| Message ID | 1787566358.8631fc262581453bbf619ec5b2062170.1a033420133000c4f3@vates.tech |
|---|---|
| State | New |
| Headers | show |
| Series | doc/bitbake-user-manual-ref-variables: document PREFERRED_RPROVIDER | expand |
Hi Yann, On 8/24/26 12:12 PM, Yann Dirson via lists.openembedded.org wrote: > [You don't often get email from yann.dirson=vates.tech@lists.openembedded.org. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ] > > Signed-off-by: Yann Dirson <yann.dirson@vates.tech> > --- > .../bitbake-user-manual-ref-variables.rst | 9 +++++++++ > 1 file changed, 9 insertions(+) > > diff --git a/doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst b/doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst > index 5395ce5e3..2756fc0c4 100644 > --- a/doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst > +++ b/doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst > @@ -1558,6 +1558,15 @@ overview of their function and contents. > PREFERRED_PROVIDER_xxx = "yyy" > PREFERRED_PROVIDER_aaa = "bbb" > > + :term:`PREFERRED_RPROVIDER` Determines which recipe should be given Please be consistent with the rest of the file. :term:`PREFERRED_RPROVIDER` should be on its own on the first line. > + preference when multiple packages rprovide the same item. Beware rprovide is not an English term so all readers won't necessarily know what this means. I can suggest: """ Determines which *recipe* generating a package runtime-providing (:term:`RPROVIDES`) the same item as other packages should be given preference. Some examples:: [...] """ > + this should contain the name of the *recipe* producing the > + package that rprovides the item, not that of the package. Some > + examples:: > + > + PREFERRED_RPROVIDER_initd-functions ?= "initscripts" > + PREFERRED_RPROVIDER_virtual-libegl-icd ?= "mesa" Are you sure this is still valid? I don't see an RPROVIDES for virtual-libegl-icd so maybe this is not doing anything? Cheers, Quentin
On 8/24/26 12:26 PM, Quentin Schulz wrote: > Hi Yann, > > On 8/24/26 12:12 PM, Yann Dirson via lists.openembedded.org wrote: >> [You don't often get email from >> yann.dirson=vates.tech@lists.openembedded.org. Learn why this is >> important at https://aka.ms/LearnAboutSenderIdentification ] >> >> Signed-off-by: Yann Dirson <yann.dirson@vates.tech> >> --- >> .../bitbake-user-manual-ref-variables.rst | 9 +++++++++ >> 1 file changed, 9 insertions(+) >> >> diff --git a/doc/bitbake-user-manual/bitbake-user-manual-ref- >> variables.rst b/doc/bitbake-user-manual/bitbake-user-manual-ref- >> variables.rst >> index 5395ce5e3..2756fc0c4 100644 >> --- a/doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst >> +++ b/doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst >> @@ -1558,6 +1558,15 @@ overview of their function and contents. >> PREFERRED_PROVIDER_xxx = "yyy" >> PREFERRED_PROVIDER_aaa = "bbb" >> >> + :term:`PREFERRED_RPROVIDER` Determines which recipe should be given > > Please be consistent with the rest of the > file. :term:`PREFERRED_RPROVIDER` should be on its own on the first line. > >> + preference when multiple packages rprovide the same item. Beware > > rprovide is not an English term so all readers won't necessarily know > what this means. I can suggest: > > """ > Determines which *recipe* generating a package runtime-providing > (:term:`RPROVIDES`) the same item as other packages should be given > preference. Some examples:: > [...] Good points. > > """ > >> + this should contain the name of the *recipe* producing the >> + package that rprovides the item, not that of the package. Some >> + examples:: >> + >> + PREFERRED_RPROVIDER_initd-functions ?= "initscripts" >> + PREFERRED_RPROVIDER_virtual-libegl-icd ?= "mesa" > > Are you sure this is still valid? I don't see an RPROVIDES for virtual- > libegl-icd so maybe this is not doing anything? Those examples are literally taken from oe-core's meta/conf/distro/include/default-providers.inc. Digging a bit, they seem to be dynamically-generated values from an anonymous python block in ./meta/recipes-graphics/mesa/mesa.inc
