From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail1.windriver.com (mail1.windriver.com [147.11.146.13]) by mail.openembedded.org (Postfix) with ESMTP id AF67560557 for ; Mon, 8 Dec 2014 23:54:01 +0000 (UTC) Received: from ALA-HCB.corp.ad.wrs.com (ala-hcb.corp.ad.wrs.com [147.11.189.41]) by mail1.windriver.com (8.14.9/8.14.5) with ESMTP id sB8Ns1Sp000739 (version=TLSv1/SSLv3 cipher=AES256-SHA bits=256 verify=FAIL) for ; Mon, 8 Dec 2014 15:54:01 -0800 (PST) Received: from msp-lpggp2.wrs.com (172.25.34.109) by ALA-HCB.corp.ad.wrs.com (147.11.189.41) with Microsoft SMTP Server id 14.3.174.1; Mon, 8 Dec 2014 15:54:00 -0800 From: Peter Seebach To: OE-core Date: Mon, 8 Dec 2014 17:53:58 -0600 Message-ID: X-Mailer: git-send-email 1.7.1 MIME-Version: 1.0 Subject: [PATCH 0/1] package.bbclass: Do fixups even when FILES_foo is set X-BeenThere: openembedded-core@lists.openembedded.org X-Mailman-Version: 2.1.12 Precedence: list List-Id: Patches and discussions about the oe-core layer List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 08 Dec 2014 23:54:04 -0000 Content-Type: text/plain There's some fixups for things like adding extra_depends to RDEPENDS which ended up in the "if not oldfiles" clause of a test against a setting for FILES_foo, but which I think should probably be done either way. The one that specifically affected us is that in some code for using a prebuilt toolchain, we set FILES_packagename from a manifest, which resulted in FILES_glibc-gconv-localename being set, which meant that RDEPENDS wasn't getting "glibc-gconv" appended for any of the glibc-gconv-localename packages, because the processing of extra_depends was conditional. Two of these variables (summary and description) looked like things where the assumption might have been that you'd define SUMMARY_foo and DESCRIPTION_foo if you were defining FILES_foo, so I've also made those settings conditional on there not being an existing setting. I ran builds of core-image-sato with and without this change, and looked at build history. 26 packages changed, but 23 were changes to PKGSIZE and nothing else. The other three changes: gst-plugins-good-gconfelements-dev has changes to RRECCOMMENDS, adding gst-plugins-good-dev, and gst-plugins-good-gconfelements has changes to RDEPENDS, adding gst-plugins-good. Both of these look reasonable to me. perl-module-cpan has a change to RDEPENDS, adding perl. I think that's also reasonable. So I don't think this breaks anything. The following changes since commit 219e793907406eabf632e784e3a11ab9acb77cfb: qemu/libc-package: Fix qemu option handling (2014-12-05 17:59:42 +0000) are available in the git repository at: git://git.yoctoproject.org/poky-contrib seebs/package_files http://git.yoctoproject.org/cgit.cgi/poky-contrib/log/?h=seebs/package_files Peter Seebach (1): package.bbclass: do variable fixups even when FILES was set meta/classes/package.bbclass | 18 ++++++++++-------- 1 files changed, 10 insertions(+), 8 deletions(-)