Re: [OE-core] [PATCH][RFC] base.bbclass: Introduce EXTRA_CONF_PACKAGECONFIG variable
On Mon, Apr 18, 2016 at 07:49:51PM -0300, Otavio Salvador wrote: > On Mon, Apr 18, 2016 at 7:45 PM, Richard Purdie >wrote: > > On Mon, 2016-04-18 at 19:31 -0300, Otavio Salvador wrote: > >> On Fri, Mar 25, 2016 at 8:55 AM, Martin Jansa >> > wrote: > >> > Was there some agreement about variable name? > >> > >> meta-qt5 5.6 is using this so we need a final decision if it will be > >> merged or not. I am in favor it as I think it is clear enough. > > > > It was an RFC and I still don't get a good feeling about the names > > used. > > > > I suggested: > > > > PACKAGECONFIG_CONFPARAMS > > or > > PACKAGECONFIG_CONFARGS > > I like CONFARGS. I'll send patch with PACKAGECONFIG_CONFARGS and wait until it's merged somewhere, before updating usages of this in other layers (so that I don't need to redo it again if the name is changed again). > > as it then makes things slightly clearer these things are coming from > > PACKAGECONFIG. We have a namespace which is a bit of a mess and I'd > > prefer to try and improve with new things if we can... > > > > The fact meta-qt5 is already using this is bad, it really shouldn't be. > > We're also very close to the final build of 2.1 right now so this would > > be a very late change :/. meta-qt5/master isn't using it, it's only in master-next, because without this change the builds are broken when people use openembedded-core/meta/conf/distro/include/no-static-libs.inc e.g. with default Poky config. > I think it is a safe change but I am fine in postponing it for 2.2. In > this case we need to revert the usage of it in meta-qt5 (which I also > support). Nothing needs to be reverted, because it wasn't merged in any permanent layer. If it doesn't go to 2.1 then I'll leave meta-qt5/krogoth version broken for Poky builds with no-static-libs.inc. -- Martin 'JaMa' Jansa jabber: martin.ja...@gmail.com signature.asc Description: Digital signature -- ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] base.bbclass: Introduce EXTRA_CONF_PACKAGECONFIG variable
On Mon, Apr 18, 2016 at 7:45 PM, Richard Purdiewrote: > On Mon, 2016-04-18 at 19:31 -0300, Otavio Salvador wrote: >> On Fri, Mar 25, 2016 at 8:55 AM, Martin Jansa > > wrote: >> > Was there some agreement about variable name? >> >> meta-qt5 5.6 is using this so we need a final decision if it will be >> merged or not. I am in favor it as I think it is clear enough. > > It was an RFC and I still don't get a good feeling about the names > used. > > I suggested: > > PACKAGECONFIG_CONFPARAMS > or > PACKAGECONFIG_CONFARGS I like CONFARGS. > as it then makes things slightly clearer these things are coming from > PACKAGECONFIG. We have a namespace which is a bit of a mess and I'd > prefer to try and improve with new things if we can... > > The fact meta-qt5 is already using this is bad, it really shouldn't be. > We're also very close to the final build of 2.1 right now so this would > be a very late change :/. I think it is a safe change but I am fine in postponing it for 2.2. In this case we need to revert the usage of it in meta-qt5 (which I also support). Martin? -- Otavio Salvador O.S. Systems http://www.ossystems.com.brhttp://code.ossystems.com.br Mobile: +55 (53) 9981-7854Mobile: +1 (347) 903-9750 -- ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] base.bbclass: Introduce EXTRA_CONF_PACKAGECONFIG variable
On Mon, 2016-04-18 at 19:31 -0300, Otavio Salvador wrote: > On Fri, Mar 25, 2016 at 8:55 AM, Martin Jansa> wrote: > > On Fri, Mar 25, 2016 at 11:28:59AM +, Richard Purdie wrote: > > > On Fri, 2016-03-25 at 11:50 +0100, Martin Jansa wrote: > > > > On Wed, Mar 02, 2016 at 12:40:31PM +, Burton, Ross wrote: > > > > > On 27 February 2016 at 22:09, Martin Jansa < > > > > > martin.ja...