On 8/24/26 3:40 PM, Yann Dirson wrote: > > > On 8/24/26 12:26 PM, Quentin Schulz wrote: >> Hi Yann, >> >> On 8/24/26 12:12 PM, Yann Dirson via lists.openembedded.org wrote: >>> [You don't often get email from >>> yann.dirson=vates.tech@lists.openembedded.org. Learn why this is >>> important at https://aka.ms/LearnAboutSenderIdentification ] >>> >>> Signed-off-by: Yann Dirson <yann.dirson@vates.tech> >>> --- >>> .../bitbake-user-manual-ref-variables.rst | 9 +++++++++ >>> 1 file changed, 9 insertions(+) >>> >>> diff --git a/doc/bitbake-user-manual/bitbake-user-manual-ref- >>> variables.rst b/doc/bitbake-user-manual/bitbake-user-manual-ref- >>> variables.rst >>> index 5395ce5e3..2756fc0c4 100644 >>> --- a/doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst >>> +++ b/doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst >>> @@ -1558,6 +1558,15 @@ overview of their function and contents. >>> PREFERRED_PROVIDER_xxx = "yyy" >>> PREFERRED_PROVIDER_aaa = "bbb" >>> >>> + :term:`PREFERRED_RPROVIDER` Determines which recipe should be given >> >> Please be consistent with the rest of the >> file. :term:`PREFERRED_RPROVIDER` should be on its own on the first line. >> >>> + preference when multiple packages rprovide the same item. Beware >> >> rprovide is not an English term so all readers won't necessarily know >> what this means. I can suggest: >> >> """ >> Determines which *recipe* generating a package runtime-providing >> (:term:`RPROVIDES`) the same item as other packages should be given >> preference. Some examples:: >> [...] > > Good points. > >> >> """ >> >>> + this should contain the name of the *recipe* producing the >>> + package that rprovides the item, not that of the package. Some >>> + examples:: >>> + >>> + PREFERRED_RPROVIDER_initd-functions ?= "initscripts" >>> + PREFERRED_RPROVIDER_virtual-libegl-icd ?= "mesa" >> >> Are you sure this is still valid? I don't see an RPROVIDES for virtual- >> libegl-icd so maybe this is not doing anything? > > Those examples are literally taken from oe-core's > meta/conf/distro/include/default-providers.inc. Digging a bit, they Doesn't mean that it's still correct. It's not unusual to discover some leftovers of past migration or deletion while writing/updating the docs :) > seem to be dynamically-generated values from an anonymous python block > in ./meta/recipes-graphics/mesa/mesa.inc > Indeed, I had missed that, thanks for the pointer. I think it'd be nice to then write in plain English what at least one of the above line does. The easiest example would be the initscripts one as we've a simple RPROVIDES counterpart in OE-Core. Cheers, Quentin
diff --git a/doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst b/doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst index 5395ce5e3..2756fc0c4 100644 --- a/doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst +++ b/doc/bitbake-user-manual/bitbake-user-manual-ref-variables.rst @@ -1558,6 +1558,15 @@ overview of their function and contents. PREFERRED_PROVIDER_xxx = "yyy" PREFERRED_PROVIDER_aaa = "bbb" + :term:`PREFERRED_RPROVIDER` Determines which recipe should be given + preference when multiple packages rprovide the same item. Beware + this should contain the name of the *recipe* producing the + package that rprovides the item, not that of the package. Some + examples:: + + PREFERRED_RPROVIDER_initd-functions ?= "initscripts" + PREFERRED_RPROVIDER_virtual-libegl-icd ?= "mesa" + :term:`PREFERRED_VERSION` If there are multiple versions of a recipe available, this variable determines which version should be given preference. You must always
Signed-off-by: Yann Dirson <yann.dirson@vates.tech> --- .../bitbake-user-manual-ref-variables.rst | 9 +++++++++ 1 file changed, 9 insertions(+)