RFC: meta-oe appends and overlayed recipes
Hi all,
I'd like to make an attempt to remove all appends and overlayed recipes from the meta-oe layer. As I've said previously, I don't believe meta-oe - as a collection of very useful additional recipes that many wish to be able to use on top of their OE-Core based build configurations - should be making any possibly unexpected changes to those configurations. Any such changes ought to be the province of distro layers alone.
We've already removed all of the obvious overlayed recipes and the meta- systemd split removed most of the pieces that were there for systemd support; there are just a few remaining recipes and appends that need to be dealt with. I'll catalogue these below with my comments.
Currently we have the following overlayed recipes:
* icon-naming-utils: meta-oe has a newer version (0.8.90 vs OE-Core's 0.8.7) and it uses BBCLASSEXTEND rather than OE-Core's native recipe. I would propose to just move the meta-oe version to OE-Core since it appears to be superior.
* libmad: both layers have the same version. The meta-oe version has some slightly different versions of the MIPS compiler flag fix and -fforce-mem removal patches but I think these can be ignored, since the OE-Core versions of these patches have proper headers and are presumably working. The OE-Core version has LICENSE_FLAGS that the meta-oe one doesn't, but is missing an avr32- specific optimisation patch that is in the meta-oe version. What would we do with the latter? Is it appropriate to add to the OE-Core recipe?
* tslib: OE-Core has the 1.0 release version, meta-oe has a git recipe that is ahead of 1.0; the OE-Core version has two patches not in the meta-oe version but that both have been merged upstream in the git revision being used in the meta-oe version. There is no newer stable release. What do we do here? Should we ask upstream (Chris) for a new stable release?
* xserver-nodm-init: the two versions are quite distinct. Not sure I understand the full history here but perhaps someone else can fill in the blanks...?
As far as bbappends go we have:
* meta-oe/recipes-core/busybox/busybox_1.20.2.bbappend
As far as I can tell this just adds an /etc/busybox-syslog.default file containing OPTIONS="-C64" and seems to have been added for systemd support. I'm not sure why this wasn't moved to meta-systemd, but I assume it needs to be merged into OE-Core now that systemd support is being added there... ?
* meta-oe/recipes-extended/polkit/polkit_0.104.bbappend
Another bbappend apparently for systemd support. Again, this should have been moved to meta-systemd; do we now need to merge it into OE-Core?
* meta-oe/recipes-qt/packagegroups/packagegroup-qte-toolchain-target.bbappend This is adding qwt to the qte toolchain. As far as I am concerned this is a distro policy decision - Qwt is a third-party library that is not part of Qt. I believe this should be moved to the layers for whichever distros want this.
* meta-oe/recipes-qt/qt4/qt4-x11-free_4.8.4.bbappend
* meta-oe/recipes-qt/qt4/qt4-embedded_4.8.4.bbappend
These two add MySQL and PostgreSQL support to Qt and Qt/Embedded. I see this as a distro policy decision; these should move to the layers for whichever distros want this. FWIW, this is particularly egregious if you've already built Qt, then add meta-oe and find Qt is being unexpectedly rebuilt.
* meta-oe/recipes-multimedia/gstreamer/gst-ffmpeg_0.10.13.bbappend Builds against external libav instead of using the builtin copy of ffmpeg, apparently for better performance on ARM (and presumably that is not the only benefit). It's less clear to me what should be done with this, but I'd still rather it could be eliminated. OE-Core does not have ffmpeg/libav; one wonders if it should or not.
Thoughts?
Cheers,
Paul
Signature · 2 lines
Paul Eggleton Intel Open Source Technology Centre
On Mon, Feb 11, 2013 at 3:09 PM, Paul Eggleton
<[email protected]> wrote:
...
Quoted 5 lines
Currently we have the following overlayed recipes:
* icon-naming-utils: meta-oe has a newer version (0.8.90 vs OE-Core's 0.8.7) and it uses BBCLASSEXTEND rather than OE-Core's native recipe. I would propose to just move the meta-oe version to OE-Core since it appears to be superior.
Agreed.
Quoted 7 lines
* libmad: both layers have the same version. The meta-oe version has some slightly different versions of the MIPS compiler flag fix and -fforce-mem removal patches but I think these can be ignored, since the OE-Core versions of these patches have proper headers and are presumably working. The OE-Core version has
LICENSE_FLAGSthat the meta-oe one doesn't, but is missing an avr32- specific optimisation patch that is in the meta-oe version. What would we do with the latter? Is it appropriate to add to the OE-Core recipe?
Well maybe clean it and keep the avr32 specific optimization?
Quoted 5 lines
* tslib: OE-Core has the 1.0 release version, meta-oe has a git recipe that is ahead of 1.0; the OE-Core version has two patches not in the meta-oe version but that both have been merged upstream in the git revision being used in the meta-oe version. There is no newer stable release. What do we do here? Should we ask upstream (Chris) for a new stable release?
It'd be good; in meanwhile we could move meta-oe version to OE-Core.
...
Quoted 7 lines
As far as bbappends go we have:
* meta-oe/recipes-core/busybox/busybox_1.20.2.bbappend
As far as I can tell this just adds an/etc/busybox-syslog.defaultfile containing OPTIONS="-C64" and seems to have been added for systemd support. I'm not sure why this wasn't moved to meta-systemd, but I assume it needs to be merged into OE-Core now that systemd support is being added there... ?
Agreed.
Quoted 3 lines
* meta-oe/recipes-extended/polkit/polkit_0.104.bbappend
Another bbappend apparently for systemd support. Again, this should have been moved to meta-systemd; do we now need to merge it into OE-Core?