@gmail.com> > > > > > wrote: > > > > > > > > > > > * add separate variable for configuration options generated > > > > > > from > > > > > > PACKAGECONFIG setting, this helps other bbclasses and > > > > > > recipes > > > > > > to take advantage of PACKAGECONFIG mechanism, without > > > > > > including > > > > > > other options from EXTRA_OECONF > > > > > > * e.g. meta-qt5 recipes are abusing EXTRA_OECONF to get > > > > > > options > > > > > > from PACKAGECONFIG: > > > > > > EXTRA_QMAKEVARS_PRE += > > > > > > but with > > > > > > conf/distro/include/no-static-libs.inc > > > > > > it means getting --disable-static as invalid option > > > > > > inside > > > > > > EXTRA_QMAKEVARS_PRE as reported by Alexandre Belloni who > > > > > > tried > > > > > > to use poky with meta-qt5. > > > > > > * once we migrate all bbclasses and recipes to > > > > > > EXTRA_CONF_PACKAGECONFIG > > > > > > we should also restrict EXTRA_OECONF append only to > > > > > > autotools.bbclass > > > > > > like I did for cmake.bbclass > > > > > > > > > > > > > > > > Sorry, some of us have been a bit busy trying to get M3 > > > > > stable. > > > > > This does > > > > > look good and I'm for squeezing it into M3. > > > > > > > > Any update on squeezing this? > > > > > > > > meta-qt5 is still broken with default poky config > > > > > > I'm a little confused, was there going to be another version with > > > some > > > tweaked variable names? > > > > Was there some agreement about variable name? > > meta-qt5 5.6 is using this so we need a final decision if it will be > merged or not. I am in favor it as I think it is clear enough. It was an RFC and I still don't get a good feeling about the names used. I suggested: PACKAGECONFIG_CONFPARAMS or PACKAGECONFIG_CONFARGS as it then makes things slightly clearer these things are coming from PACKAGECONFIG. We have a namespace which is a bit of a mess and I'd prefer to try and improve with new things if we can... The fact meta-qt5 is already using this is bad, it really shouldn't be. We're also very close to the final build of 2.1 right now so this would be a very late change :/. Cheers, Richard -- ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] base.bbclass: Introduce EXTRA_CONF_PACKAGECONFIG variable
On Fri, Mar 25, 2016 at 8:55 AM, Martin Jansawrote: > On Fri, Mar 25, 2016 at 11:28:59AM +, Richard Purdie wrote: >> On Fri, 2016-03-25 at 11:50 +0100, Martin Jansa wrote: >> > On Wed, Mar 02, 2016 at 12:40:31PM +, Burton, Ross wrote: >> > > On 27 February 2016 at 22:09, Martin Jansa >> > > wrote: >> > > >> > > > * add separate variable for configuration options generated from >> > > > PACKAGECONFIG setting, this helps other bbclasses and recipes >> > > > to take advantage of PACKAGECONFIG mechanism, without including >> > > > other options from EXTRA_OECONF >> > > > * e.g. meta-qt5 recipes are abusing EXTRA_OECONF to get options >> > > > from PACKAGECONFIG: >> > > > EXTRA_QMAKEVARS_PRE += >> > > > but with >> > > > conf/distro/include/no-static-libs.inc >> > > > it means getting --disable-static as invalid option inside >> > > > EXTRA_QMAKEVARS_PRE as reported by Alexandre Belloni who tried >> > > > to use poky with meta-qt5. >> > > > * once we migrate all bbclasses and recipes to >> > > > EXTRA_CONF_PACKAGECONFIG >> > > > we should also restrict EXTRA_OECONF append only to >> > > > autotools.bbclass >> > > > like I did for cmake.bbclass >> > > > >> > > >> > > Sorry, some of us have been a bit busy trying to get M3 stable. >> > > This does >> > > look good and I'm for squeezing it into M3. >> > >> > Any update on squeezing this? >> > >> > meta-qt5 is still broken with default poky config >> >> I'm a little confused, was there going to be another version with some >> tweaked variable names? > > Was there some agreement about variable name? meta-qt5 5.6 is using this so we need a final decision if it will be merged or not. I am in favor it as I think it is clear enough. -- Otavio Salvador O.S. Systems http://www.ossystems.com.brhttp://code.ossystems.com.br Mobile: +55 (53) 9981-7854Mobile: +1 (347) 903-9750 -- ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] base.bbclass: Introduce EXTRA_CONF_PACKAGECONFIG variable
