[oe] [meta-oe][v2] 00/22] systemd rework

[Andreas Müller]

On Wed, Feb 22, 2012 at 2:15 PM, Martin Jansa <martin.jansa at gmail.com> wrote:
> On Wed, Feb 22, 2012 at 01:47:07PM +0100, Andreas Müller wrote:
>> On Mon, Feb 20, 2012 at 5:04 PM, Martin Jansa <martin.jansa at gmail.com> wrote:
>> > On Mon, Feb 20, 2012 at 04:56:46PM +0100, Andreas Müller wrote:
>> >> On Mon, Feb 20, 2012 at 1:40 PM, Martin Jansa <martin.jansa at gmail.com> wrote:
>> >> > On Thu, Feb 16, 2012 at 03:02:54AM +0100, Andreas Müller wrote:
>> >> >> V1 -> V2:
>> >> >> [2]  do not check configuration variables for cross and nativesdk
>> >> >> [3]  rename systemd_package to systemd_generate_package_scripts
>> >> >> [6]  parse unit files to determine the files to pack
>> >> >> [7]  create warning if *-systemd package already rdepends
>> >> >>      it's base package
>> >> >> added in V2:
>> >> >> [8]-[20] remove obsolet systemd code from recipes (these were used as
>> >> >>      test cases)
>> >> >> [21] dhcp-(relay) test case for multiple systemd packages
>> >> >> [22] modified to start service
>> >> >>
>> >> >> The following tests were performed:
>> >> >> * build from scratch with [1]-[7] / run test
>> >> >> * build from scratch / apply patches / build -> check buildhistory diff
>> >> >> * build from scratch w/o patches / build from scratch with patches -> check
>> >> >>   buildhistory diff
>> >> >>
>> >> >> Additional patches were prepared for other recipes inheriting systemd. These are
>> >> >> not yet send because they move systemd support from ${PN} to ${PN}-systemd which
>> >> >> currently needs modifications on tasks/images. I'd be happy if a common accepted
>> >> >> solution would be suggested/implemented (image-feature?).
>> >> >
>> >> > Today I've noticed one more issue (not sure if it's caused by this
>> >> > patchset or not).
>> >> I think it is
>> >> >
>> >> > Some images are failing for me with:
>> >> > For package xserver-nodm-init-systemd SYSTEMD_SERVICE-entry xserver-nodm.service does not exist
>> >> > from systemd.bbclass:populate_packages_prepend
>> >> >
>> >> > I guess it's because sstate package:
>> >> > sstate-xserver-nodm-init-all-oe-linux-2.0-r13-all-2-2ac030a94ddbd202f54b99c2e32c6312_package.tgz
>> >> >
>> >> > has only
>> >> > package/lib/systemd/system/xserver-nodm.service
>> >> >
>> >> > but systemd.bbclass:populate_packages_prepend is looking only in ${D}
>> >> > (image directory).
>> >> >
>> >> I assume you are working with temp files deleted (don't know the exact
>> >
>> > Yes temp files were deleted (and inherit rm_work).
>> >
>> >> config option now). To fix: Is there a variable pointing to package
>> >> dir?
>> >
>> > classes/package.bbclass:PKGD    = "${WORKDIR}/package"
>> >
>> > should help
>> >
>> I am working on this and checked rm_work.bbclass. Seems that complete
>> package contents are retained / respectively are recovered by sstate.
>> Hence I have another question:
>>
>> Will it cause problems to run the whole code in
>> populate_packages_prepend only (once) in case ${D} exists? Any fallout
>> to expect by this simple solution?
>
> If I understand you right, then looks good to me.
>
>> Please let me know if so because this would save me from long build tests.
>
> For me this whole series can go in as it is now and my comments can be
> implemented as followup patches (which will be easier to review then
> whole patchset again).
>
> Maybe only add that variable for /lib/systemd/system so I can push my
> "systemd rework" patchset for meta-smartphone with this variable
> already.
>
I am afraid I have embedded all comments already - will send a new
version after test...

* the modifications will be documented
* commits untouched will remain Otavio's ack

so efforts for another review are few.

Cheers

Andreas