Agreed.
Quoted 4 lines
* meta-oe/recipes-qt/packagegroups/packagegroup-qte-toolchain-target.bbappend This is adding qwt to the qte toolchain. As far as I am concerned this is a distro policy decision - Qwt is a third-party library that is not part of Qt. I believe this should be moved to the layers for whichever distros want this.
I think people using qte toolchain ought to comment on it.
Quoted 6 lines
* meta-oe/recipes-qt/qt4/qt4-x11-free_4.8.4.bbappend
* meta-oe/recipes-qt/qt4/qt4-embedded_4.8.4.bbappend
These two add MySQL and PostgreSQL support to Qt and Qt/Embedded. I see this as a distro policy decision; these should move to the layers for whichever distros want this. FWIW, this is particularly egregious if you've already built Qt, then add meta-oe and find Qt is being unexpectedly rebuilt.
Yes; I think I agree also because *most* people won't use these backends so better to handle this in the specific layers/distros/whatever needs it.
Quoted 6 lines
* meta-oe/recipes-multimedia/gstreamer/gst-ffmpeg_0.10.13.bbappend Builds against external libav instead of using the builtin copy of ffmpeg, apparently for better performance on ARM (and presumably that is not the only benefit). It's less clear to me what should be done with this, but I'd still rather it could be eliminated. OE-Core does not have ffmpeg/libav; one wonders if it should or not.
Dropping it would be a regression IMO so we might try to work in a solution to OE-Core?
Signature · 3 lines
Otavio Salvador O.S. Systems E-mail: [email protected] http://www.ossystems.com.br Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br
Quoted 11 lines · Paul Eggleton
* meta-oe/recipes-qt/packagegroups/packagegroup-qte-toolchain-target.bbappend This is adding qwt to the qte toolchain. As far as I am concerned this is a distro policy decision - Qwt is a third-party library that is not part of Qt. I believe this should be moved to the layers for whichever distros want this.
* meta-oe/recipes-qt/qt4/qt4-x11-free_4.8.4.bbappend
* meta-oe/recipes-qt/qt4/qt4-embedded_4.8.4.bbappend
These two add MySQL and PostgreSQL support to Qt and Qt/Embedded. I see this as a distro policy decision; these should move to the layers for whichever distros want this. FWIW, this is particularly egregious if you've already built Qt, then add meta-oe and find Qt is being unexpectedly rebuilt.
If these were implemented as PACKAGECONFIG options, then distros would just need to set the appropriate PACKAGECONFIG for the package in the distro config and we wouldn't even need the appends...
Cheers,
Richard
Quoted 10 lines · Otavio Salvador
On Mon, Feb 11, 2013 at 3:09 PM, Paul Eggleton
<[email protected]> wrote:
...Quoted 6 lines
* meta-oe/recipes-qt/qt4/qt4-x11-free_4.8.4.bbappend
* meta-oe/recipes-qt/qt4/qt4-embedded_4.8.4.bbappend
These two add MySQL and PostgreSQL support to Qt and Qt/Embedded. I see this as a distro policy decision; these should move to the layers for whichever distros want this. FWIW, this is particularly egregious if you've already built Qt, then add meta-oe and find Qt is being unexpectedly rebuilt.
Quoted 3 lines
Yes; I think I agree also because *most* people won't use these backends so better to handle this in the specific layers/distros/whatever needs it.
Definitely in favour of this.
That would simplify my own bbappends a lot, by letting me remove quite a few oe_filter_out's...
Cheers,
Anders
Signature · 2 lines
Anders Darander ChargeStorm AB / eStorm AB
Quoted 15 lines · On Monday 11 February 2013 22:35:47 Richard Purdie wrote:
Quoted 7 lines · Paul Eggleton
* meta-oe/recipes-qt/packagegroups/packagegroup-qte-toolchain-target.bbappe nd This is adding qwt to the qte toolchain. As far as I am concerned this is a distro policy decision - Qwt is a third-party library that is not part of Qt. I believe this should be moved to the layers for whichever distros want this.
Quoted 7 lines
* meta-oe/recipes-qt/qt4/qt4-x11-free_4.8.4.bbappend
* meta-oe/recipes-qt/qt4/qt4-embedded_4.8.4.bbappend
These two add MySQL and PostgreSQL support to Qt and Qt/Embedded. I see this as a distro policy decision; these should move to the layers for whichever distros want this. FWIW, this is particularly egregious if you've already built Qt, then add meta-oe and find Qt is being unexpectedly rebuilt.
Quoted 3 lines
If these were implemented as PACKAGECONFIG options, then distros would just need to set the appropriate PACKAGECONFIG for the package in the distro config and we wouldn't even need the appends...
The thing is the Qt configure options are a little more complicated - many of them are three-state switches (enable built-in, enable as a plugin or disabled). Thus we've opted to split the configuration options into variables for each type. We don't get PACKAGECONFIG's DEPENDS handling, but if we used PACKAGECONFIG we'd lose some flexibility.
Cheers,
Paul
Signature · 2 lines
Paul Eggleton Intel Open Source Technology Centre
Quoted 6 lines · Paul Eggleton
I'd like to make an attempt to remove all appends and overlayed recipes from the meta-oe layer. As I've said previously, I don't believe meta-oe - as a collection of very useful additional recipes that many wish to be able to use on top of their OE-Core based build configurations - should be making any possibly unexpected changes to those configurations. Any such changes ought to be the province of distro layers alone.
Hear hear!
