Commit Graph

104 Commits (fc8f0f6716b4c08fbeff0af997121c575fc2c37c)

Author SHA1 Message Date
Peter Hutterer fc8f0f6716 __init__: remove empty line that snuck in during rebasing 2018-01-24 21:39:56 +10:00
Peter Hutterer 2ee934a81d kete: don't let a second client start searching 2018-01-24 19:25:00 +10:00
Peter Hutterer 8bb971cdcd tuhi: don't use a timeout for searching
We track the searching client now, so the timeout is no longer required
2018-01-24 19:25:00 +10:00
Peter Hutterer cfa4aca2df dbus: limit StartSearch to one client only
Basically copied from the device's Listening approach.

While it's possible to have multiple clients searching at the same time it's a
niche case and the effort at fixing the race conditions that come from that is
likely not worth the effort.

Let's add multiple simultaneous clients when we have a real need for it.
2018-01-24 19:23:20 +10:00
Peter Hutterer cce63d267b dbus: stop discovery when the searching client disappears
This is the single-client version only, we can deal with multiple clients
later.

Fixes #21
2018-01-24 19:23:20 +10:00
Peter Hutterer 71dd70cc95 README: fix the API documentation, Pairing* is now Search* 2018-01-24 19:23:20 +10:00
Peter Hutterer daf2693927 Add setup.py
Fixes #17
2018-01-24 19:09:35 +10:00
Peter Hutterer eb5efd2e1c tuhi: move everything to base.py
Let tuhi.py just be the script that calls main. This way we're somewhat
setup.py compatible.
2018-01-24 19:08:12 +10:00
Benjamin Tissoires 6fa14b65f2 kete: Move the file in a tools folder
This is an unbelievable commit. This just moves one file in a new dir,
and this fixes bug #19.
2018-01-23 19:42:23 +10:00
Peter Hutterer 5e32dc4872 dbus: don't update self.paired unless it changes
TuhDevice.paired is set on every device update (RSSI changes!) and that sends
a GObject.notify() for the property. Filter those, otherwise
we're just spamming dbus with PropertiesChanged notify events even
though nothing has actually changed.
2018-01-23 14:57:00 +10:00
Peter Hutterer 7aa4314c3d wacom: use .format instead of hex to generate the timestring
hex() skips leading zeroes, so 180123 turns into 18, 12, 3
2018-01-23 14:44:01 +10:00
Peter Hutterer 790f2337dc kete: allow pairing on existing devices 2018-01-23 14:22:33 +10:00
Peter Hutterer d9459ef3f9 kete: quit after a successful pairing 2018-01-23 14:22:33 +10:00
Peter Hutterer 08ab047d0a tuhi: set pairing to False on known devices too 2018-01-23 14:22:33 +10:00
Peter Hutterer fc23247ffa kete: handle the 'searching' property correctly in the manager 2018-01-23 14:22:33 +10:00
Peter Hutterer efbae8c50c kete: add a basic comment to fetch drawings
All it does is print information about the drawings, but it has the data, so
converting to svgs in $PWD is doable now.
2018-01-23 14:22:33 +10:00
Peter Hutterer 415ecd39a5 kete: move the MainLoop into the manager so we can watch the bus name
And the basic mainloop code is the same everywhere anyway
2018-01-23 14:22:33 +10:00
Peter Hutterer 44a2d5e8eb ble: if we connect twice, log the error as debug
We likely get multiple 'udpated' notifications as the RSSI property changes,
all causing a Connect() on the device and an ugly error message (that we used
to catch and print). Make that error message prettier.
2018-01-23 14:22:33 +10:00
Peter Hutterer f9ba9fe6ca ble: don't crash if a device doesn't have a Name attribute
Hello, tradies next door with your weird phones...
2018-01-23 14:22:33 +10:00
Peter Hutterer 63ba0462ea dbus: send out PropertiesChanged when we update the drawings 2018-01-23 14:22:33 +10:00
Benjamin Tissoires ca82af78de tuhi: handle cold-plugged devices
When bluez restarts (or the tuhi daemon restarts), the values we have
in the bluez device's field ManufacturerData are quite not accurate.

When bluez restarts they are empty, and if the last time we saw the
device was for the pairing process, the device will still be marked
as in the pairing mode.

So we should mark the cold-plug sequence differently from the hot-plug
one, and we should be more confident in the current configuration we
have stored to export the currently known devices over dbus.

