| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With some peculiar device we suddenly (during the service details discovery)
got a crash with CBDescriptor suddenly becoming something else - NSString,
NSMutableArray etc. - meaning the object was deleted and its memory re-used.
It would appear, CBPeripheral can suddenly change it's services tree and
it informs its delegate (aka 'us') about this change using the (previously)
missing method. In this method we cannot do much, due to the specificity of our
public API that allows concurrent discoveries, it's 'non-monolitic' (in several
steps) discoveries etc. etc. So the only thing we can do - stop everything,
remove all services, transition to QLowEnergyController::ConnectedState and
wait for a user to re-discover services.
Fixes: QTBUG-75043
Change-Id: Ie98d90aea112e40b4c6771e3f7315772dfd92b39
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| |
|
|
|
|
|
|
|
|
|
|
| |
From CoreBluetooth's point of view - this is just a callback, not a
descriptor write request (they even consider all descriptors immutable
in peripheral), but for Qt it should be a descriptor write operation,
and we must emit a signal.
Oh, and set endHandle correctly on a LE service object!
Change-Id: I71922507a6ece987ad8b5c317ef618301ae240c2
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Instead of hardcoded 10 second timeout use the latest Qt API enabling
arbitrary or infinite length for LE device discovery.
- Since now scan can take long or unlimited time, the previous approach
with detached CBCentralManager and DDA is not working anymore -
we can not report devices only at the end of scan.
Re-use LENotifier and signals/slots instead.
- Remove DDA timer, timeout logic is completely inside
osxbtledeviceinquiry object now.
Change-Id: I07735581ac175d1d5e5d788aaf697dcb41429e31
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
|
| |
|
|
|
|
|
|
|
| |
CoreBluetooth has CBPeripheralManager/CBMutableService both on iOS (since 6.0) and OS X
(>= 10.9). This lets me implement the Qt's BTLE 'advertisement'
API and peripheral role for both iOS and OS X.
Change-Id: I3e69a5870535a45bc16bbd9862ca84300b01daca
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@theqtcompany.com>
|
| |
|
|
|
|
|
|
|
|
| |
From Qt 5.7 -> LGPL v2.1 isn't an option anymore, see
http://blog.qt.io/blog/2016/01/13/new-agreement-with-the-kde-free-qt-foundation/
Updated license headers to use new LGPL header instead of LGPL21 one
(in those files which will be under LGPL v3)
Change-Id: I856c13e2a6d4d12c46e1286b0ca1c092ee4608f8
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
ATM CBCentralManager's delegate does its work on the main queue. With
CoreFoundation event dispatcher it's now possible to use QtBluetooth from
non-gui thread (more generally - from a thread other than main).
This makes main queue useless - and we have to move to a dedicated dispatch
queue. Also, we have to make sure we do not have race conditioins/dead-locks.
This patch:
1. decouples OSXBTLECentralManager and QLowEnergyController so that these
two objects working (potentially) on different threads do not share any data
and do not have to use locks, removes the explicit 'delegate' interface/inheritance
and replaces them with LECentralNotifier class - to be able to use Qt's signal/slot mechanics
for inter-thread communication.
2. all OSXBTLECentralManager's are now executed on qt_LE_queue queue
to avoid any race-conditions (since they potentially update manager's
internal state).
3. Results/errors are now reported using LECentralNotifier's object
(QLowEnergyController has corresponding slots connected to the notifier)
Task-number: QTBUG-49476
Change-Id: Ie07cdc13ad559c96a7d2ff010843fb7bcce07c99
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
|
