On Sat, 2020-05-23 at 02:01 +0900, Masahiro Yamada wrote: > Hi Ben, > > On Sun, May 10, 2020 at 1:06 PM Jason Self <j@xxxxxxxxxx> wrote: > > The official Debian kernel packages, > > I'd like to consult you. > Does this patch look OK? No, it doesn't. > > in addition to running the > > scripts in /etc/kernel, also runs linux-update-symlinks via the > > postinst and postrm scripts. This maintains the /vmlinuz and > > /initrd.img (if it uses an initramfs) symlinks. This change makes the > > builddeb script generate those same calls. > > > > Signed-off-by: Jason Self <j@xxxxxxxxxx> > > --- > > scripts/package/builddeb | 8 ++++++++ > > 1 file changed, 8 insertions(+) > > > > diff --git a/scripts/package/builddeb b/scripts/package/builddeb > > index 6df3c9f8b2da..f34c11c63ede 100755 > > --- a/scripts/package/builddeb > > +++ b/scripts/package/builddeb > > @@ -195,6 +195,14 @@ export DEB_MAINT_PARAMS="\$*" > > # Tell initramfs builder whether it's wanted > > export INITRD=$(if_enabled_echo CONFIG_BLK_DEV_INITRD Yes No) > > > > +$(case $script in > > +postinst) > > + echo linux-update-symlinks install $version $installed_image_path > > + ;; > > +postrm) > > + echo linux-update-symlinks remove $version $installed_image_path > > +esac) This doesn't handle upgrades properly. The extra conditions in the maintainer scripts for official Debian packages really are needed. It's also not correct to assume that linux-update-symlinks is always available. If the generated linux-image package depends on "linux-base (>= 4.1~)" then the postinst script can use it unconditionally. However, the postrm script will always have to check whether it is present. Ben. > > test -d $debhookdir/$script.d && run-parts --arg="$version" --arg="/$installed_image_path" $debhookdir/$script.d > > exit 0 > > EOF > > -- > > 2.26.0 > > > > -- Ben Hutchings Every program is either trivial or else contains at least one bug
Attachment:
signature.asc
Description: This is a digitally signed message part
