This reverts commit de4d00ddf5eecffa1e20fb5e1b3eeb7c91d641d2. We
gave up on the 'run gnome-initial-setup on live boot' approach a
couple of years ago, with GNOME 47 Beta, because the patches
were not merged upstream and maintaining them downstream became
too onerous:
so we should have reverted this at that time, too. We only have
not noticed any problems more or less by coincidence: creation
of /var/lib/gnome-initial-setup has not been working in the live
environment, which means we wind up creating a liveuser account
anyway because the check for /var/lib/gnome-initial-setup in
this code fails.
Recently on ELN, though, /var/lib/gnome-initial-setup started to
get successfully created in the live env, so we don't create a
liveuser account, and so g-i-s runs. This should fix that, and
also mean we don't suddenly started seeing g-i-s on Rawhide if
/var/lib/gnome-initial-setup starts showing up again there.
This reverts commit de4d00ddf5eecffa1e20fb5e1b3eeb7c91d641d2. We
gave up on the 'run gnome-initial-setup on live boot' approach a
couple of years ago, with GNOME 47 Beta, because the patches
were not merged upstream and maintaining them downstream became
too onerous:
https://src.fedoraproject.org/rpms/gnome-initial-setup/c/3cd24ee0efb277932c28e3e39fd09b49870fe128?branch=rawhide
so we should have reverted this at that time, too. We only have
not noticed any problems more or less by coincidence: creation
of /var/lib/gnome-initial-setup has not been working in the live
environment, which means we wind up creating a liveuser account
anyway because the check for /var/lib/gnome-initial-setup in
this code fails.
Recently on ELN, though, /var/lib/gnome-initial-setup started to
get successfully created in the live env, so we don't create a
liveuser account, and so g-i-s runs. This should fix that, and
also mean we don't suddenly started seeing g-i-s on Rawhide if
/var/lib/gnome-initial-setup starts showing up again there.