From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from vms173005pub.verizon.net ([206.46.173.5]) by linuxtogo.org with esmtp (Exim 4.72) (envelope-from ) id 1QKcbs-000400-RT; Thu, 12 May 2011 22:35:49 +0200 Received: from gandalf.denix.org ([unknown] [108.18.140.4]) by vms173005.mailsrvcs.net (Sun Java(tm) System Messaging Server 7u2-7.02 32bit (built Apr 16 2009)) with ESMTPA id <0LL300CR6NQC9TCT@vms173005.mailsrvcs.net>; Thu, 12 May 2011 15:32:42 -0500 (CDT) Received: by gandalf.denix.org (Postfix, from userid 1000) id A5EA014AF5B; Thu, 12 May 2011 16:32:36 -0400 (EDT) Date: Thu, 12 May 2011 16:32:36 -0400 From: Denys Dmytriyenko To: openembedded-devel@lists.openembedded.org Message-id: <20110512203236.GE21678@denix.org> References: <1305013151-17485-1-git-send-email-Martin.Jansa@gmail.com> <1305019695.8000.132.camel@phil-desktop> <20110510093616.GB3673@jama.jama.net> <20110512163919.GD21678@denix.org> <20110512182158.GA3162@jama.jama.net> MIME-version: 1.0 In-reply-to: <20110512182158.GA3162@jama.jama.net> User-Agent: Mutt/1.5.16 (2007-06-09) Cc: openembedded-core@lists.openembedded.org Subject: Re: [oe] [PATCH 1/2] recipes: fix LICENSE fields X-BeenThere: openembedded-core@lists.openembedded.org X-Mailman-Version: 2.1.11 Precedence: list Reply-To: Patches and discussions about the oe-core layer List-Id: Patches and discussions about the oe-core layer List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 12 May 2011 20:35:49 -0000 Content-type: text/plain; charset=us-ascii Content-disposition: inline On Thu, May 12, 2011 at 08:21:58PM +0200, Martin Jansa wrote: > On Thu, May 12, 2011 at 12:39:19PM -0400, Denys Dmytriyenko wrote: > > On Tue, May 10, 2011 at 11:36:16AM +0200, Martin Jansa wrote: > > > On Tue, May 10, 2011 at 10:28:15AM +0100, Phil Blundell wrote: > > > > On Tue, 2011-05-10 at 09:39 +0200, Martin Jansa wrote: > > > > > | /OE/tmpdir-shr/work/armv4t-oe-linux-gnueabi/pixman-0.22.0-r5.0/temp/run.SRC_DISTRIBUTECOMMAND.17340: line 95: /pixman: No such file or directory > > > > > > > > If '&' in ${LICENSE} is breaking SRC_DISTRIBUTECOMMAND then that > > > > suggests there is a bug in that code (probably some sort of missing > > > > quotes). Removing the &s might be a fine thing to do anyway, but it > > > > would also be a good idea to track down and fix the underlying bug. > > > > > > it calls cd tmp/deploy/sources/$license/pixman > > > where with $license == '&' (because LICENSE field is splitted by ' ' and > > > then '|'), it ends doing cd on background and then complaing about /pixman > > > > > > Directory named & even when created with right backslash > > > is still hard to use, so I think it would be better to have QA check > > > that LICENSE field has sane value, even better when we have list of > > > possible values etc. > > > > I would agree with Phil that fixing distribute_sources is required, but I'd go > > even further as to say that the current implementation of distribute_sources > > is quite bad. It splits the LICENSE field into tokens based on the wrong > > assumptions that all the whitespaces are used to separate multiple licenses. > > In other words, "Public Domain" is a valid license and shouldn't be replaced > > with "PD". What are you going to do with "GPLv2+ w/ GCC RLE"? Which stands for > > "GPL version 2 (or later) with GCC Runtime Library Exception" and is a valid > > > license with an exception... Sorry, but "GPL+libraryexception" is just bogus > > (and it's not a personal remark - I do understand your constrains, see below). > > Agreed, but do you find > LICENSE = "GPL + library exception" LICENSE = "GPL w/ library exception" > better? > > > I understand that distribute_sources is quite important and useful. So, maybe > > the correct way is to adopt a proper and stricter syntax for the LICENSE > > field, which uses some other characters to separate multiple licenses. Say, > > something like ":" or "|" should work. Using "," would be even better, but I'm > > '|' is alredy used for split in distribute_sources, but I'm not sure if > my interpretation that it means logical OR in LICENSE context is right > (and space as logical AND). Yes, I had already previously suggested using Debian's syntax for the license field, which allows complex combinations, such as OR/AND: http://thread.gmane.org/gmane.comp.handhelds.openembedded/38585/focus=38590 > > afraid it can easily be used inside the license name on its own... Basically, > > something along the lines of "GPL:MIT:Public Domain:Texas A&M Proprietary" etc. > > > > Any comments or objections? > > oe-core does more with LICENSEs (like LIC_FILES_CHKSUMs and keeping > common licenses in > /OE/shr-core/openembedded-core/meta/files/common-licenses which could be > used for QA check on LICENSE field) so I think it should be discussed on > oe-core ML first and then use the same syntax in oe.dev or just switch > to oe-core.. Yes, I'm quite aware of Yocto/oe-core handling of licenses and checksum verification mechanism for them, but unfortunately we don't have that in mainline classic OE... We could try backporting it from oe-core to classic OE or just wait until with switch to oe-core completely. In that case I'd suggest avoid mangling license fields just to satisfy broken distribute_sources... -- Denys