| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |\
| |
| | |
Fix/sdl crash during set app icon
|
| | |
| |
| |
| | |
Update SetAppIcon request according to the changes in file system
|
| | | |
|
| | |
| |
| | |
* Fix versioning appliance for vehicle data
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Fix misplaced OnDataVideoStreaming
OnDataVideoStreaming notification was sent for each application
independently. Now one would be sent from application_manager with
respect to service status
* Fix OnDataStreamingNotification
* rename parameters
* fixup! rename parameters
|
| |\ \
| | |
| | | |
Fix BC.PolicyUpdate double sending
|
| | | | |
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | | |
* Fix wrong success code for ReadDID request truncated data
* fixup! Fix wrong success code for ReadDID request truncated data
|
| | | | |
|
| | | | |
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
App. (#2365)
* Bugfix `SDL forwards OnButtonEvent notification`
Bugfix `SDL forwards OnButtonEvent notification
of CUSTOM_BUTTON to BACKGROUND App.`
Issue number #967
* Check that SDL forwards OnButtonEvent notification of CUSTOM_BUTTON to BACKGROUND App
* fixup! Bugfix `SDL forwards OnButtonEvent notification`
Co-authored-by: Ira Lytvynenko (GitHub) <ILytvynenko@luxoft.com>
Co-authored-by: Igor Gapchuk (GitHub) <41586842+IGapchuk@users.noreply.github.com>
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
A crash occurs because we use an overload of
boost::beast::websocket::stream::write without error_code.
After the error occurs we add an error message to the logger
and continue working.
In SendMessage_UnpreparedConnection_WithoutFall test we use sleep because we can't ability to catch processed event.
This issue is that LoopThreadDelegate requires a WebsocketSession to work.
WebsocketSession init LoopThreadDelegate with 'this', so we have no way to simulate it without additional installers.
In future, to avoid sleep, we can:
Create a wrapper class for the socket, then provide the socket layout for the WebsocketSession in unit tests and wrap the following interfaces:
async_read
write down
get_executor
async_accept
|
| | | |
| | |
| | |
| | |
| | |
| | | |
In case of duplicated name of application SDL didn't increment
'count_of_rejections_duplicate_name' counter in policy db.
Added incrementing of counter.
|
| | | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
There is the issue when SDL incorrect validates display capabilities
as a Smart Object. The route cause of the issue is the missed "alertIcon"
parameter in mapping <string value, hmi_apis::Common_ImageFieldName::eType>
. During validation, a Smart Object SDL can't find mandatory parameter and
fail validation. As a result, SDL doesn't set display capabilities into
the HMI capabilities entity.
|
| | |/
|/|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Fix undefined behavior after SO cleanup
After cleanup_data() call, delete is used to deallocate
memory by pointer, however pointer still hold a value
of address of deallocated memory. By that reason SO
still have an ability to access that data what might
cause UB including core crash in the random places.
This function has been updated to set pointers to NULL
as well as smart object type to prevent accident access
to deallocated memory.
* fixup! Fix undefined behavior after SO cleanup
* Fix CSmartSchema corruption while copying
The issue is that HMICapabilities class is storing
different capability fields by raw pointer. In some
cases when one thread is trying to copy smart object
by pointer, another thread may try to set the same
field and destroys a previous one. This may cause a
half-copied smart object instance, like smart object
with the schema pointing to already destroyed schema
item. When such smart object is destroyed, we may see
a core crash or unexpected SDL behavior.
To avoid such issues, raw pointers were raplced with
shared pointers, set by swap() operation. This provide
a guarantee that object will not be destroyed while
copying by some thread.
|
| |\ \
| |/
|/| |
Update develop with hotfix 6.0.1
|
| | |\
|/ /
| | |
Hotfix Release 6.0.1
|
| | |\
| | |
| | | |
Fix application of hybrid app preference parameter for cloud applications
|
| | | | |
|
| | | | |
|
| | | | |
|
| | |/ |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Add the new ApplicationEvent.
Add the new ApplicationEvent (kRCStatusChanged) for informing RcRPCPlugin
that OnRCStatus notification should be sent.
Split adding RcExtension to application and sending OnRCStatus notification
to the separate actions.
* Fix SDL crash during application registration.
There is an issue when SDL crashes during registration second application
in the time when already exists a registered other one application.
In the current implementation ApplicationManager during application
registering by the ApplicationManager::RegisterApplication method creates
a new instance of the application and adds it to the collection of all
registered applications. This method calls in the
RegisterAppInterfaceRequest, but all needed extensions will be added only
by the ApplicationManager::OnApplicationRegistered method during sending
RegisterAppInterface Response to the mobile.
According to the described behavior could appear case when a first
application is registered successfully and all needed application
extensions are successfully added too. The second application starts
the registration process in someone thread, but in the other one thread,
SDL receives PolicyEvent::kApplicationPolicyUpdated event and tries to
unsubscribe all applications from removed in policy table Vehicle Data
Items. All actions go on in the next sequence:
Precondition:
The first application is successfully registered and all needed
extensions are added.
