On Tue, 2026-06-09 at 10:40 +0200, Jörg Sommer wrote: > Richard Purdie schrieb am Di 09. Jun, 08:41 (+0100): > > On Mon, 2026-06-08 at 18:54 +0200, Jörg Sommer via lists.openembedded.org > > wrote: > > > From: Jörg Sommer <[email protected]> > > > > > > Fixes: bee29dc822c611 ("rpcbind: Remove IPv6 listener, if ipv6 not in > > > DISTRO_FEATURES") > > > Signed-off-by: Jörg Sommer <[email protected]> > > > --- > > > meta/recipes-extended/rpcbind/rpcbind_1.2.8.bb | 3 ++- > > > 1 file changed, 2 insertions(+), 1 deletion(-) > > > > > > diff --git a/meta/recipes-extended/rpcbind/rpcbind_1.2.8.bb > > > b/meta/recipes-extended/rpcbind/rpcbind_1.2.8.bb > > > index bd8ba1e88a..6604774c1f 100644 > > > --- a/meta/recipes-extended/rpcbind/rpcbind_1.2.8.bb > > > +++ b/meta/recipes-extended/rpcbind/rpcbind_1.2.8.bb > > > @@ -55,7 +55,8 @@ do_install:append () { > > > install -d ${D}${systemd_system_unitdir}/rpcbind.service.d > > > install -m 0644 ${UNPACKDIR}/rpcbind.systemd > > > ${D}${systemd_system_unitdir}/rpcbind.service.d/rpcbind.conf > > > > > > - if ! ${@bb.utils.contains('DISTRO_FEATURES', 'ipv6', 'true', 'false', > > > d)} > > > + if ${@bb.utils.contains('DISTRO_FEATURES', 'systemd', 'true', 'false', > > > d)} \ > > > + && ! ${@bb.utils.contains('DISTRO_FEATURES', 'ipv6', 'true', > > > 'false', d)} > > > then > > > sed -i '/^Listen.*=\[.*\]/s/^/# /' > > > ${D}${systemd_system_unitdir}/*.socket > > > fi > > > > That can be simplified to: > > > > bb.utils.contains('DISTRO_FEATURES', 'systemd ipv6', 'true', 'false', d) > > Actually, it is 'systemd ! ipv6'. I do not think this syntax is supported. > > This pattern with true/false is used pretty often: > > % for i in oe-core meta-openembedded; git -C $i grep "'true', 'false'" |wc -l > 83 > 131 > > Is it worth to create a helper? Something like contains_tf?
Sorry, I missed the ! hidden in there! I'm torn on the function. "contains" does have special optimisation within bitbake and even if you wrap it, we'd lose that unless we add more specialist code to bitbake to handle the new function. I started wondering if a "to_shell()" type converter might be better if we could then have truevalue default to True. The function could convert True to "true" for shell usage. Having d as a positional parameter in the arguments for contains makes that hard to transition to though :( Cheers, Richard
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#238278): https://lists.openembedded.org/g/openembedded-core/message/238278 Mute This Topic: https://lists.openembedded.org/mt/119709226/21656 Group Owner: [email protected] Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
