| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
cb19e1f7e3cc2cfee3b8fcbdfa5cd151826dadac fixed a bogus comparison in
an if statement.
However, the correct condition is actually a bug, and should be removed
altogether. The previously incorrect condition was always testing
to false so the problem wasn't triggered.
Task-number: QTBUG-72180
Change-Id: Ia00b55596173474e8c0fe3b0c374490620c6917f
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This patch introduces a new macro, QT_LOCATION_DEBUG, that, when
fed via qmake command line, changes the scope of some members,
adds some extra members and some extra data dumping into these
members.
In this way external code can inspect the state of the renderer
in order to debug it.
The patch also moves a few classes into separate files, as well
as privately exporting them.
Change-Id: I32a042863f78dcd10f94b0b7fe8b90f3fc486891
Reviewed-by: Paolo Angelelli <paolo.angelelli@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
According to MSDN GeoLocator's default ReportInterval is 1 second or as
frequent as the hardware can support – whichever is shorter. By using that
functionality we can support minimumUpdateInterval "properly".
Change-Id: Id54197babeec04c8fc0a309930b192e36d9138f8
Reviewed-by: Miguel Costa <miguel.costa@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Windows' backend will complain if we try to set the report interval
while being subscribed to position changes so we have to remove this
registration and redo it when the new interval is in effect.
Change-Id: I5b3f23a7b6e530dd0a3f673e2b8ce6f5102b7eb1
Reviewed-by: Andre de la Rocha <andre.rocha@qt.io>
Reviewed-by: Miguel Costa <miguel.costa@qt.io>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We do not use the native callback's result but Windows' backend will
complain if we try to set some of GeoLocator's properties while we are
registered for status changes. By removing the callback registration we
get rid of these warnings.
Change-Id: Ic3829ee438c708fb4411ba94ba1202bb427e8815
Reviewed-by: Andre de la Rocha <andre.rocha@qt.io>
Reviewed-by: Miguel Costa <miguel.costa@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
For some reason the native status is always Disabled for desktop builds
(even though the location service is running).
So with desktop builds we have to use requestAccess to gain information
about the state of the location service. As the functionality is async the
important parts of QWinRTFunctions are shadowed locally.
Change-Id: I575ee0b161de734c29453a7a07350bc8d09720e8
Reviewed-by: Andre de la Rocha <andre.rocha@qt.io>
Reviewed-by: Miguel Costa <miguel.costa@qt.io>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Windows 8.1 is no longer supported
Change-Id: I00d103057a63abbcbc3ad3baa8ef93b35e07e819
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Gets rid of duplicated code.
Change-Id: I4635b23c5ae431867299eeeafd4bd260f8ba156f
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: I641690a87dab2527493c3c16ffb476e4121d9ecb
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: Ic9df2c2870184eacc2d1400fc51354ee1da0c288
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
If the QGeoPositionInfoSource is created in a command line app we cannot
rely on CoInitialize having been called by the event dispatcher before.
Fixes: QTBUG-71194
Change-Id: Id2a8fd22f7b4cf6e5c2629bf7dcd8319b585666e
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
get_Altitude and the like are marked deprecated and thus should not be used.
Use proper replacement functions as advised in Microsoft's
documentation.
Change-Id: I2166611a3133cadf9fb110242f231d76ca517932
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: I74ce6e9433eb7ad0f065ea5c554a8c11dd690604
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
- Use nullptr instead of 0
- Use recommended way of including Qt headers
Change-Id: I170dbb5aa41a2dd4d10ff99c421da8a037cbe153
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
CategoryModel was missing an update call in the ::componentComplete
callback, as well as reacting on the name changed signal of the plugin.
An update call was also missing in the setPlugin method.
This patch makes it react on the attached signal instead, and
adds missing update calls.
Fixes: QTBUG-70254
Change-Id: I1a8de0137b4fe4af7c5ffc848799061147febba2
Reviewed-by: Michael Brasser <michael.brasser@live.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
A dangerous local reference with the same name of a member variable
caused the latter not to be evaluated and the former not to be set.
Change-Id: Iaadcc3bce36f56ed8eeb77374687cf33632abd2f
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Or else there's no reactive way to tell navigator objects when the user
changes this property.
Change-Id: Ief78facf537a50c15611127c5282c2bcfd59f8fd
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The incubating indices bookkeeping was wrong.
This patch removes the bookkeeping and relies on the delegate
model doing the right thing when a row is gone but the
delegate still incubating.
Fixes: QTBUG-71264
Change-Id: Ibf5e525aa7ac79faf2fa149b52def05893d0bcc0
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Versioned win32-msvc legacy scopes have been deprecated and might yield
wrong results. As we do not support any versions older than the previously
checked ones an unversioned check is fine.
