nativesdk-coreutils: a lot of warnings fixed
cover nativesdk-coreutils: a lot of warnings fixed
The following changes since commit 0f62b3e5344f63deca43091fa7909448d4f87279:
quilt: fix ptest faildiff.test failed (2016-03-14 21:24:00 -0400)
are available in the git repository at:
git://git.openembedded.org/openembedded-core-contrib dengke/nativesdk-coreutils_many_warnings_fixed
http://cgit.openembedded.org/cgit.cgi/openembedded-core-contrib/log/?h=dengke/nativesdk-coreutils_many_warnings_fixed
Dengke Du (1):
nativesdk-coreutils: a lot of warnings fixed
meta/recipes-core/coreutils/coreutils_8.25.bb | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
Signature · 1 line
1.9.1
PATCH 1/1 nativesdk-coreutils: a lot of warnings fixed
When we create nativesdk-coreutils, a lot of warnings appear,it show many files can't find. Because in the coreutils recipe, it didn't contain the do_install for the nativesdk, so when the alternative system check the files in the following process, it can't find the files. So we should add the do_install for the nativesdk, change the function do_install_append_class-target() to do_install_append() in the file:
meta/recipes-core/coreutils/coreutils_8.25.bb
in this way, the alternative system can find the files, the warnings disappear.
--- meta/recipes-core/coreutils/coreutils_8.25.bb | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
meta/recipes-core/coreutils/coreutils_8.25.bb +2 -1
@@ -27,6 +27,7 @@ SRC_URI[manpages.sha256sum] = "2ee31c3a6d2276f49c5515375d4a0c1047580da6ac1053689 EXTRA_OECONF_class-native = "--without-gmp" EXTRA_OECONF_class-target = "--enable-install-program=arch --libexecdir=${libdir}" +EXTRA_OECONF_class-nativesdk = "--enable-install-program=arch" # acl and xattr are not default features #
@@ -67,7 +68,7 @@ do_compile_prepend () { mkdir -p ${B}/src } -do_install_append_class-target() { +do_install_append() { for i in df mktemp base64; do mv ${D}${bindir}/$i ${D}${bindir}/$i.${BPN}; done install -d ${D}${base_bindir}
Signature · 1 line
1.9.1
Quoted 3 lines · Dengke Du
When we create nativesdk-coreutils, a lot of warnings appear,it show many files can't find.
I don't think anywhere where nativesdk packages are used actually runs the postinst scripts to create the alternative links, so maybe we should be stopping update-alternatives do anything in nativesdk (just like it doesn't do anything for native)?
Ross
Quoted 8 lines · Ross
Quoted 2 lines · Dengke Du
When we create nativesdk-coreutils, a lot of warnings appear,it show many files can't find.
I don't think anywhere where nativesdk packages are used actually runs the postinst scripts to create the alternative links, so maybe we should be stopping update-alternatives do anything in nativesdk (just like it doesn't do anything for native)?
Actually I checked this by adding a dummy line which creates a file in $D to the dbus postinst and then adding nativesdk-dbus to buildtools-tarball - it seems that the postinst does actually get run. As long as the postinst can successfully run at SDK construction time then apparently it should work.
Cheers,
Paul
Signature · 2 lines
Paul Eggleton Intel Open Source Technology Centre
After a talk with Hongxu jia, we decided to change the do_install_append_class-target() to do_install_append(), in this function, added the following:
if [ "${CLASSOVERRIDE}" = "class-native" ]; then
return
fi
in this way, avoid the conflict with the following do_install_append_class-native function.
//dengke
Quoted 16 lines · Paul Eggleton
Quoted 7 lines · Ross
Quoted 2 lines · Dengke Du
When we create nativesdk-coreutils, a lot of warnings appear,it show many files can't find.
I don't think anywhere where nativesdk packages are used actually runs the postinst scripts to create the alternative links, so maybe we should be stopping update-alternatives do anything in nativesdk (just like it doesn't do anything for native)?
Actually I checked this by adding a dummy line which creates a file in $D to the dbus postinst and then adding nativesdk-dbus to buildtools-tarball - it seems that the postinst does actually get run. As long as the postinst can successfully run at SDK construction time then apparently it should work.
Cheers,
Paul