Quoted 5 lines
* tslib: OE-Core has the 1.0 release version, meta-oe has a git recipe that is ahead of 1.0; the OE-Core version has two patches not in the meta-oe version but that both have been merged upstream in the git revision being used in the meta-oe version. There is no newer stable release. What do we do here? Should we ask upstream (Chris) for a new stable release?
Is anyone actually using tslib these days? oe-core dropped kdrive, and we don't package the apparently unmaintained input driver for Xorg. I guess a new upstream would be good, and then move to meta-oe.
Quoted 3 lines
* xserver-nodm-init: the two versions are quite distinct. Not sure I understand the full history here but perhaps someone else can fill in the blanks...?
I don't understand the full history either yet, but this is clearly something that needs to be sorted.
Quoted 5 lines
* meta-oe/recipes-core/busybox/busybox_1.20.2.bbappend
As far as I can tell this just adds an/etc/busybox-syslog.defaultfile containing OPTIONS="-C64" and seems to have been added for systemd support. I'm not sure why this wasn't moved to meta-systemd, but I assume it needs to be merged into OE-Core now that systemd support is being added there... ?
When it's understood *what* that does, then we can evaluate it for oe-core.
Quoted 3 lines
* meta-oe/recipes-extended/polkit/polkit_0.104.bbappend
Another bbappend apparently for systemd support. Again, this should have been moved to meta-systemd; do we now need to merge it into OE-Core?
Yes, half of it has been merged to master already. The rest should be in Radu's branch, we can sort that today.
Quoted 6 lines
* meta-oe/recipes-multimedia/gstreamer/gst-ffmpeg_0.10.13.bbappend Builds against external libav instead of using the builtin copy of ffmpeg, apparently for better performance on ARM (and presumably that is not the only benefit). It's less clear to me what should be done with this, but I'd still rather it could be eliminated. OE-Core does not have ffmpeg/libav; one wonders if it should or not.
libav/gst-ffmpeg/gst-av (as it's called in gst1.0) has interesting legal issues, but I do think it should be in oe-core.
Ross
Quoted 10 lines · Ross
Quoted 6 lines
* tslib: OE-Core has the 1.0 release version, meta-oe has a git recipe that is ahead of 1.0; the OE-Core version has two patches not in the meta-oe version but that both have been merged upstream in the git revision being used in the meta-oe version. There is no newer stable release. What do we do here? Should we ask upstream (Chris) for a new stable release?
Is anyone actually using tslib these days? oe-core dropped kdrive, and we don't package the apparently unmaintained input driver for Xorg. I guess a new upstream would be good, and then move to meta-oe.
Qt Embedded as we build it is currently configured to use tslib, as is SDL. If the alternative (evdev?) is supported they could presumably be switched over though. I don't know how practical that is however.
Quoted 7 lines
Quoted 3 lines
* xserver-nodm-init: the two versions are quite distinct. Not sure I understand the full history here but perhaps someone else can fill in the blanks...?
I don't understand the full history either yet, but this is clearly something that needs to be sorted.
Quoted 8 lines
Quoted 7 lines
* meta-oe/recipes-core/busybox/busybox_1.20.2.bbappend
As far as I can tell this just adds an/etc/busybox-syslog.defaultfile containing OPTIONS="-C64" and seems to have been added for systemd support.
I'm not sure why this wasn't moved to meta-systemd, but I assume it needs to be merged into OE-Core now that systemd support is being added there... ?When it's understood *what* that does, then we can evaluate it for oe-core.
The following commit introduced this:
http://cgit.openembedded.org/meta-openembedded/commit/?id=c48a6a605c6d8d38cfbc5df39b3dc310bffc07c1
Otavio can you explain further?
Quoted 6 lines
Quoted 3 lines
* meta-oe/recipes-extended/polkit/polkit_0.104.bbappend Another bbappend apparently for systemd support. Again, this should have been moved to meta-systemd; do we now need to merge it into OE-Core?
Yes, half of it has been merged to master already. The rest should be in Radu's branch, we can sort that today.
OK, great.
Quoted 9 lines
Quoted 6 lines
* meta-oe/recipes-multimedia/gstreamer/gst-ffmpeg_0.10.13.bbappend Builds against external libav instead of using the builtin copy of ffmpeg, apparently for better performance on ARM (and presumably that is not the only benefit). It's less clear to me what should be done with this, but I'd still rather it could be eliminated. OE-Core does not have ffmpeg/libav; one wonders if it should or not.
libav/gst-ffmpeg/gst-av (as it's called in gst1.0) has interesting legal issues, but I do think it should be in oe-core.
Well, we already have gst-ffmpeg in OE-Core so I can't imagine libav would be any worse as long as it is similarly protected with LICENSE_FLAGS.
Cheers,
Paul
Signature · 2 lines
Paul Eggleton Intel Open Source Technology Centre
Quoted 20 lines · Paul Eggleton
Quoted 15 lines · On Monday 11 February 2013 22:35:47 Richard Purdie wrote:
Quoted 7 lines · Paul Eggleton
* meta-oe/recipes-qt/packagegroups/packagegroup-qte-toolchain-target.bbappe nd This is adding qwt to the qte toolchain. As far as I am concerned this is a distro policy decision - Qwt is a third-party library that is not part of Qt. I believe this should be moved to the layers for whichever distros want this.
Quoted 7 lines
* meta-oe/recipes-qt/qt4/qt4-x11-free_4.8.4.bbappend
* meta-oe/recipes-qt/qt4/qt4-embedded_4.8.4.bbappend
These two add MySQL and PostgreSQL support to Qt and Qt/Embedded. I see this as a distro policy decision; these should move to the layers for whichever distros want this. FWIW, this is particularly egregious if you've already built Qt, then add meta-oe and find Qt is being unexpectedly rebuilt.Quoted 3 lines
If these were implemented as PACKAGECONFIG options, then distros would just need to set the appropriate PACKAGECONFIG for the package in the distro config and we wouldn't even need the appends...