On Fri, Mar 25, 2016 at 11:28:59AM +, Richard Purdie wrote: > On Fri, 2016-03-25 at 11:50 +0100, Martin Jansa wrote: > > On Wed, Mar 02, 2016 at 12:40:31PM +, Burton, Ross wrote: > > > On 27 February 2016 at 22:09, Martin Jansa> > > wrote: > > > > > > > * add separate variable for configuration options generated from > > > > PACKAGECONFIG setting, this helps other bbclasses and recipes > > > > to take advantage of PACKAGECONFIG mechanism, without including > > > > other options from EXTRA_OECONF > > > > * e.g. meta-qt5 recipes are abusing EXTRA_OECONF to get options > > > > from PACKAGECONFIG: > > > > EXTRA_QMAKEVARS_PRE += > > > > but with > > > > conf/distro/include/no-static-libs.inc > > > > it means getting --disable-static as invalid option inside > > > > EXTRA_QMAKEVARS_PRE as reported by Alexandre Belloni who tried > > > > to use poky with meta-qt5. > > > > * once we migrate all bbclasses and recipes to > > > > EXTRA_CONF_PACKAGECONFIG > > > > we should also restrict EXTRA_OECONF append only to > > > > autotools.bbclass > > > > like I did for cmake.bbclass > > > > > > > > > > Sorry, some of us have been a bit busy trying to get M3 stable. > > > This does > > > look good and I'm for squeezing it into M3. > > > > Any update on squeezing this? > > > > meta-qt5 is still broken with default poky config > > I'm a little confused, was there going to be another version with some > tweaked variable names? Was there some agreement about variable name? :) -- Martin 'JaMa' Jansa jabber: martin.ja...@gmail.com signature.asc Description: Digital signature -- ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] base.bbclass: Introduce EXTRA_CONF_PACKAGECONFIG variable
On Fri, 2016-03-25 at 11:50 +0100, Martin Jansa wrote: > On Wed, Mar 02, 2016 at 12:40:31PM +, Burton, Ross wrote: > > On 27 February 2016 at 22:09, Martin Jansa> > wrote: > > > > > * add separate variable for configuration options generated from > > > PACKAGECONFIG setting, this helps other bbclasses and recipes > > > to take advantage of PACKAGECONFIG mechanism, without including > > > other options from EXTRA_OECONF > > > * e.g. meta-qt5 recipes are abusing EXTRA_OECONF to get options > > > from PACKAGECONFIG: > > > EXTRA_QMAKEVARS_PRE += > > > but with > > > conf/distro/include/no-static-libs.inc > > > it means getting --disable-static as invalid option inside > > > EXTRA_QMAKEVARS_PRE as reported by Alexandre Belloni who tried > > > to use poky with meta-qt5. > > > * once we migrate all bbclasses and recipes to > > > EXTRA_CONF_PACKAGECONFIG > > > we should also restrict EXTRA_OECONF append only to > > > autotools.bbclass > > > like I did for cmake.bbclass > > > > > > > Sorry, some of us have been a bit busy trying to get M3 stable. > > This does > > look good and I'm for squeezing it into M3. > > Any update on squeezing this? > > meta-qt5 is still broken with default poky config I'm a little confused, was there going to be another version with some tweaked variable names? Cheers, Richard -- ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] base.bbclass: Introduce EXTRA_CONF_PACKAGECONFIG variable
On Wed, Mar 02, 2016 at 12:40:31PM +, Burton, Ross wrote: > On 27 February 2016 at 22:09, Martin Jansawrote: > > > * add separate variable for configuration options generated from > > PACKAGECONFIG setting, this helps other bbclasses and recipes > > to take advantage of PACKAGECONFIG mechanism, without including > > other options from EXTRA_OECONF > > * e.g. meta-qt5 recipes are abusing EXTRA_OECONF to get options > > from PACKAGECONFIG: > > EXTRA_QMAKEVARS_PRE += > > but with > > conf/distro/include/no-static-libs.inc > > it means getting --disable-static as invalid option inside > > EXTRA_QMAKEVARS_PRE as reported by Alexandre Belloni who tried > > to use poky with meta-qt5. > > * once we migrate all bbclasses and recipes to EXTRA_CONF_PACKAGECONFIG > > we should also restrict EXTRA_OECONF append only to autotools.bbclass > > like I did for cmake.bbclass > > > > Sorry, some of us have been a bit busy trying to get M3 stable. This does > look good and I'm for squeezing it into M3. Any update on squeezing this? meta-qt5 is still broken with default poky config -- Martin 'JaMa' Jansa jabber: martin.ja...@gmail.com signature.asc Description: Digital signature -- ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] base.bbclass: Introduce EXTRA_CONF_PACKAGECONFIG variable
