| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
Also no need to shell out to logger as this is typically started by
systemd and thus any stderr output will end up in the journal (and
by extension syslog if it's installed).
Might help avoid numerous reports of the service 'failing' as
indicated in mga#8472.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
LESS=-MM breaks colorized output of git, and maybe perl-doc as well
(Mdv #57143).
It seems no other distro is shipping this as default,
should be safe to remove.
|
| |
|
|
|
|
|
|
| |
This can happen on the ifcfg-Auto* files created by NM. It may be
more sensible overall to just skip such files, but incase any other
invalid files are found, they are now simply skipped to avoid errors.
|
|
|
|
|
|
|
|
|
|
|
| |
As harddrake is run by mandrake_everytime (which starts Before=basic.target)
and it, in turn may try to start cpufreq (which starts After=basic.target)
a deadlock can occur.
To work around this, we pass the --no-block option to systemctl
to return flow and allow us to continue.
Hopefully solves mga#4772
|
|
|
|
|
|
|
| |
Trust link detection to Network Manager rather than doing more work ourselves.
Also ensure we wait for resolvconf even when we are using NM as this is still
needed.
|
|
|
|
|
|
|
|
| |
This change ensures that systemd will wait until this script
completes (and thus network is up) before we allow any logins.
Also provide $named such that this translates to nss-lookup.target in
systemd.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
Both scripts will check to see if NetworkManager is running before deciding
to ignore interfaces. This directive will ensure that systemd starts
NetworkManager before running these scripts and thus any checks and probes
will be valid.
This should prevent the double running of wpa_supplicant if the network
script was run before NetworkManager which can cause constant disconnects.
|
|
|
|
| |
Also do not wait for NM devices with no link.
|
|
|
|
| |
Like done upstream in commit 300aaee62cd7a3166969e795b53b065c510e09c1
|
| |
|
|
|
|
| |
https://bugs.mageia.org/show_bug.cgi?id=1216
|
|
|
|
|
|
|
| |
When building gcc/ecj1, TMP/TMPDIR is used and this results in debugedit
being unhappy.
(patch originally from svn: 77112 rtp - Remove trailing whitespace at the end of TMPDIR/TMP)
|
| |
|
| |
|
|
|