| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
|
|
| |
Add an option to set the TimeThreshold.
https://bugs.freedesktop.org/show_bug.cgi?id=105993
|
|
|
|
|
|
|
|
|
|
| |
Some sources will be able to make use of this to reduce battery life so we
should forward this information to them, via the Locator class. Currently,
only ModemManager can do that but that's also where it makes most sense
since GPS is the only source that can continuously fetch location updates
and hence drain the most power.
https://bugs.freedesktop.org/show_bug.cgi?id=105993
|
|
|
|
|
|
|
|
|
|
|
| |
Unlike other (real) location sources, Locator instances are unique for each
client application. Which means we only need just one time-threshold value
so we now provide getter and setter for time-threshold as guint, which
is what will be set by the clients, instead of the time-threshold property
of the parent class LocationSource (they can still do that if they wish
though).
https://bugs.freedesktop.org/show_bug.cgi?id=105993
|
|
|
|
|
|
|
|
| |
Implement the 'time-threshold' property of Modem interface. This
implementation translates this to ModemLocation::gps-refresh-rate property
of ModemManager service.
https://bugs.freedesktop.org/show_bug.cgi?id=105993
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add 'time-threshold' property of type guint, not GClueMinUINT to Modem
interface. In a following patch, ModemManager implementation will
translate this to ModemLocation::gps-refresh-rate property of
ModemManager service.
There is no reason to use the more complex GClueMinUINT type here since
the Modem implementations are at the last end of the time-threshold users
and hence just needs one specific value to be set on them.
https://bugs.freedesktop.org/show_bug.cgi?id=105993
|
|
|
|
|
|
|
|
| |
In the following patches, this property will be used to set the GPS
refresh rate on the ModemManager, based on the D-Bus TimeThreshold
property.
https://bugs.freedesktop.org/show_bug.cgi?id=105993
|
|
|
|
|
|
|
|
|
| |
This is a helper class that keeps a list of guint values and the minimum
value from this list. In the following patches, it will be used by
location sources to use the minimum time-threshold (location update rate)
from all the time-thresholds requested by different applications.
https://bugs.freedesktop.org/show_bug.cgi?id=105993
|
| |
|
| |
|
|
|
|
|
| |
In a19221e5fe (modem-manager: Instant GPS location update), we started
using the API introduced in 1.6.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A client process could create and destroy GClueSimple multiple times in its
lifetime. Since the service keeps the Client (and it's associated location)
object alive until the D-Bus connection from the client process remains
alive and gclue_simple_new*() awaits for a location fix on the client it
gets from service, subsequent calls to gclue_simple_new*() (after first
one) will likely not return anytime soon.
With this patch, GClueSimple now checks for any existing location object on
the client object and doesn't keep waiting for location update if one is
available already.
|
|
|
|
|
|
|
|
| |
GPS location updates are every 30s by default in ModemManager. This commit
sets this rate to 0 (update as soon as the GPS sends a new location) so
Geoclue can be used in real-time applications (like routing).
https://bugs.freedesktop.org/show_bug.cgi?id=101975
|
|
|
|
|
| |
Explain further in a comment that the apps white-list is not meant to list all
the apps in the world.
|
|
|
|
|
|
|
|
|
|
| |
Commit d3893e6fac09064682e880c7972f2fcf82a6fb87 introduced a constant
TIME_DIFF_THRESHOLD with value 60. In the summary of the commit this is said to
be 60s, however it is compared to the output of g_date_time_difference, which
is in microseconds. Therefore, if we want the threshold to correspond to 1
minute, TIME_DIFF_THRESHOLD should have value 60000000.
https://bugs.freedesktop.org/show_bug.cgi?id=101357
|
| |
|
| |
|
| |
|
|
|
|
|
| |
We were requiring altitude from NMEA GGA sentences and that later ends up
in a crash cause we return a NULL location without setting an error.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
gnome-shell now uses Geoclue to show weather information in the calendar
drop-down. While it can easily bypass the authorization dialog (being
the agent itself), the location indicator loses its usefulness when
permanently visible, so add it to the other system components in the
whitelist.
https://bugs.freedesktop.org/show_bug.cgi?id=100263
|
|
|
|
|
|
| |
This fixes a leak.
https://bugs.freedesktop.org/show_bug.cgi?id=100175
|
|
|
|
|
|
| |
Correct a copy&paste mistake in a doc comment.
https://bugs.freedesktop.org/show_bug.cgi?id=100176
|
|
|
|
|
|
|
|
|
|
| |
The gnome-settings-daemon color plugin uses Geoclue to guess the current
location and deduce the sunset and sunrise times. Make sure that the
functionality does not require a dialogue on startup when enabled.