Quoted 5 lines
The thing is the Qt configure options are a little more complicated - many of them are three-state switches (enable built-in, enable as a plugin or disabled). Thus we've opted to split the configuration options into variables for each type. We don't get PACKAGECONFIG's DEPENDS handling, but if we used PACKAGECONFIG we'd lose some flexibility.
Is there a way we can at least have it behave in a similar way to PACKAGECONFIG? Can those configuration variables be set from the distro?
My main point is that I'd like to not need bbappend files just for configuration.
Cheers,
Richard
Quoted 31 lines · On Tuesday 12 February 2013 13:10:05 Richard Purdie wrote:
Quoted 24 lines · Paul Eggleton
Quoted 19 lines · On Monday 11 February 2013 22:35:47 Richard Purdie wrote:
Quoted 9 lines · Paul Eggleton
* meta-oe/recipes-qt/packagegroups/packagegroup-qte-toolchain-target.bba ppe
nd This is adding qwt to the qte toolchain. As far as I am concerned this
is a distro policy decision - Qwt is a third-party library that is not part of Qt. I believe this should be moved to the layers for whichever distros want this.Quoted 8 lines
* meta-oe/recipes-qt/qt4/qt4-x11-free_4.8.4.bbappend
* meta-oe/recipes-qt/qt4/qt4-embedded_4.8.4.bbappend
These two add MySQL and PostgreSQL support to Qt and Qt/Embedded. I see
this as a distro policy decision; these should move to the layers for whichever distros want this. FWIW, this is particularly egregious if you've already built Qt, then add meta-oe and find Qt is being unexpectedly rebuilt.Quoted 3 lines
If these were implemented as PACKAGECONFIG options, then distros would just need to set the appropriate PACKAGECONFIG for the package in the distro config and we wouldn't even need the appends...
Quoted 5 lines
The thing is the Qt configure options are a little more complicated - many of them are three-state switches (enable built-in, enable as a plugin or disabled). Thus we've opted to split the configuration options into variables for each type. We don't get PACKAGECONFIG's DEPENDS handling, but if we used PACKAGECONFIG we'd lose some flexibility.
Quoted 3 lines
Is there a way we can at least have it behave in a similar way to PACKAGECONFIG? Can those configuration variables be set from the distro?
Well, the default values are set using ?= in qt4.inc, so there's nothing stopping them from being set from the distro config.
Cheers,
Paul
Signature · 2 lines
Paul Eggleton Intel Open Source Technology Centre
Quoted 7 lines · Ross
Quoted 3 lines
* meta-oe/recipes-extended/polkit/polkit_0.104.bbappend
Another bbappend apparently for systemd support. Again, this should have been moved to meta-systemd; do we now need to merge it into OE-Core?Yes, half of it has been merged to master already. The rest should be in Radu's branch, we can sort that today.
I take that back - polkit in oe-core supports systemd if enabled, so this append can be removed.
Ross
Quoted 8 lines · Paul Eggleton
Quoted 7 lines · Ross
Quoted 6 lines
* tslib: OE-Core has the 1.0 release version, meta-oe has a git recipe that is ahead of 1.0; the OE-Core version has two patches not in the meta-oe version but that both have been merged upstream in the git revision being used in the meta-oe version. There is no newer stable release. What do we do here? Should we ask upstream (Chris) for a new stable release?
Quoted 7 lines
Quoted 3 lines
Is anyone actually using tslib these days? oe-core dropped kdrive, and we don't package the apparently unmaintained input driver for Xorg. I guess a new upstream would be good, and then move to meta-oe.
Qt Embedded as we build it is currently configured to use tslib, as is SDL. If the alternative (evdev?) is supported they could presumably be switched over though. I don't know how practical that is however.
Yes Qt Embedded uses it; not sure evdev is an alternative here.
Quoted 14 lines
Quoted 14 lines
Quoted 3 lines
* xserver-nodm-init: the two versions are quite distinct. Not sure I understand the full history here but perhaps someone else can fill in the blanks...?
I don't understand the full history either yet, but this is clearly something that needs to be sorted.
Quoted 7 lines
* meta-oe/recipes-core/busybox/busybox_1.20.2.bbappend
As far as I can tell this just adds an/etc/busybox-syslog.defaultfile containing OPTIONS="-C64" and seems to have been added for systemd support.
I'm not sure why this wasn't moved to meta-systemd, but I assume it needs to be merged into OE-Core now that systemd support is being added there... ?
Quoted 7 lines
Quoted 1 line
When it's understood *what* that does, then we can evaluate it for oe-core.
The following commit introduced this:
http://cgit.openembedded.org/meta-openembedded/commit/?id=c48a6a605c6d8d38cfbc5df39b3dc310bffc07c1
Otavio can you explain further?
This was to make the service to use the /etc/default/busybox-syslog to configure syslog. Today with the availability of journald I not sure it still has an use. It might be dropped I think.
Quoted 30 lines
Quoted 6 lines
Quoted 3 lines
* meta-oe/recipes-extended/polkit/polkit_0.104.bbappend Another bbappend apparently for systemd support. Again, this should have been moved to meta-systemd; do we now need to merge it into OE-Core?
Yes, half of it has been merged to master already. The rest should be in Radu's branch, we can sort that today.
OK, great.
