|
|
|
|
|
|
by hyperion2010
1468 days ago
|
|
|
I love the point in there that it isn't actually xorg vs wayland, it is xorg vs dbus. Wayland is so deficient that basically everything has to depend on dbus if it doesn't want to use X. This framing clarifies the issue substantially, because whatever people think about wayland, they might have some slightly different opinions about dbus. |
|
|
Recently i updated my machine and it would fail to boot because NetworkManager-wait-online.service's invocation of `nm-online -s` would fail even when NM was connected, even when `nm-online` actual liveness check would succeed.
I spent hours reading NM code, wading through auto-generated GObject introspections and their XML bullshit to try to figure out why org.freedesktop.NetworkManager.startup was true, what the magic numbers in org.freedesktop.NetworkManager.Connection.Active.StateFlags meant, why my desktop wouldn't boot and couldn't even find what could could cause the state to change before I finally gave up and patched nm-wait-online to just invoke the codepath which did an actual liveness check rather than bumble through a bunch of dbus interfaces.
Gotta say I was missing the old KDE3 dcop after that... IPC is still such a PITA on linux.