From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ozlabs.org ([103.22.144.67]:45265 "EHLO ozlabs.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751610AbaHTTrE (ORCPT ); Wed, 20 Aug 2014 15:47:04 -0400 From: Rusty Russell Subject: Re: [PATCH 1/3] modinst: wrap long lines in order to enhance cmd_modules_install In-Reply-To: References: <1408474660-24999-1-git-send-email-beber@meleeweb.net> <1408474660-24999-2-git-send-email-beber@meleeweb.net> Date: Thu, 21 Aug 2014 01:38:10 +0930 Message-ID: <87tx57i1p1.fsf@rustcorp.com.au> MIME-Version: 1.0 Content-Type: text/plain Sender: linux-kbuild-owner@vger.kernel.org List-ID: To: Jim Davis , Bertrand Jacquin Cc: linux-kbuild@vger.kernel.org, linux-kernel , Willy Tarreau Jim Davis writes: > On Tue, Aug 19, 2014 at 11:57 AM, Bertrand Jacquin wrote: >> Note: shouldn't we use 'install -D $(2)/$@ $@' instead of mkdir >> and cp ? > > Will an install with -D in that sense always be available? That's not > a flag in the original (BSD) program, and I did find one SPARC system > -- admittedly an ancient one -- where the install program had a -D > option but it did something else. install isn't listed as a required > program in Documentation/Changes, for that matter. Yes, we don't count on install. And ISTR that install -D would try to change perms on directories in the path, which can fail if you don't own them and aren't root... Cheers, Rusty.