On Wed, Mar 02, 2016 at 01:11:45PM +, Richard Purdie wrote: > On Wed, 2016-03-02 at 13:38 +0100, Martin Jansa wrote: > > On Sat, Feb 27, 2016 at 11:09:47PM +0100, Martin Jansa wrote: > > > * add separate variable for configuration options generated from > > > PACKAGECONFIG setting, this helps other bbclasses and recipes > > > to take advantage of PACKAGECONFIG mechanism, without including > > > other options from EXTRA_OECONF > > > * e.g. meta-qt5 recipes are abusing EXTRA_OECONF to get options > > > from PACKAGECONFIG: > > > EXTRA_QMAKEVARS_PRE += > > > but with > > > conf/distro/include/no-static-libs.inc > > > it means getting --disable-static as invalid option inside > > > EXTRA_QMAKEVARS_PRE as reported by Alexandre Belloni who tried > > > to use poky with meta-qt5. > > > * once we migrate all bbclasses and recipes to > > > EXTRA_CONF_PACKAGECONFIG > > > we should also restrict EXTRA_OECONF append only to > > > autotools.bbclass > > > like I did for cmake.bbclass > > > > No comments? Should I resend without [RFC] tag? > > > > This is needed to fix couple components when > > conf/distro/include/no-static-libs.inc is used. > > I can see the need for it, I'm just not 100% sure I like the form of > the patch. No one particular thing is doing that, just a general > feeling of unease which I can't quite put into words :(. > > We continue to have a need to differentiate between "proper" autotools > recipes and non-autotools recipes which would make this kind of issue > easier. I guess I'm trying to weigh up whether we should consider > something a bit more invasive to try and improve things and if we do > that whether this patch helps or hinders that (it probably does help). I've considered the invasive part of moving EXTRA_OECONF append to autotools.bbclass (like I did for cmake.bbclass) but after grepping for EXTRA_OECONF I've decided to leave it for separate step (e.g. waf-samba.bbclass and meta-oe/recipes-benchmark/fio/fio_2.2.6.bb are abusing EXTRA_OECONF and would break if we remove this). EXTRA_OECMAKE wasn't afaik abused anywhere and fix for qt5 was relatively simple: http://patchwork.openembedded.org/patch/116981/ so I went with compromise to fix what's really failing now and leave future cleanup/improvement for later when more recipes adapt EXTRA_CONF_PACKAGECONFIG variable. > I'm also not 100% convinced EXTRA_CONF_PACKAGECONFIG is the right name, > but I can see how you got here and I'm not sure I have a better > suggestion (PACKAGECONFIG_CONFPARAMS? _CONFARGS?) I was expecting this discussion, I have no strong opinion either way. Namespacing with with PACKAGECONFIG_ prefix is good idea though. -- Martin 'JaMa' Jansa jabber: martin.ja...@gmail.com signature.asc Description: Digital signature -- ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] base.bbclass: Introduce EXTRA_CONF_PACKAGECONFIG variable
On Wed, 2016-03-02 at 13:38 +0100, Martin Jansa wrote: > On Sat, Feb 27, 2016 at 11:09:47PM +0100, Martin Jansa wrote: > > * add separate variable for configuration options generated from > > PACKAGECONFIG setting, this helps other bbclasses and recipes > > to take advantage of PACKAGECONFIG mechanism, without including > > other options from EXTRA_OECONF > > * e.g. meta-qt5 recipes are abusing EXTRA_OECONF to get options > > from PACKAGECONFIG: > > EXTRA_QMAKEVARS_PRE += > > but with > > conf/distro/include/no-static-libs.inc > > it means getting --disable-static as invalid option inside > > EXTRA_QMAKEVARS_PRE as reported by Alexandre Belloni who tried > > to use poky with meta-qt5. > > * once we migrate all bbclasses and recipes to > > EXTRA_CONF_PACKAGECONFIG > > we should also restrict EXTRA_OECONF append only to > > autotools.bbclass > > like I did for cmake.bbclass > > No comments? Should I resend without [RFC] tag? > > This is needed to fix couple components when > conf/distro/include/no-static-libs.inc is used. I can see the need for it, I'm just not 100% sure I like the form of the patch. No one particular thing is doing that, just a general feeling of unease which I can't quite put into words :(. We continue to have a need to differentiate between "proper" autotools recipes and non-autotools recipes which would make this kind of issue easier. I guess I'm trying to weigh up whether we should consider something a bit more invasive to try and improve things and if we do that whether this patch helps or hinders that (it probably does help). I'm also not 100% convinced EXTRA_CONF_PACKAGECONFIG is the right name, but I can see how you got here and I'm not sure I have a better suggestion (PACKAGECONFIG_CONFPARAMS? _CONFARGS?) Cheers, Richard -- ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] base.bbclass: Introduce EXTRA_CONF_PACKAGECONFIG variable
