be identified as config fragments and generate false warnings. I was also being worried about this. We could improve the sed pattern. How about this? https://patchwork.kernel.org/patch/10665119/ > > > > Kind regards, > > Petr > Kind regards, > Nasser > > [1] https://patchwork.ozlabs.org/comment/2018492/