uninative.bbclass decides whether to enable itself by comparing
`ldd --version` against UNINATIVE_MAXGLIBCVERSION. buildtools-extended
ships its own gcc and glibc, and native recipes built with it link
against that glibc, but it ships no ldd, so the check reads the host's
ldd instead. On a host whose glibc is newer than uninative's, uninative
is disabled even though the build never uses the host glibc, and native
sstate is then keyed to the host distro instead of "universal".

Ship ldd so the check sees the glibc the build actually uses.

Reviewed-by: Tibault Damman <[email protected]>
Signed-off-by: Lander Noterman <[email protected]>
---
 meta/recipes-core/meta/buildtools-extended-tarball.bb | 1 +
 1 file changed, 1 insertion(+)

diff --git a/meta/recipes-core/meta/buildtools-extended-tarball.bb 
b/meta/recipes-core/meta/buildtools-extended-tarball.bb
index bd4c19c7ea..c9b2881d50 100644
--- a/meta/recipes-core/meta/buildtools-extended-tarball.bb
+++ b/meta/recipes-core/meta/buildtools-extended-tarball.bb
@@ -28,6 +28,7 @@ TOOLCHAIN_HOST_TASK += "\
     nativesdk-libtool \
     nativesdk-pkgconfig \
     nativesdk-glibc-utils \
+    nativesdk-ldd \
     nativesdk-glibc-gconvs \
     nativesdk-libxcrypt-dev \
     "
-- 
2.55.0

-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#246875): 
https://lists.openembedded.org/g/openembedded-core/message/246875
Mute This Topic: https://lists.openembedded.org/mt/121491795/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub 
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to