Quoted 9 lines
Quoted 6 lines
* meta-oe/recipes-multimedia/gstreamer/gst-ffmpeg_0.10.13.bbappend Builds against external libav instead of using the builtin copy of ffmpeg, apparently for better performance on ARM (and presumably that is not the only benefit). It's less clear to me what should be done with this, but I'd still rather it could be eliminated. OE-Core does not have ffmpeg/libav; one wonders if it should or not.
libav/gst-ffmpeg/gst-av (as it's called in gst1.0) has interesting legal issues, but I do think it should be in oe-core.
Well, we already have gst-ffmpeg in OE-Core so I can't imagine libav would be any worse as long as it is similarly protected with
LICENSE_FLAGS.Cheers,
Paul--
Paul Eggleton
Intel Open Source Technology Centre
Signature · 3 lines
Otavio Salvador O.S. Systems E-mail: [email protected] http://www.ossystems.com.br Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br
Quoted 10 lines · Ross Burton
Quoted 7 lines · Ross
Quoted 3 lines
* meta-oe/recipes-extended/polkit/polkit_0.104.bbappend
Another bbappend apparently for systemd support. Again, this should have been moved to meta-systemd; do we now need to merge it into OE-Core?Yes, half of it has been merged to master already. The rest should be in Radu's branch, we can sort that today.
I take that back - polkit in oe-core supports systemd if enabled, so this append can be removed.
In fact we cannot drop the bbappend or we break the upgrade path (except if we bump PR in OE-Core to keep it).
Signature · 3 lines
Otavio Salvador O.S. Systems E-mail: [email protected] http://www.ossystems.com.br Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br
Quoted 7 lines · On Tuesday 12 February 2013 14:44:58 Otavio Salvador wrote:
Quoted 6 lines · Ross Burton
Quoted 5 lines · Ross
Quoted 4 lines
* meta-oe/recipes-extended/polkit/polkit_0.104.bbappend Another bbappend apparently for systemd support. Again, this should have been moved to meta-systemd; do we now need to merge it into OE-Core?
Quoted 8 lines
Quoted 5 lines
Quoted 2 lines
Yes, half of it has been merged to master already. The rest should be in Radu's branch, we can sort that today.
I take that back - polkit in oe-core supports systemd if enabled, so this append can be removed.
In fact we cannot drop the bbappend or we break the upgrade path (except if we bump PR in OE-Core to keep it).
These do need to be dropped; if bumping PR in OE-Core is what has to happen to achieve that then so be it.
Cheers,
Paul
Signature · 2 lines
Paul Eggleton Intel Open Source Technology Centre
Quoted 28 lines · Paul Eggleton
Hi all,
I'd like to make an attempt to remove all appends and overlayed recipes from the meta-oe layer. As I've said previously, I don't believe meta-oe - as a collection of very useful additional recipes that many wish to be able to use on top of their OE-Core based build configurations - should be making any possibly unexpected changes to those configurations. Any such changes ought to be the province of distro layers alone.
We've already removed all of the obvious overlayed recipes and the meta- systemd split removed most of the pieces that were there for systemd support; there are just a few remaining recipes and appends that need to be dealt with. I'll catalogue these below with my comments.
Currently we have the following overlayed recipes:
* icon-naming-utils: meta-oe has a newer version (0.8.90 vs OE-Core's 0.8.7) and it uses BBCLASSEXTEND rather than OE-Core's native recipe. I would propose to just move the meta-oe version to OE-Core since it appears to be superior.
* libmad: both layers have the same version. The meta-oe version has some slightly different versions of the MIPS compiler flag fix and -fforce-mem removal patches but I think these can be ignored, since the OE-Core versions of these patches have proper headers and are presumably working. The OE-Core version has
LICENSE_FLAGSthat the meta-oe one doesn't, but is missing an avr32- specific optimisation patch that is in the meta-oe version. What would we do with the latter? Is it appropriate to add to the OE-Core recipe?
Quoted 5 lines
* tslib: OE-Core has the 1.0 release version, meta-oe has a git recipe that is ahead of 1.0; the OE-Core version has two patches not in the meta-oe version but that both have been merged upstream in the git revision being used in the meta-oe version. There is no newer stable release. What do we do here? Should we ask upstream (Chris) for a new stable release?
tslib is also requested on devices where kernel driver returns too much noise, tslib can filter that, evdev needs separate filter like
http://atrey.karlin.mff.cuni.cz/~metan/evfilter/
but that isn't integrated in OE.
Quoted 3 lines
* xserver-nodm-init: the two versions are quite distinct. Not sure I understand the full history here but perhaps someone else can fill in the blanks...?
The biggest difference is integrated xinput-calibrator and use of xserver-common_1.34.bb
Quoted 17 lines
As far as bbappends go we have:
* meta-oe/recipes-core/busybox/busybox_1.20.2.bbappend
As far as I can tell this just adds an/etc/busybox-syslog.defaultfile containing OPTIONS="-C64" and seems to have been added for systemd support. I'm not sure why this wasn't moved to meta-systemd, but I assume it needs to be merged into OE-Core now that systemd support is being added there... ?* meta-oe/recipes-extended/polkit/polkit_0.104.bbappend
Another bbappend apparently for systemd support. Again, this should have been moved to meta-systemd; do we now need to merge it into OE-Core?* meta-oe/recipes-qt/packagegroups/packagegroup-qte-toolchain-target.bbappend This is adding qwt to the qte toolchain. As far as I am concerned this is a distro policy decision - Qwt is a third-party library that is not part of Qt. I believe this should be moved to the layers for whichever distros want this.
