Newer development versions of libgudev now strip the linefeeds by
themselves, so fix our naive linefeed-stripping which munched on the
last character of the string if there was no linefeed.
Could not find a percentage for capacity level 'Ful'
See https://gitlab.gnome.org/GNOME/libgudev/-/merge_requests/26
This fixes the serial number not being set on Bluetooth devices which
might not have a serial number but should always have a Bluetooth
address to differentiate them.
We were correctly handling an "Alias" property changing, but never
passed the property to that code as we were dropping anything that
wasn't a battery related update.
otherwise the native build on openbsd complains:
../src/openbsd/up-backend.c:278:23: warning: variable 'new_state' is uninitialized when used here [-Wuninitialized]
new_time_to_empty = (new_state == UP_DEVICE_STATE_DISCHARGING && a.minutes_left > 0 ? a.minutes_left : 0);
^~~~~~~~~
regression from 8be73b98 ?
The version number has been bumped to be able to maintain multiple
branches without conflict. This version bump is not associated with a
API/ABI break.
This switches to always use the power/current reading if we had a
sensible reading at some point in the past.
In contrast to the older UPower code, will will however ignore the
reading for 10 seconds after a discontinuity (power plug/unplug or
resume) unless we read an explicit zero value after the event happened.
We do this under the assumption that the readings may be wildly off, and
it is better to not show an estimate rather than one that is wildly
incorrect.
Note that this commit also normalises negative power readings while
discharging to be positive.
Closes: #199
Estimation code can request a battery poll if the value is not good
enough at the point. Make this a little bit more explicit by renaming
the intenral variable to "repoll_needed" and automatically resetting it
to FALSE.
The up_device_battery_estimate function did more than just estimating
the current power consumption (and doing some state guessing). Move the
time to full/empty checking out of the function. Also, let it directly
modify the reported state before it is pushed into the ring-buffer.
When polling is resumed the timeout needs to be reevaluated. This
requires running the polling handler once (in the next mainloop
iteration).
Set the ready time to zero to ensure this is happening. Without this, we
would be stuck without actually polling until we get a uevent from the
kernel on one of the power supplies.
Fixes: #198
When a battery is swapped the old history needs to be saved and the
other history should be loaded.
Change the code to load the history lazily when needed. Then, to reload,
we purely need to clear the history object and it'll be loaded again
when required.
It seems that the test was still flaky, the reason for that would be
that we did not explicitly wait for the log line saying that the
aggregate state was calculated.
The only reason that it did not consistently fail appears to be that
searching for the state conflict caused messages to be skipped. That is
wrong, we should account for every "Calculating percentage" message to
ensure that upowerd and the test is in sync.
Assuming we have some estimation for the current battery capacity (i.e.
percentage), we can infer a FULL/EMPTY state. Do so if the battery state
is unknown.
Related: #196
This should be quite robust, in particular as we should be getting
notifications about AC plug/unplug.
The value for the battery will lag a few seconds. However, the
DisplayDevice will do some guessing taking the AC state into account,
and as such the user should get at least some immediate feedback.
Closes: #196
This makes the switch. There are a few behaviour changes with regard to
estimations (which hopefully got both simpler and more robust at the
same time).
We keep giving people these commands for bug triaging. So, lets hope
that adding them to the README removes some of the overhead and can be
helpful to users.
The state guessing code based on the AC state was not tested well.
Improve the test by testing both 1 and 2 batteries and checking the
reported state in more detail.
There is no reason to not guess the state if the device has no AC power
and there is more than one battery. Remove the corresponding constraint.
Related: #146
Otherwise we may not have the percentage to work with, rendering the
guessing useless. This also moves the time estimation down (after the
state guessing) and does it unconditionally. This is, however, not an
issue, as the calculation matches with other places.
Related: #146
Using energy is broken as the value might be zero if it is not provided.
However, either energy or percentage ("capacity") should have been read
from the sysfs. And, in both cases the percentage should reflect
something reasonable.
Related: #146
The state aggregation test requires an idle handler to run, which can be
a bit unreliable as it may or may not run twice.
Force running it twice and add code to wait for it to complete. Do so
properly by waiting for the correct log messages rather than sleeping so
that everything is ordered nicely while not slowing down the test a lot.
Closes: #193
If we have an input node with both a keyboard and a mouse in one node,
then prefer the mouse. However, if we have separate mouse and keyboard
device nodes, prefer the keyboard.
Closes: #189
This adds a test that checks all possible inputs for the battery state
aggregation. Not all states are actually tested in this case, as some
states may be handled through state inference and others might not be
well defined.
Stop trying to guess a battery state based on the AC information for a
single battery. Instead, just do the guessing in the display device.
This means that cases with more than one battery work fine. It also
means that we still report an UNKNOWN state for the battery itself.
Also get the on_battery information from the display devices discharging
state, there is no need to iterate all of the batteries to figure that
out.
Note that we don't really have a well defined state for the display
device in all cases. We'll add a test in the next step, marking those
cases as TBD or ANY to show where the issues are. AFAICT we do not have
a regression though, so this should be fine for now.
Fixes: #146
This fixes the previous commmit to 190d12e292 ("common: Add a file
with common helper") which broke the string handling for the freebsd
backend.
Related: !144
The umockdev code to disable it by setting the variable does not
propagate through python. Set the variable here until there is a more
permament solution available.
This removes some code duplication between the linux and freebsd
backend. And, this file could become home to other small helper
functions in the future.
Using tearDown is brittle, as an assertion will stop it from running
through completely. So move cleanup into a helper that is called via
addCleanup for logind and bluez.
