Probably not. I’m just pointing out the reason for those denials. The loader tries the rpath first, and thus we get a denial.
The render queue problem seems related to this:
starting kdenlive_render process using: "/snap/kdenlive/39/usr/bin/kdenlive_render"
Qt: Session management error: Could not open network socket
No org.kde.JobViewServer registered, trying to start kuiserver
Failed to start kuiserver
Ah very insightful, this seems to imply that the kde-frameworks-5-qt-5-15-3-core20 snap is built incorrectly, I’m not sure why it was built such that the dynamic libraries there have rpath setup to look in /snap/core20/current/… unless that content snap was built to be consumed by classic snaps somehow?
Do you see any journal denials on your system when this happens? I don’t see this message on the terminal when I tried testing kdenlive last week by just doing snap run kdenlive
I think it’s a dbus permission problem. If you install with devmode and set QDBUS_DEBUG=1, you can see the working communication between kdenlive_render and kdenlive.
I think it fails here:
QDBusConnectionPrivate(0x7f52e0003c00) sending message: QDBusMessage(type=MethodCall, service="org.freedesktop.DBus", path="/org/freedesktop/DBus", interface="org.freedesktop.DBus", member="ListNames", signature="", contents=() )
QDBusConnectionPrivate(0x7f52e0003c00) got message reply: QDBusMessage(type=Error, service="org.freedesktop.DBus", error name="org.freedesktop.DBus.Error.AccessDenied", error message="An AppArmor policy prevents this sender from sending this message to this recipient; type=\"method_call\", sender=\":1.286\" (uid=1001 pid=68175 comm=\"/snap/kdenlive/x4/usr/bin/kdenlive_render /snap/kd\" label=\"snap.kdenlive.kdenlive (enforce)\") interface=\"org.freedesktop.DBus\" member=\"ListNames\" error name=\"(unset)\" requested_reply=\"0\" destination=\"org.freedesktop.DBus\" (bus)", signature="s", contents=("An AppArmor policy prevents this sender from sending this message to this recipient; type="method_call", sender=":1.286" (uid=1001 pid=68175 comm="/snap/kdenlive/x4/usr/bin/kdenlive_render /snap/kd" label="snap.kdenlive.kdenlive (enforce)") interface="org.freedesktop.DBus" member="ListNames" error name="(unset)" requested_reply="0" destination="org.freedesktop.DBus" (bus)") )
which means it probably needs the system-observe interface. I tried and this seems to solve the problem. Is that an interface that makes sense for a video editor, and would autoconnect be granted?
Yes I think system-observe is reasonable to be granted auto-connection for a video editor, if one of the snap collaborators can add system-observe to the snap and then start a forum topic in #store-requests for auto-connection we can get that moving.
I’m still very confused why the qt content snap has these rpaths set like that, it probably needs to be looked at more
From: Jonathan Riddell jr@jriddell.org
Date: Tue, 9 Feb 2021 10:52:34 +0000
Subject: [PATCH] enable patchelf as we can not do the hack of patchelf
manually as we did in core18 (see Rakefile) cos now snapcraft is installed as
a readonly snap
I can still reproduce the problem with the snap “latest/stable: 21.12.2 2022-02-10 (48)”.
The rendering also gets stuck in “waiting…”.
Kdenlive log:
Qt: Session management error: Could not open network socket
dbus[2237125]: arguments to dbus_message_new_method_call() were incorrect, assertion "_dbus_check_is_valid_path (path)" failed in file ../../../dbus/dbus-message.c line 1366.
This is normally a bug in some application using the D-Bus library.
D-Bus not built with -rdynamic so unable to print a backtrace
For me, kdenlive works fine on 22.04 as long as there are clips in the timeline. The usual warning dialog if you don’t have any clips in the timeline does not appear in the snap though.
But I had a clip in my timeline.
BTW I have a AMD GPU so there is no NVIDIA.
Is there a command to run kdenlive on dev mode so I can send more logs for you guys?