Quoted 6 lines
* meta-oe/recipes-qt/qt4/qt4-x11-free_4.8.4.bbappend
* meta-oe/recipes-qt/qt4/qt4-embedded_4.8.4.bbappend
These two add MySQL and PostgreSQL support to Qt and Qt/Embedded. I see this as a distro policy decision; these should move to the layers for whichever distros want this. FWIW, this is particularly egregious if you've already built Qt, then add meta-oe and find Qt is being unexpectedly rebuilt.
MySQL and PostgreSQL are not in oe-core so it cannot be replaced with simple PACKAGECONFIG option in oe-core recipe, but I also prefer to share such .bbappend somewhere instead of every distribution reinventing the wheel when trying to enable something as simple as db in qt.
Quoted 6 lines
* meta-oe/recipes-multimedia/gstreamer/gst-ffmpeg_0.10.13.bbappend Builds against external libav instead of using the builtin copy of ffmpeg, apparently for better performance on ARM (and presumably that is not the only benefit). It's less clear to me what should be done with this, but I'd still rather it could be eliminated. OE-Core does not have ffmpeg/libav; one wonders if it should or not.
Signature · 1 line
Martin 'JaMa' Jansa jabber: [email protected]
Quoted 71 lines · Martin Jansa
Quoted 28 lines · Paul Eggleton
Hi all,
I'd like to make an attempt to remove all appends and overlayed recipes from the meta-oe layer. As I've said previously, I don't believe meta-oe - as a collection of very useful additional recipes that many wish to be able to use on top of their OE-Core based build configurations - should be making any possibly unexpected changes to those configurations. Any such changes ought to be the province of distro layers alone.
We've already removed all of the obvious overlayed recipes and the meta- systemd split removed most of the pieces that were there for systemd support; there are just a few remaining recipes and appends that need to be dealt with. I'll catalogue these below with my comments.
Currently we have the following overlayed recipes:
* icon-naming-utils: meta-oe has a newer version (0.8.90 vs OE-Core's 0.8.7) and it uses BBCLASSEXTEND rather than OE-Core's native recipe. I would propose to just move the meta-oe version to OE-Core since it appears to be superior.
* libmad: both layers have the same version. The meta-oe version has some slightly different versions of the MIPS compiler flag fix and -fforce-mem removal patches but I think these can be ignored, since the OE-Core versions of these patches have proper headers and are presumably working. The OE-Core version has
LICENSE_FLAGSthat the meta-oe one doesn't, but is missing an avr32- specific optimisation patch that is in the meta-oe version. What would we do with the latter? Is it appropriate to add to the OE-Core recipe?Quoted 5 lines
* tslib: OE-Core has the 1.0 release version, meta-oe has a git recipe that is ahead of 1.0; the OE-Core version has two patches not in the meta-oe version but that both have been merged upstream in the git revision being used in the meta-oe version. There is no newer stable release. What do we do here? Should we ask upstream (Chris) for a new stable release?
tslib is also requested on devices where kernel driver returns too much noise, tslib can filter that, evdev needs separate filter like
http://atrey.karlin.mff.cuni.cz/~metan/evfilter/
but that isn't integrated in OE.Quoted 3 lines
* xserver-nodm-init: the two versions are quite distinct. Not sure I understand the full history here but perhaps someone else can fill in the blanks...?
The biggest difference is integrated xinput-calibrator and use of xserver-common_1.34.bb
Quoted 17 lines
As far as bbappends go we have:
* meta-oe/recipes-core/busybox/busybox_1.20.2.bbappend
As far as I can tell this just adds an/etc/busybox-syslog.defaultfile containing OPTIONS="-C64" and seems to have been added for systemd support. I'm not sure why this wasn't moved to meta-systemd, but I assume it needs to be merged into OE-Core now that systemd support is being added there... ?* meta-oe/recipes-extended/polkit/polkit_0.104.bbappend
Another bbappend apparently for systemd support. Again, this should have been moved to meta-systemd; do we now need to merge it into OE-Core?* meta-oe/recipes-qt/packagegroups/packagegroup-qte-toolchain-target.bbappend This is adding qwt to the qte toolchain. As far as I am concerned this is a distro policy decision - Qwt is a third-party library that is not part of Qt. I believe this should be moved to the layers for whichever distros want this.
Quoted 6 lines
* meta-oe/recipes-qt/qt4/qt4-x11-free_4.8.4.bbappend
* meta-oe/recipes-qt/qt4/qt4-embedded_4.8.4.bbappend
These two add MySQL and PostgreSQL support to Qt and Qt/Embedded. I see this as a distro policy decision; these should move to the layers for whichever distros want this. FWIW, this is particularly egregious if you've already built Qt, then add meta-oe and find Qt is being unexpectedly rebuilt.
Quoted 4 lines
MySQL and PostgreSQL are not in oe-core so it cannot be replaced with simple PACKAGECONFIG option in oe-core recipe, but I also prefer to share such .bbappend somewhere instead of every distribution reinventing the wheel when trying to enable something as simple as db in qt.
So maybe meta-oe ought to add PACKAGECONFIG options?
Quoted 15 lines
Quoted 6 lines
* meta-oe/recipes-multimedia/gstreamer/gst-ffmpeg_0.10.13.bbappend Builds against external libav instead of using the builtin copy of ffmpeg, apparently for better performance on ARM (and presumably that is not the only benefit). It's less clear to me what should be done with this, but I'd still rather it could be eliminated. OE-Core does not have ffmpeg/libav; one wonders if it should or not.