On 27 February 2016 at 22:09, Martin Jansawrote: > * add separate variable for configuration options generated from > PACKAGECONFIG setting, this helps other bbclasses and recipes > to take advantage of PACKAGECONFIG mechanism, without including > other options from EXTRA_OECONF > * e.g. meta-qt5 recipes are abusing EXTRA_OECONF to get options > from PACKAGECONFIG: > EXTRA_QMAKEVARS_PRE += > but with > conf/distro/include/no-static-libs.inc > it means getting --disable-static as invalid option inside > EXTRA_QMAKEVARS_PRE as reported by Alexandre Belloni who tried > to use poky with meta-qt5. > * once we migrate all bbclasses and recipes to EXTRA_CONF_PACKAGECONFIG > we should also restrict EXTRA_OECONF append only to autotools.bbclass > like I did for cmake.bbclass > Sorry, some of us have been a bit busy trying to get M3 stable. This does look good and I'm for squeezing it into M3. Ross -- ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
Re: [OE-core] [PATCH][RFC] base.bbclass: Introduce EXTRA_CONF_PACKAGECONFIG variable
On Sat, Feb 27, 2016 at 11:09:47PM +0100, Martin Jansa wrote: > * add separate variable for configuration options generated from > PACKAGECONFIG setting, this helps other bbclasses and recipes > to take advantage of PACKAGECONFIG mechanism, without including > other options from EXTRA_OECONF > * e.g. meta-qt5 recipes are abusing EXTRA_OECONF to get options > from PACKAGECONFIG: > EXTRA_QMAKEVARS_PRE += > but with > conf/distro/include/no-static-libs.inc > it means getting --disable-static as invalid option inside > EXTRA_QMAKEVARS_PRE as reported by Alexandre Belloni who tried > to use poky with meta-qt5. > * once we migrate all bbclasses and recipes to EXTRA_CONF_PACKAGECONFIG > we should also restrict EXTRA_OECONF append only to autotools.bbclass > like I did for cmake.bbclass No comments? Should I resend without [RFC] tag? This is needed to fix couple components when conf/distro/include/no-static-libs.inc is used. Cheers, signature.asc Description: Digital signature -- ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
[OE-core] [PATCH][RFC] base.bbclass: Introduce EXTRA_CONF_PACKAGECONFIG variable
* add separate variable for configuration options generated from PACKAGECONFIG setting, this helps other bbclasses and recipes to take advantage of PACKAGECONFIG mechanism, without including other options from EXTRA_OECONF * e.g. meta-qt5 recipes are abusing EXTRA_OECONF to get options from PACKAGECONFIG: EXTRA_QMAKEVARS_PRE += but with conf/distro/include/no-static-libs.inc it means getting --disable-static as invalid option inside EXTRA_QMAKEVARS_PRE as reported by Alexandre Belloni who tried to use poky with meta-qt5. * once we migrate all bbclasses and recipes to EXTRA_CONF_PACKAGECONFIG we should also restrict EXTRA_OECONF append only to autotools.bbclass like I did for cmake.bbclass Signed-off-by: Martin Jansa--- meta/classes/base.bbclass | 9 ++--- meta/classes/cmake.bbclass | 2 ++ 2 files changed, 8 insertions(+), 3 deletions(-) diff --git a/meta/classes/base.bbclass b/meta/classes/base.bbclass index 1372f38..6e94995 100644 --- a/meta/classes/base.bbclass +++ b/meta/classes/base.bbclass @@ -428,9 +428,12 @@ python () { extraconf.append(items[1]) appendVar('DEPENDS', extradeps) appendVar('RDEPENDS_${PN}', extrardeps) -if bb.data.inherits_class('cmake', d): -appendVar('EXTRA_OECMAKE', extraconf) -else: +appendVar('EXTRA_CONF_PACKAGECONFIG', extraconf) + +# TODO once are all recipes/bbclass abusign EXTRA_OECONF +# to get PACKAGECONFIG options fixed to use EXTRA_CONF_PACKAGECONFIG +# move this append to autotools.bbclass +if not bb.data.inherits_class('cmake', d): appendVar('EXTRA_OECONF', extraconf) pn = d.getVar('PN', True) diff --git a/meta/classes/cmake.bbclass b/meta/classes/cmake.bbclass index 22cd61e..122718e 100644 --- a/meta/classes/cmake.bbclass +++ b/meta/classes/cmake.bbclass @@ -27,6 +27,8 @@ OECMAKE_EXTRA_ROOT_PATH ?= "" OECMAKE_FIND_ROOT_PATH_MODE_PROGRAM = "ONLY" OECMAKE_FIND_ROOT_PATH_MODE_PROGRAM_class-native = "BOTH" +EXTRA_OECMAKE_append = " ${EXTRA_CONF_PACKAGECONFIG}" + # CMake expects target architectures in the format of uname(2), # which do not always match TARGET_ARCH, so all the necessary # conversions should happen here. -- 2.7.1 -- ___ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core