Change-Id: I234c5204c58e76a1a685f24c2aba316b1fe50508
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
While dbus might be available on other platforms, the plugin might not be
functional and thus should not be built on these.
Change-Id: I7671a3ecaba6944f99317eee4f7738a7df98ba10
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This patch fixes rendering artifacts with polylines appearing
when rendering large polylines and/or at high zoom levels.
Two problems caused the artifacts:
1.A too close near plane when calculating the projectable region in
QGeoProjectionWebMercator, which presumably introduced numerical
errors when then using such a region to clip data to be projected
using a projection transformation based on the same frustum.
2.Projected polylines too large for qTriangulatingStroker, that would
then introduce artifacts at screen.
To solve 1., as a temporary solution, the distance of the near plane has
been increased to a value that seems safe for zoom levels < 19.
This while a better formula that scales further is being researched.
To solve 2., screen-space line clipping has been brought back from 5.8,
and added on top of mercator-space clipping. This, in theory, should
also increase the performance, allowing qTriangulatingStroker to process
less data.
Task-number: QTBUG-71607
Change-Id: Id774419dde819931e2fdd78b02081695a91302ef
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We can now pass parameters to qdoc commands - use them to automate
the version bumps for documented QML imports.
Also, improve the structure and language on QML module pages.
Change-Id: Id56d36a9bc550ab9a1aea1af6ff8da83e5e1feb9
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Paolo Angelelli <paolo.angelelli@qt.io>
|
| | |
| | |
| | |
| | |
| | | |
Change-Id: Ic49ab959a028adc9edd60018351386d51b2a9bca
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Removing a local variable never used.
Change-Id: Id2517d0f326b16b4b4ad26e53e4115a33f44c7c2
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
|/ /
| |
| |
| |
| |
| |
| |
| |
| | |
Adds support for setting up Mapbox China API base URL + China
government-approved styles.
Change-Id: I66d9e15f336804466285e0996b0e326ef288be44
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Reviewed-by: Paolo Angelelli <paolo.angelelli@qt.io>
|
|\ \
| |/
| |
| | |
Change-Id: I0f4a3b7bc7c5c7f8da5af631000a6d5156884efe
|
| |
| |
| |
| |
| |
| |
| | |
Fixes: QTBUG-71355
Change-Id: I47b058891a47cd6d020c4e192d40233c57814313
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Paolo Angelelli <paolo.angelelli@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
So plugins can use it, also stand alone ones, and no ugly
"import ../" anymore.
Change-Id: I4e53815e041e13d495026d60a826a31bccaef2b4
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Reviewed-by: Bramastyo Harimukti Santoso <bramastyo.harimukti.santoso@pelagicore.com>
|
| |
| |
| |
| |
| | |
Change-Id: Ie8af7ed00a2f97439f6d801b284a49cfcf281ab0
Reviewed-by: Paolo Angelelli <paolo.angelelli@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Since Qt 5.12, \code command accepts parameters, and in-line usage
no longer works (and was never documented to work).
Also, use \badcode for qmake snippets to make it appear as plain text.
Task-number: QTBUG-70980
Change-Id: I53cafdc912b4663d03c19b0a0003693330c94f15
Reviewed-by: Martin Smith <martin.smith@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
that result in a warning nowadays, saying we must have a description
in Info.plist to be able to call these methods.
Task-number: QTBUG-41827
Change-Id: Ib1997a609553345358aded419a403582ccfdc0fb
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| |
| |
| |
| |
| | |
Change-Id: I136ef0a3125b42b484e36413d26aa49e59a43b57
Reviewed-by: Paolo Angelelli <paolo.angelelli@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
The name was changed from html5 to wasm, so the condition needed
to be updated.
Change-Id: I94851dae5674c4bac7eb02f876a8251c09045f39
Reviewed-by: Paolo Angelelli <paolo.angelelli@qt.io>
|
|\ \
| |/
| |
| |
| |
| |
| | |
Conflicts:
src/imports/locationlabs/locationlabs.cpp
Change-Id: I362b903966b5087022fd125b511b236323c265ed
|
| |
| |
| |
| |
| | |
Change-Id: I3813b65b18877f5030aea4511ca7cb850874d7c7
Reviewed-by: Paolo Angelelli <paolo.angelelli@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Since it is possible that a category has unspecified visibility then
when comparing we still want to match on those. Otherwise it can end up
not finding any matches as a result.
Change-Id: I586bc3c1c91502ff56ea2974ab87a5dddc9d1263
Reviewed-by: Paolo Angelelli <paolo.angelelli@qt.io>
|
| |
| |
| |
| |
| |
| | |
Task-number: QTBUG-69617
Change-Id: I98bd56c07a5897cae62a9c287dab1cacd5f8d07d
Reviewed-by: Paolo Angelelli <paolo.angelelli@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
We were previously not requesting all relevant attributes, and deriving
some required info by ourself, in the wrong way.