--
Martin 'JaMa' Jansa jabber: [email protected]_______________________________________________
Openembedded-devel mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
Signature · 3 lines
Otavio Salvador O.S. Systems E-mail: [email protected] http://www.ossystems.com.br Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br
Quoted 8 lines · On Tuesday 12 February 2013 20:19:02 Martin Jansa wrote:
Quoted 6 lines · Paul Eggleton
* tslib: OE-Core has the 1.0 release version, meta-oe has a git recipe that is ahead of 1.0; the OE-Core version has two patches not in the meta-oe version but that both have been merged upstream in the git revision being used in the meta-oe version. There is no newer stable release. What do we do here? Should we ask upstream (Chris) for a new stable release?
Quoted 4 lines
tslib is also requested on devices where kernel driver returns too much noise, tslib can filter that, evdev needs separate filter like
http://atrey.karlin.mff.cuni.cz/~metan/evfilter/
but that isn't integrated in OE.
OK. I think the best course of action is to petition upstream for a stable release and then update OE-Core to that, then we can drop the recipe from meta-oe without ill effect.
Quoted 6 lines
Quoted 3 lines
* xserver-nodm-init: the two versions are quite distinct. Not sure I understand the full history here but perhaps someone else can fill in the blanks...?
The biggest difference is integrated xinput-calibrator and use of xserver-common_1.34.bb
So can you see a means for us to move this into OE-Core? Where are the points of conflict?
Quoted 8 lines
Quoted 7 lines
* meta-oe/recipes-qt/qt4/qt4-x11-free_4.8.4.bbappend
* meta-oe/recipes-qt/qt4/qt4-embedded_4.8.4.bbappend
These two add MySQL and PostgreSQL support to Qt and Qt/Embedded. I see this as a distro policy decision; these should move to the layers for whichever distros want this. FWIW, this is particularly egregious if you've already built Qt, then add meta-oe and find Qt is being unexpectedly rebuilt.
Quoted 4 lines
MySQL and PostgreSQL are not in oe-core so it cannot be replaced with simple PACKAGECONFIG option in oe-core recipe, but I also prefer to share such .bbappend somewhere instead of every distribution reinventing the wheel when trying to enable something as simple as db in qt.
We kind of discussed this already - PACKAGECONFIG can't work for this anyway because at least for the database drivers it's a tri-state switch. At the moment all these bbappends really do is set QT_SQL_DRIVER_FLAGS, DEPENDS and set the correct include directory. This is not a particularly advanced mechanism, is unlikely to change, and I don't think it's too much to ask for this to be set in each distro layer. This is distro policy and should not be expressed in meta-oe.
Cheers,
Paul
Signature · 2 lines
Paul Eggleton Intel Open Source Technology Centre
Quoted 5 lines · Martin Jansa
Quoted 3 lines
* meta-oe/recipes-qt/qt4/qt4-x11-free_4.8.4.bbappend
* meta-oe/recipes-qt/qt4/qt4-embedded_4.8.4.bbappend
These two add MySQL and PostgreSQL support to Qt and Qt/Embedded. I seethis
Quoted 1 line
as a distro policy decision; these should move to the layers for
Quoted 8 lines
whichever
Quoted 2 lines
distros want this. FWIW, this is particularly egregious if you've already built Qt, then add meta-oe and find Qt is being unexpectedly rebuilt.
MySQL and PostgreSQL are not in oe-core so it cannot be replaced with simple PACKAGECONFIG option in oe-core recipe, but I also prefer to share such .bbappend somewhere instead of every distribution reinventing the wheel when trying to enable something as simple as db in qt.
I don't see any reason the oe-core recipe couldn't provide the packageconfig options, even missing the dependencies. They are valid package configuration options, so they should exist in PACKAGECONFIG. If a distro wants to enable one of them, however, they'd better include the layers that provide the deps :)
Signature · 1 line
Christopher Larson
Quoted 9 lines · Paul Eggleton
Quoted 8 lines · On Tuesday 12 February 2013 20:19:02 Martin Jansa wrote:
Quoted 6 lines · Paul Eggleton
* tslib: OE-Core has the 1.0 release version, meta-oe has a git recipe that is ahead of 1.0; the OE-Core version has two patches not in the meta-oe version but that both have been merged upstream in the git revision being used in the meta-oe version. There is no newer stable release. What do we do here? Should we ask upstream (Chris) for a new stable release?
Quoted 8 lines
Quoted 4 lines
tslib is also requested on devices where kernel driver returns too much noise, tslib can filter that, evdev needs separate filter like
http://atrey.karlin.mff.cuni.cz/~metan/evfilter/
but that isn't integrated in OE.OK. I think the best course of action is to petition upstream for a stable release and then update OE-Core to that, then we can drop the recipe from meta-oe without ill effect.
Agreed,
Quoted 9 lines
Quoted 6 lines
Quoted 3 lines
* xserver-nodm-init: the two versions are quite distinct. Not sure I understand the full history here but perhaps someone else can fill in the blanks...?
The biggest difference is integrated xinput-calibrator and use of xserver-common_1.34.bb
So can you see a means for us to move this into OE-Core? Where are the points of conflict?
xinput-calibrator is now being reworked by Andreas so we should at least wait for this rework to settle down and xserver-common is similar to formfactor, a bit more flexible but a lot worse to maintain for new MACHINES (FWIW all local patches are waiting in florian mbox).