The first thread:
1. The Mobile device sends RegisterAppInterfaceRequest RPC to the SDL for
registration of the second application.
2. SDL receives the RegisterAppInterfaceRequest and calls method Run.
Within method RegisterAppInterfaceRequest::Run ApplicationManager runs
3. the method RegisterApplication.
The second thread:
4. The SDL receives the PolicyEvent::kApplicationPolicyUpdated event and
forwards the event to each plugin for processing it.
5. VehicleInfoPlugin tries to unsubscribe all registered applications
from removed in policy table Vehicle Data Items.
The plugin finds the first application and unsubscribes it from all
removed Vehicle Data Items.
The plugin finds the second application and doesn't find any extensions
for the application and crashes on the DCHECK with an uninitialized
pointer to the extension.
|
| | | |
|
| | | |
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Check for conflicting app resumption level
* Change variable name
* Restore widgets before HMI level conflict check
* Add check for number of restored widget windows
* Address comments
* Fix active app check
* Fix typo
* Fix style
* Add helper method
* Add new app manager predicate finder
|
| |\
| |
| | |
Release 6.0.0
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Add unit tests
* Fix subscription to app services
* fixup! Fix subscription to app services
* fixup! Add unit tests
* fixup! Add unit tests
* fixup! Fix subscription to app services
* fixup! Fix subscription to app services
* fixup! Add unit tests
* fixup! Add unit tests
* fixup! Add unit tests
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
* Provide operation result to function file_system::CreateDirectory
The function CreateDirectory signature (return value) was changed
in file_system.h
* Added const cv for the creating directory path
into unit tests(when uses file_system::CreateDirectory)
* Fix merge conflict
|
| | |\
| | |
| | | |
Fix custom vehicle data limits
|
| | | | |
|
| | |\ \
| | | |
| | | |
| | | |
| | | | |
smartdevicelink/fix/fix_synchronize_backup_and_reset_functions_in_cache
Synchronize backup and reset policy functions
|
| | | | | |
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This is an issue specific for EXTERNAL_PROPRIETARY flow
only. This issue is exposed by PolicyHandlerTest unit test:
ResetPolicyTable_WithPreloadedFile_ExpectPolicyTableReset
which sometimes is failing.
The main problem was that database reset(drop) and backup
can be done simultaneously in different threads which were
not synchronized. In that case, database drop often may be
failed because another thread is trying to write something
into the DB table which is going to be dropped. This causes
above unit test to fail sometimes.
To make that unit test and policy component more stable,
these two threads were synchronized by additing a waiting
function, so one thread will be able to wait for another
thread to finish its own staff.
|
| | |\ \ \
| | | | |
| | | | |
| | | | |
| | | | | |
smartdevicelink/fix/custom_vehicle_data_array_key_translation
Fix custom VD key translation for struct arrays
|
| | | | | | |
|
| | | | |/
| | |/| |
|
| | | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* Fix log severity level in on_vehicle_data_notification.cc
Fixed log severity level in on_vehicle_data_notification.cc:
LOG4CXX_ERROR is changed on LOG4CXX_DEBUG and corrected a
mistake in string.
* Resolve merge conflict
* Fix Connection Typo
* Revert disabled unit test
|
| | |\ \ \
| | | | |
| | | | | |
Update copyrights for generated files
|
| | | | | | |
|
| | | | | | |
|
| | | |/ /
| |/| |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
StartService (#3082)
* Fix HMI level on StartServiceNAck
* fixup! Fix HMI level on StartServiceNAck
|
| | | | | |
|
| | |\ \ \ |
|
| | | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
There was found one corner case when SDL does
not change the internal status when service
start was refused due to current HMI level of
application. Was added missing check to set
status properly and send expected notification.
|
| | | | |/
| | |/|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
* Fix revoked VD items subscription after PTU
There was found an issue that SDL still kept
subscriptions for custom vehicle data items even
for case when they were removed during PTU. By
that reason, SDL was crashing with DCHECK on
attempt to unsubscribe from non existing custom
data.
There was added missing logic of internal
unsubscribe in case some custom vehicle data
was removed by policies.
* Fix sending of UnsubscribeVehicleData
Was added a missing part of logic of subscriptions
control - cache manager was updated to keep the
vehicle data which was removed by last PTU.
Custom vehicle data manager was updated to
consider the removed vehicle data and also prepare
and send unsubscribe requests to HMI for case when
some vehicle data was removed and some application
was subscribed on it.
* fixup! Fix sending of UnsubscribeVehicleData
* fixup! Fix sending of UnsubscribeVehicleData
* fixup! Fix sending of UnsubscribeVehicleData
* fixup! Fix sending of UnsubscribeVehicleData
|
| | | |/
| |/| |
|
| | |\ \
| | |/ |
|
| | | | |
|
