From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from relay1.mentorg.com (relay1.mentorg.com [192.94.38.131]) by mail.openembedded.org (Postfix) with ESMTP id 4E662619DD for ; Mon, 15 Jul 2013 11:41:43 +0000 (UTC) Received: from svr-orw-fem-01.mgc.mentorg.com ([147.34.98.93]) by relay1.mentorg.com with esmtp id 1Uyh9z-00027V-Dj from Muhammad_Shakeel@mentor.com ; Mon, 15 Jul 2013 04:41:43 -0700 Received: from SVR-IES-FEM-01.mgc.mentorg.com ([137.202.0.104]) by svr-orw-fem-01.mgc.mentorg.com over TLS secured channel with Microsoft SMTPSVC(6.0.3790.4675); Mon, 15 Jul 2013 04:41:43 -0700 Received: from [137.202.157.119] (137.202.0.76) by SVR-IES-FEM-01.mgc.mentorg.com (137.202.0.104) with Microsoft SMTP Server (TLS) id 14.2.247.3; Mon, 15 Jul 2013 12:41:40 +0100 Message-ID: <51E3E01D.9050905@mentor.com> Date: Mon, 15 Jul 2013 16:42:21 +0500 From: Muhammad Shakeel User-Agent: Mozilla/5.0 (X11; Linux i686; rv:17.0) Gecko/20130623 Thunderbird/17.0.7 MIME-Version: 1.0 To: "Burton, Ross" References: <1373528728-27268-1-git-send-email-muhammad_shakeel@mentor.com> <51DEF360.8030509@linux.intel.com> <53609789-437E-429A-997E-7A1BECF21A48@mentor.com> <51DEFA34.70405@linux.intel.com> In-Reply-To: X-Originating-IP: [137.202.0.76] X-OriginalArrivalTime: 15 Jul 2013 11:41:43.0038 (UTC) FILETIME=[471701E0:01CE8150] Cc: "openembedded-core@lists.openembedded.org" Subject: Re: [PATCH] [v2] openssh: Enabling with systemd 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, 15 Jul 2013 11:41:43 -0000 Content-Type: multipart/alternative; boundary="------------010708000201040200040905" --------------010708000201040200040905 Content-Type: text/plain; charset="ISO-8859-1"; format=flowed Content-Transfer-Encoding: 7bit On 07/12/2013 09:32 PM, Burton, Ross wrote: > On 11 July 2013 19:32, Saul Wold wrote: >>> We are seeing different checks for different recipes. Do you >>> guys think that it is possible to handle them in bbclass. >>> >> How different are they, there is the DISTRO_FEATURE check and then >> typically the service file or the init file is sed'ed and >> installed. > > My initial thoughts were to keep it simple and have a > do_install_append() that deleted the sysvinit or systemd files > depending on DISTRO_FEATURES, leaving installing the files themselves > (from upstream, from a file, whatever) up to the recipe because > that's where the complications are. We can also easily move 'sed' part to systemd class because we will only be looking for all the service files in ${systemd_unitdir}, irrespective of how these got installed. I gave a little thought to move the installation part into the class and there can be a solution but probably little complicated. The neatest solution would be to define an another variable say 'SYSTEMD_SERVICE_FILES_{PN}' which will hold the name of all the service files, need to be installed, along with their relative path from WORKDIR. We can check in systemd class that if ${systemd_unitdir} is not present, means upstream hasn't installed anything, then install whatever is in 'SYSTEMD_SERVICE_FILES_{PN}'. Other option is to reuse current 'SYSTEMD_SERVICE_${PN}' that lists the service files in the package but doesn't include the path of service files. If we do not want a new variable, which seems like a duplication, then we can add the path with the service files in this variable and tweak systemd class to handle this in current implementation. We will add the path only if service file to be installed is not present in WORKDIR but buried somewhere in the package source. Please let me know of your thoughts about installation part. Best Regards, Shakeel --------------010708000201040200040905 Content-Type: text/html; charset="ISO-8859-1" Content-Transfer-Encoding: 7bit On 07/12/2013 09:32 PM, Burton, Ross wrote:
> On 11 July 2013 19:32, Saul Wold <sgw@linux.intel.com> wrote:
>>> We are seeing different checks for different recipes. Do you
>>> guys think that it is possible to handle them in bbclass.
>>>
>> How different are they, there is the DISTRO_FEATURE check and then
>> typically the service file or the init file is sed'ed and
>> installed.
>
> My initial thoughts were to keep it simple and have a
> do_install_append() that deleted the sysvinit or systemd files
> depending on DISTRO_FEATURES, leaving installing the files themselves
> (from upstream, from a file, whatever) up to the recipe because
> that's where the complications are.

We can also easily move 'sed' part to systemd class because we will only be
looking for all the service files in ${systemd_unitdir}, irrespective of
how these got installed.

I gave a little thought to move the installation part into the class and there
can be a solution but probably little complicated. The neatest solution would
be to define an another variable say 'SYSTEMD_SERVICE_FILES_{PN}'
which will hold the name of all the service files, need to be installed, along
with their relative path from WORKDIR. We can check in systemd class that if
${systemd_unitdir} is not present, means upstream hasn't installed anything,
then install whatever is in 'SYSTEMD_SERVICE_FILES_{PN}'.

Other option is to reuse current 'SYSTEMD_SERVICE_${PN}' that lists the service
files in the package but doesn't include the path of service files. If we do not
want a new variable, which seems like a duplication, then we can add the path
with the service files in this variable and tweak systemd class to handle this in
current implementation. We will add the path only if service file to be installed is
not present in WORKDIR but buried somewhere in the package source.

Please let me know of your thoughts about installation part.

Best Regards,
Shakeel

--------------010708000201040200040905--