From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from vms173005pub.verizon.net ([206.46.173.5]) by linuxtogo.org with esmtp (Exim 4.72) (envelope-from ) id 1QKcbs-000400-RT; Thu, 12 May 2011 22:35:49 +0200 Received: from gandalf.denix.org ([unknown] [108.18.140.4]) by vms173005.mailsrvcs.net (Sun Java(tm) System Messaging Server 7u2-7.02 32bit (built Apr 16 2009)) with ESMTPA id <0LL300CR6NQC9TCT@vms173005.mailsrvcs.net>; Thu, 12 May 2011 15:32:42 -0500 (CDT) Received: by gandalf.denix.org (Postfix, from userid 1000) id A5EA014AF5B; Thu, 12 May 2011 16:32:36 -0400 (EDT) Date: Thu, 12 May 2011 16:32:36 -0400 From: Denys Dmytriyenko To: openembedded-devel@lists.openembedded.org Message-id: <20110512203236.GE21678@denix.org> References: <1305013151-17485-1-git-send-email-Martin.Jansa@gmail.com> <1305019695.8000.132.camel@phil-desktop> <20110510093616.GB3673@jama.jama.net> <20110512163919.GD21678@denix.org> <20110512182158.GA3162@jama.jama.net> MIME-version: 1.0 In-reply-to: <20110512182158.GA3162@jama.jama.net> User-Agent: Mutt/1.5.16 (2007-06-09) Cc: openembedded-core@lists.openembedded.org Subject: Re: [PATCH 1/2] recipes: fix LICENSE fields X-BeenThere: openembedded-devel@lists.openembedded.org X-Mailman-Version: 2.1.11 Precedence: list Reply-To: openembedded-devel@lists.openembedded.org List-Id: Using the OpenEmbedded metadata to build Distributions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 12 May 2011 20:35:49 -0000 Content-type: text/plain; charset=us-ascii Content-disposition: inline On Thu, May 12, 2011 at 08:21:58PM +0200, Martin Jansa wrote: > On Thu, May 12, 2011 at 12:39:19PM -0400, Denys Dmytriyenko wrote: > > On Tue, May 10, 2011 at 11:36:16AM +0200, Martin Jansa wrote: > > > On Tue, May 10, 2011 at 10:28:15AM +0100, Phil Blundell wrote: > > > > On Tue, 2011-05-10 at 09:39 +0200, Martin Jansa wrote: > > > > > | /OE/tmpdir-shr/work/armv4t-oe-linux-gnueabi/pixman-0.22.0-r5.0/temp/run.SRC_DISTRIBUTECOMMAND.17340: line 95: /pixman: No such file or directory > > > > > > > > If '&' in ${LICENSE} is breaking SRC_DISTRIBUTECOMMAND then that > > > > suggests there is a bug in that code (probably some sort of missing > > > > quotes). Removing the &s might be a fine thing to do anyway, but it > > > > would also be a good idea to track down and fix the underlying bug. > > > > > > it calls cd tmp/deploy/sources/$license/pixman > > > where with $license == '&' (because LICENSE field is splitted by ' ' and > > > then '|'), it ends doing cd on background and then complaing about /pixman > > > > > > Directory named & even when created with right backslash > > > is still hard to use, so I think it would be better to have QA check > > > that LICENSE field has sane value, even better when we have list of > > > possible values etc. > > > > I would agree with Phil that fixing distribute_sources is required, but I'd go > > even further as to say that the current implementation of distribute_sources > > is quite bad. It splits the LICENSE field into tokens based on the wrong > > assumptions that all the whitespaces are used to separate multiple licenses. > > In other words, "Public Domain" is a valid license and shouldn't be replaced > > with "PD". What are you going to do with "GPLv2+ w/ GCC RLE"? Which stands for > > "GPL version 2 (or later) with GCC Runtime Library Exception" and is a valid > > > license with an exception... Sorry, but "GPL+libraryexception" is just bogus > > (and it's not a personal remark - I do understand your constrains, see below). > > Agreed, but do you find > LICENSE = "GPL + library exception" LICENSE = "GPL w/ library exception" > better? > > > I understand that distribute_sources is quite important and useful. So, maybe > > the correct way is to adopt a proper and stricter syntax for the LICENSE > > field, which uses some other characters to separate multiple licenses. Say, > > something like ":" or "|" should work. Using "," would be even better, but I'm > > '|' is alredy used for split in distribute_sources, but I'm not sure if > my interpretation that it means logical OR in LICENSE context is right > (and space as logical AND). Yes, I had already previously suggested using Debian's syntax for the license field, which allows complex combinations, such as OR/AND: http://thread.gmane.org/gmane.comp.handhelds.openembedded/38585/focus=38590 > > afraid it can easily be used inside the license name on its own... Basically, > > something along the lines of "GPL:MIT:Public Domain:Texas A&M Proprietary" etc. > > > > Any comments or objections? > > oe-core does more with LICENSEs (like LIC_FILES_CHKSUMs and keeping > common licenses in > /OE/shr-core/openembedded-core/meta/files/common-licenses which could be > used for QA check on LICENSE field) so I think it should be discussed on > oe-core ML first and then use the same syntax in oe.dev or just switch > to oe-core.. Yes, I'm quite aware of Yocto/oe-core handling of licenses and checksum verification mechanism for them, but unfortunately we don't have that in mainline classic OE... We could try backporting it from oe-core to classic OE or just wait until with switch to oe-core completely. In that case I'd suggest avoid mangling license fields just to satisfy broken distribute_sources... -- Denys