It's definitely critical (you can wind up with an unusable system
if it's broken). The attempt to land the anaconda webUI change
highlighted that we need to make this change.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
With the advent of alternative bootloaders (systemd-boot) whether
or not grubby gets installed should be dependent on whether grub
is selected as the bootloader. This currently happens correctly
with anaconda and the grub dependencies, so it can be removed
safely. That allows systemd-boot to be installed cleanly without
grubby drippings confusing it.
Sync F40 with F39 since the branch happened before the commit was
merged.
Signed-off-by: Jeremy Linton <jeremy.linton@arm.com>
It's critical for build because we always build packages in mock,
and it's critical for compose because we build live images in
mock.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
With the advent of alternative bootloaders (systemd-boot) whether
or not grubby gets installed should be dependent on whether grub
is selected as the bootloader. This currently happens correctly
with anaconda and the grub dependencies, so it can be removed
safely. That allows systemd-boot to be installed cleanly without
grubby drippings confusing it.
While anaconda may now request sdubby and systemd-boot, i'm
told the comment in the anaconda section isn't 100% accurate
as anaconda may try to install all those selections to the live
images. So, lets not add those to the anaconda comps yet.
Signed-off-by: Jeremy Linton <jeremy.linton@arm.com>
We merged the libertas WiFi firmware into a single package as
combined they're less that 2Mb, and we split out the collection
of QCom firmware to their own package.
Signed-off-by: Peter Robinson <pbrobinson@gmail.com>
The iwlwifi WiFi firmwares have been somewhat reorganised so
the firmware are collected into the three main groups as
per the upstream linux driver categories which means if
the driver is enabled all the supported devices will have
appropriate and make things a little easier to manage.
Signed-off-by: Peter Robinson <pbrobinson@gmail.com>
KIO should be installed by default as they provide functionality shared
between core apps (Dolphin in this case) and can not be packaged as a
Flatpak.
See: https://pagure.io/fedora-kde/SIG/issue/354
This reverts commit f380d49d25.
We have to untag the new libproxy for now as it broke other
stuff - https://pagure.io/releng/issue/11434 - so we shouldn't
make this change in comps yet.
We probably don't need to explicitly list libproxy as its
replacement, as gnome-shell requires it.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
The various DVB firmware are used by some old DVB terrestrial/satellite
recievers. They're all pretty old and the drivers were nearly removed
from the kernel but were kept at the last minute.
Signed-off-by: Peter Robinson <pbrobinson@gmail.com>
`make sort` won't usually *fail*, even if it sorts stuff (it'll
only fail if it actually can't sort for some reason). So when
either the PR or the existing state of the repo isn't properly
sorted, what we get is the "git status failed!" case. I *think*
this is likely the only time we'll get that, I don't think
"make validate" would actually change files without failing. So
let's gloss that a git status failure probably means a sorting
issue.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
In util-linux-2.39.1.fc39 the separate util-linux-user subpackage
is dropped and the things it contained moved (back) into the
main util-linux package. These were initially split out in 2016
to avoid a libuser dependency in util-linux; not sure why this is
not considered useful any more.
Signed-off-by: Adam Williamson <awilliam@redhat.com>