Signature · 1 line
Martin 'JaMa' Jansa jabber: [email protected]
Quoted 17 lines · Chris Larson
On Tue, Feb 12, 2013 at 12:19 PM, Martin Jansa <[email protected] <mailto:[email protected]>> wrote:
> * meta-oe/recipes-qt/qt4/qt4-x11-free_4.8.4.bbappend
> * meta-oe/recipes-qt/qt4/qt4-embedded_4.8.4.bbappend
> These two add MySQL and PostgreSQL support to Qt and Qt/Embedded. I see this > as a distro policy decision; these should move to the layers for whichever > distros want this. FWIW, this is particularly egregious if you've already > built Qt, then add meta-oe and find Qt is being unexpectedly rebuilt.MySQL and PostgreSQL are not in oe-core so it cannot be replaced with simple PACKAGECONFIG option in oe-core recipe, but I also prefer to share such .bbappend somewhere instead of every distribution reinventing the wheel when trying to enable something as simple as db in qt.
Quoted 4 lines
I don't see any reason the oe-core recipe couldn't provide the packageconfig options, even missing the dependencies. They are valid package configuration options, so they should exist in PACKAGECONFIG. If a distro wants to enable one of them, however, they'd better include the layers that provide the deps :)
RPM and Smart integrations both have configuration options that are invalid for most configurations or sometimes need additional software provided via a layer.
I think this is acceptable, as long as the configuration selects the proper
DEPENDS/RDEPENDS to ensure the items work properly in the end.
--Mark
Quoted 9 lines
--
Christopher Larson_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core
W dniu 12.02.2013 20:51, Martin Jansa pisze:
Quoted 2 lines
xserver-common is similar to formfactor, a bit more flexible but a lot worse to maintain for new MACHINES
IIRC xserver-common allows to set all MACHINE related variables in /etc/default/something file. I added that years ago when merged xserver-common with xserver-kdrive-common with all OE/Poky patches etc.
Quoted 5 lines · Otavio Salvador
Quoted 3 lines
Qt Embedded as we build it is currently configured to use tslib, as is SDL. If the alternative (evdev?) is supported they could presumably be switched over though. I don't know how practical that is however.
Yes Qt Embedded uses it; not sure evdev is an alternative here.
I'd be surprised if QtE doesn't support evdev, but I've been surprised a lot recently. Can you investigate quickly to see if QtE supports evdev?
As far as X and oe-core is concerned, tslib isn't involve at all anymore. There are xorg-input-tslib drivers in Debian but they've not been touched for years, the upstream source URL doesn't exist any more, and I'd actually be somewhat surprised if they worked. Xorg is all about evdev + xinput for touchscreens.
Ross
Quoted 11 lines · Ross
Quoted 5 lines · Otavio Salvador
Quoted 3 lines
Qt Embedded as we build it is currently configured to use tslib, as is SDL. If the alternative (evdev?) is supported they could presumably be switched over though. I don't know how practical that is however.
Yes Qt Embedded uses it; not sure evdev is an alternative here.
I'd be surprised if QtE doesn't support evdev, but I've been surprised a lot recently. Can you investigate quickly to see if QtE supports evdev?
Quoted 5 lines
As far as X and oe-core is concerned, tslib isn't involve at all anymore. There are xorg-input-tslib drivers in Debian but they've not been touched for years, the upstream source URL doesn't exist any more, and I'd actually be somewhat surprised if they worked. Xorg is all about evdev + xinput for touchscreens.
Yes, xf86-input-tslib works
http://git.openembedded.org/meta-openembedded/commit/?id=506d5f781f5ee93e3c9b8ca51a7fb65f40d4a8b8
Signature · 1 line
Martin 'JaMa' Jansa jabber: [email protected]
Quoted 2 lines · Martin Jansa
Yes, xf86-input-tslib works
http://git.openembedded.org/meta-openembedded/commit/?id=506d5f781f5ee93e3c9b8ca51a7fb65f40d4a8b8
[ surprised face ]
You totally should push those tarballs to a git repo on fd.o, pengutronix has clearly abandoned it.
In a harsh mood (its 2130 and I'm working) I'm leaning towards making tslib a packageconfig for qte, and moving tslib into meta-oe...
Ross
Quoted 11 lines · Ross
Quoted 2 lines · Martin Jansa
Yes, xf86-input-tslib works
http://git.openembedded.org/meta-openembedded/commit/?id=506d5f781f5ee93e3c9b8ca51a7fb65f40d4a8b8[ surprised face ]
You totally should push those tarballs to a git repo on fd.o, pengutronix has clearly abandoned it.
In a harsh mood (its 2130 and I'm working) I'm leaning towards making tslib a packageconfig for qte, and moving tslib into meta-oe...
Not if qte does not support evdev; this needs to be checked first or it'll be a regression.
Signature · 3 lines
Otavio Salvador O.S. Systems E-mail: [email protected] http://www.ossystems.com.br Mobile: +55 53 9981-7854 http://projetos.ossystems.com.br
Hi Ross,
Quoted 4 lines · Ross
Quoted 2 lines · Martin Jansa
Yes, xf86-input-tslib works
http://git.openembedded.org/meta-openembedded/commit/?id=506d5f781f5ee93e3c9b8ca51a7fb65f40d4a8b8
Quoted 8 lines
[ surprised face ]
You totally should push those tarballs to a git repo on fd.o, pengutronix has clearly abandoned it.
In a harsh mood (its 2130 and I'm working) I'm leaning towards making tslib a packageconfig for qte, and moving tslib into meta-oe...
if you have a how-to to get Qt/E 4.x working with evdev, please send it before removing tslib so that we can test it on several boards which are actually running with Qt/E+tslib.
Thanks
Eric
Quoted 3 lines · Eric Bénard
if you have a how-to to get Qt/E 4.x working with evdev, please send it before removing tslib so that we can test it on several boards which are actually running with Qt/E+tslib.
Totally, not regressing badly is a prerequisite.
Ross