On 7/1/20 10:29 PM, Richard Purdie wrote:
On Wed, 2020-07-01 at 21:43 +0800, hongxu wrote:
Since commit [24b80d211f nativesdk.bbclass: set sbindir to bindir]
applied, sbindir = bindir in nativesdk, then routine `rmt' is
installed to bindir which break `rmdir ${D}${bindir}/' failed with
`Directory not empty'
Don't rmdir while nativesdk
Signed-off-by: Hongxu Jia <[email protected]>
---
meta/recipes-extended/cpio/cpio_2.13.bb | 6 ++++--
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/meta/recipes-extended/cpio/cpio_2.13.bb b/meta/recipes-
extended/cpio/cpio_2.13.bb
index 9e35a80f8b..44036f07f9 100644
--- a/meta/recipes-extended/cpio/cpio_2.13.bb
+++ b/meta/recipes-extended/cpio/cpio_2.13.bb
@@ -23,7 +23,9 @@ do_install () {
if [ "${base_bindir}" != "${bindir}" ]; then
install -d ${D}${base_bindir}/
mv "${D}${bindir}/cpio" "${D}${base_bindir}/cpio"
- rmdir ${D}${bindir}/
+ if [ "${PN}" != "nativesdk-cpio" ]; then
+ rmdir ${D}${bindir}/
+ fi
fi
Please make the check "bindir != sbindir".
Got it, V2 incoming
# Avoid conflicts with the version from tar
@@ -46,4 +48,4 @@ ALTERNATIVE_LINK_NAME[cpio] = "${base_bindir}/cpio"
ALTERNATIVE_PRIORITY[rmt] = "50"
ALTERNATIVE_LINK_NAME[rmt] = "${sbindir}/rmt"
-BBCLASSEXTEND = "native"
+BBCLASSEXTEND = "native nativesdk"
So something which was never enabled broke? I guess this was hiding in
some other layer?
With a simple search, there is a low version cpio 2.8 in meta-gplv2
without nativesdk support
//Hongxu
Cheers,
Richard
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#140188):
https://lists.openembedded.org/g/openembedded-core/message/140188
Mute This Topic: https://lists.openembedded.org/mt/75235440/21656
Group Owner: [email protected]
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub
[[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-