See https://bugzilla.gnome.org/show_bug.cgi?id=779343
https://bugs.freedesktop.org/show_bug.cgi?id=100008
|
| |
|
| |
|
|
|
|
| |
We were unrefing an object that didn't belong to us.
|
| |
|
|
|
|
| |
This fixes cross-builds.
|
|
|
|
| |
Rather then asking build system to link to libgeoclue. This fixes cross-builds.
|
| |
|
|
|
|
|
|
|
|
| |
If the time since the last update is below the threshold,
it won't emit the LocationUpdated signal. The default value is 0.
When TimeThreshold is zero, it always emits the signal.
https://bugs.freedesktop.org/show_bug.cgi?id=94460
|
|
|
|
|
|
|
| |
Add a debug log when the distance from previous location is below
distance_threshold.
https://bugs.freedesktop.org/show_bug.cgi?id=94460
|
|
|
|
|
|
|
| |
We're about to introduce time-based threshold so let's rename
*threshold* to *distance_threshold*.
https://bugs.freedesktop.org/show_bug.cgi?id=94460
|
|
|
|
|
|
| |
The default value of DistanceThreshold is zero.
https://bugs.freedesktop.org/show_bug.cgi?id=94460
|
|
|
|
| |
https://bugs.freedesktop.org/show_bug.cgi?id=96521
|
|
|
|
|
|
|
| |
Now that xdg-app has been renamed, we also need to check for the
"flatpak-" prefix for the cgroup.
https://bugs.freedesktop.org/show_bug.cgi?id=96655
|
|
|
|
|
|
|
|
|
| |
This ensures g-ir-scanner uses the configured compiler. Fixes the
build in GContinuous, which is currently somewhat broken in that
`/usr/bin/cc` is a cross compiler by default. But this propagation is
the correct thing to do always.
A similar patch landed in GStreamer.
|
|
|
|
|
|
| |
It's not required to unref for the return of gclue_location_get_timestamp.
https://bugs.freedesktop.org/show_bug.cgi?id=96178
|
|
|
|
|
|
|
|
|
|
|
| |
Instead of just checking if some network is available, let's check if we
are fully connected to internet or not.
If glib >= 2.44 is available, we make use the new API added in that
version to check for connectivity but if it's not, we check if Mozilla
Location Services hostname is reachable.
https://bugs.freedesktop.org/show_bug.cgi?id=95493
|
|
|
|
| |
https://bugs.freedesktop.org/show_bug.cgi?id=95493
|
| |
|
| |
|
|
|
|
|
|
| |
If available.
https://bugs.freedesktop.org/show_bug.cgi?id=94157
|
|
|
|
| |
https://bugs.freedesktop.org/show_bug.cgi?id=94157
|
|
|
|
|
|
|
|
|
|
| |
And drop the now unused gclue_location_new_with_description().
This will allow to also set the timestamp in the next commit,
as that property is construct only and couldn't be set if using
gclue_location_new_with_description().
https://bugs.freedesktop.org/show_bug.cgi?id=94157
|
|
|
|
| |
https://bugs.freedesktop.org/show_bug.cgi?id=94521
|
|
|
|
|
|
|
|
|
|
| |
g_variant_n_children() is unsigned, so comparing < 0 will always be
false; this looks like a typo for <= 0, which would avoid allocating the
GVariantIter unnecessarily.
Spotted by Coverity (CID: #78509).
https://bugs.freedesktop.org/show_bug.cgi?id=94669
|
|
|
|
|
|
|
|
| |
This fixes a memory leak on exit.
Spotted by Coverity (CID: #78510).
https://bugs.freedesktop.org/show_bug.cgi?id=94669
|
|
|
|
|
|
|
| |
When run from outside the srcdir, autogen.sh was failing because various
of the tools it calls expect to be called from within srcdir.
https://bugs.freedesktop.org/show_bug.cgi?id=94649
|