So just request what's needed and use it.
Task-number: QTBUG-70499
Change-Id: I6e0d4b28da1412a4c13460d25e16e383f0aca9a0
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| |
| |
| |
| |
| |
| | |
Task-number: QTBUG-70264
Change-Id: I2922b855d08e42bbd771cc959ea9b6e97eed60b9
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This reverts commit 63c8b3bca03586994c2b59c262e9cb3a9b12b363.
With vSphere based VMs moved to KVM based, we can try
if the networking issues have been solved.
Change-Id: I41fd6845341df5722c1ba2e0cd76e01bf16df4fb
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Liang Qi <liang.qi@qt.io>
|
| |
| |
| |
| |
| |
| | |
Task-number: QTBUG-52660
Change-Id: I528bde71dd0ad5ad58e26a9d484364b35b6c4b62
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Relies on automatic service activation:
https://dbus.freedesktop.org/doc/dbus-specification.html#message-bus-starting-services
This means that the service isn't explicitly started but gets started
automatically when a request to the service is made. It's also possible
that the geoclue2 service gets shutdown during runtime (for example when
no client is requesting updates for some time). Automatic service
activation ensures, that the service gets (re)started whenever needed.
Create the manager interface right in the constructor as there's no
benefit in creating it lazily. Also, since automatic service activation
is used, m_manager->isValid() can't be used, because it always returns
false until the service has been (automatically) started.
Always return correct values from supportedPositioningMethods(). The use
of the m_manager->availableAccuracyLevel() shorthand helper doesn't tell
the difference between an invalid property (for example if geoclue2
isn't available at all) and GCLUE_ACCURACY_LEVEL_NONE since both return
0. Getting the QVariant of the property is needed to distinguish those
cases. An invalid QVariant means that we couldn't communicate to
geoclue2. A QVariant that can be turned into 0 means
GCLUE_ACCURACY_LEVEL_NONE. In this commit GCLUE_ACCURACY_LEVEL_NONE is
also correctly turned into NoPositioningMethods.
If client doesn't exist when startClient() is called, create a client
first. It's also possible that a client is crerated more than once
during runtime. For example, when the above mentioned automatic geoclue2
service shutdown takes place, the old client is no longer usable. Now
in this case the old client gets discarded and a new client is created.
Call startClient() in startUpdates() – because otherwise startUpdates()
wouldn't really work.
Remove the warning message from the beginning of configureClient(). It's
quite OK that configureClient() gets called before m_client exists (it
just doesn't do anything then). This was probably in place with the
assumption that setUpdateInterval(), setPreferredPositioningMethods()
and startUpdates() would get called in some specific order. But no such
assumptions can be made.
Delete client interface object once it's not usable anymore (either
gives an error or is explicitly stopped). This way, a new client gets
created once a client is needed again.
Other minor changes:
- In startClient(), check whether starting the client is needed before
doing it. This removes the need to do any checks before calling this
function.
- In stopClient(), check whether stopping the client is OK before doing
it. This removes the need to do any checks before calling this
function.
- Use QPointer for the D-Bus interface objects -> no need to set the
pointers to nullptr after deletion.
- Add explicit casts to few places that gave implicit signedness change
warnings.
Change-Id: If9db25775710fe64f272db222030e481533d9732
Reviewed-by: Volker Krause <volker.krause@kdab.com>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
Convenience method to test if a property is present.
Change-Id: I01c6748e35aadfd89eecebf8b8cc64a278bba87f
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
This has been missing until now.
Change-Id: Ida8f7aa0d14e549d1ad1a04a85d096ff5801b71e
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Change the clear method name from clearPath to just clear to follow
more closely the other Qt APIs.
Change-Id: I98d4ea43e69e3d365c2c470e9b8b14479f288fe6
Reviewed-by: Paolo Angelelli <paolo.angelelli@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|\ \
| |/
| |
| | |
Change-Id: I6b28d02c3bb589b4ea6672693d4978a6c342d181
|
| |\
| | |
| | |
| | | |
Change-Id: I1e2854bf5a4220d7c9252562c0583b704896df09
|
| |/
| |
| |
| |
| | |
Change-Id: I7045270108795be9d8a4dc1073748e1b9bb7d56f
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Up till now QGeoManeuvers inside QGeoRouteSegments were generated as
starting maneuvers, that is, related to the beginning of the segment.
But they were documented as being related to the end of the segment.
This patch fixes the documentation.
Change-Id: Ia08ed2bd597b4f63e6b8a7b3518755d895446a92
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Reviewed-by: Christopher Adams <chris.adams@jollamobile.com>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
All route segments after the second waypoint contain broken paths.
This patch fixes it refactoring the parsing logic, also adding support
for route legs.
Change-Id: I30ec9e889a5e16bf81c4a0caf1b59fc895b7ea5e
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|