GIT_SILENT

Unfortunately the statistics refresh rate of a device is a global property. So
when it is changed from the outside we just force it back to one second. The
initial rate is saved and updated when something tries to change it in order
to restore it upon destruction.
Because NetworkmanagerQt just forwards the DBus signal we have to guard the
slot with a boolean to we do not trigger it ourselves.
BUG:433063
FIXED-IN:5.21.1


(cherry picked from commit 4b5251ad)

These should be 'read' and 'write' instead of readRate and 'writeRate'.
This matches the old sensor names and what the applet expects.
BUG:432851


(cherry picked from commit ccca59fd)

GIT_SILENT

David Redondo authored 3 years ago and ngraham committed 3 years ago

BUG:430676
FIXED-IN:5.21.0


(cherry picked from commit 09205b22)

Classical case of off-by-one. The documentation starts numbering at one.
I used the numbers from it instead of the proper ones.
BUG:432637


(cherry picked from commit 34705df7)

In order to read the file it not only does need to exist but we also have to open it.
This broke some systems where a file existed but was not readable with current permission.
If we check if we could open a file we can save the check if the file exists.


(cherry picked from commit b3bee7ef)

We are searching for a StorageDrive to see if it is on a HardDisk or or some
other type of device. However on some setups a Volume will not have StorageDrive
as an ancestor, this is for example the case when using lvm. So let's include
every volume by default and exclude only those for which we find a Drive that is
not a hard disk.


(cherry picked from commit 8747897a)

Fix copy and paste mistake
GIT_SILENT


(cherry picked from commit 11fc8be0)

(cherry picked from commit 0e5a96ca)

GIT_SILENT

GIT_SILENT

In certain cases, NetworkManager can go on a rampage sending active
connection change signals in very rapid succession. This results in
memory usage ballooning because lambda connections can not be unique.

Rather than destroying and recreating the NetworkManagerDevice objects
on active connection change, this changes things so that we only remove
them from the SensorContainer but keep the actual objects around. This
allows NetworkManagerDevice to track its active connection state,
removing the need for constant reconnects.

BUG: 430003

This prevents a crash where the SensorProperty could disappear without
the AggregateSensor detecting it.

Otherwise clients looking for the interface will receive an error when the name
is announced but there is not the wanted interface.

Do not search for libpcap and consequently do not build the plugin

We rely on the prefixes on the front page of plasma-systemmonitor; spare every
subclass from having to do it.

Queries information about current network devices using the rtnetlink kernel
interface. This exposed all the info except a pretty name for the current
connection and wireless strength. We can use this as a fallback on Linux if
NetworkManager is not available.
Because this is our fallback on Linux libnl is now required (from recommended).
BUG:425992

This allows us falling back to something else if our preferred backend
is not supported at runtime.

Since we want to report speed per second and the daemon is now running
faster than this.

Since 2000ms is rather slow

GIT_SILENT

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

The property is deprecated and does not make a visible difference,
because the names of the tabs are set in the downloaded XML file.

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

GIT_SILENT

In case of conflict in i18n, keep the version of the branch "ours"
To resolve a particular conflict, "git checkout --ours path/to/file.desktop"

We need to update sensor properties if either the physical object or the
swap object is subscribed (or both).

We do not anymore use a subsystem which was using this method.

Uses mostly the same methods to get the information as the old code.
On FreeBSD the method to get the swap values is much simpler with just one
function call.