Fixes #13
2018-01-23 14:22:33 +10:00
Benjamin Tissoires ea890958d6 Start discovery mode when one device requests it
We need to check when the discovery stops (timeout from StartSearch)
if we should keep the discover one or not.
2018-01-23 14:22:33 +10:00
Peter Hutterer 45196fbdca kete: print when drawings are available 2018-01-23 14:22:33 +10:00
Peter Hutterer 59ab21a5e8 kete: add the 'listen' command
Doesn't do anything but start/stop listening but hey, that's what it says on
the box
2018-01-23 14:22:33 +10:00
Benjamin Tissoires a51663342f dbus: keep track of the senders of the StartListening() method
If the sender disappear, we should stop listening for incoming events.

We match the Start/StopListening() that way, but if a client forgets
to call StopListening() before leaving and some data are being retrieved,
it's not our problem.
2018-01-23 14:22:33 +10:00
Benjamin Tissoires e3fff4015a implement a basic Start/StopListening
This implements the ListeningStopped signal (especially the EAGAIN if we're
already listening) but does not yet actually trigger the listening on the
device.

There is still no timeout, and no signal gets emitted.
The current way of testing this is:
- call StartListening() on the DBus device
- start discovery on the adapter by some other mean
- press the button on the device -> the sync will happen
- call StopDiscovery()
- press the button on the device -> the sync will *not* happen
2018-01-23 14:22:33 +10:00
Benjamin Tissoires 0e6d82dd14 README: document the new listening process
We need to be able to selectively enable/disable listening mode per
device, and the Listening property needs to be updated.

The current approach uses a Listening property as a public state, and the
StartListening()/StopListening() calls to start/stop. The ListeningStopped
signal is just there to carry error codes.

The current design only allows for a single client to trigger listening,
anyone else will see -EAGAIN and has to wait for the property to update.
2018-01-23 14:22:33 +10:00
Peter Hutterer 010c651d6f dbus: add a __repr__ for the TuhiDBusDevice
And remove a now obsolete debugging statement, this was just put there to
debug the PropertyChanged code
2018-01-23 14:22:33 +10:00
Peter Hutterer e5d3da5590 kete: default to 'list' if we don't have a subcommand 2018-01-23 09:05:17 +10:00
Peter Hutterer fc81356d25 kete: don't return self from __exit__
This supresses any exceptions we get during the handling. Just continue and
let the stack deal with the exception (if any)
2018-01-23 09:03:10 +10:00
Peter Hutterer e974ca17cc kete: use a Devices property changed as indicator for success
Remember the devices we are pairing and subscribe to the Manager.Devices
property changed notification. If a device moves from just existing to be part
of Manager.Devices it has been successfully paired.
2018-01-22 17:14:37 +10:00
Peter Hutterer d3192dc070 dbus: Send out a PropertyChanged event when a device is paired
Just updating the array isn't enough, we need to bubble up the property
change and then manually send the PropertyChanged event
2018-01-22 17:14:37 +10:00
Peter Hutterer daac6e6d96 dbus: drop the "paired" flag from create_device
We can get this from the device now
2018-01-22 17:14:37 +10:00
Peter Hutterer ec87503209 tuhi: keep a reference to the Tuhi dbus device in the TuhiDevice
Because we need to notify it when we have paired so the daemon can update
2018-01-22 17:14:37 +10:00
Peter Hutterer 645c7577fe config: use uppercase 'UUID' in debug message 2018-01-22 17:14:37 +10:00
Peter Hutterer e21f04bb8b Add tuhi-kete as a CLI interface to tuhi 2018-01-22 17:14:33 +10:00
Peter Hutterer a2fd3cd8d1 dbus: on StartSearch(), emit PairableDevice signals for already known devices
If we created a device before StartSearch was called, we need to manually
send the signal to make sure the client sees it.
2018-01-22 15:25:34 +10:00
Peter Hutterer 0b7f756ba0 dbus: add the device's Address to the introspection XML
Code was already there, but the XML was missing the address
2018-01-22 15:24:56 +10:00
Benjamin Tissoires 56dc8741dc defer the creation of the wacom device object
If bluez is restarted, the services are not resolved.
If the data has not been connected, the manufacturerData might be
correctly set, but the services are still not resolved, meaning that
we can not start listening on the various GATT characteristics.

Defer the creation of the wacom device object after connect, when
we are sure the services are resolved.
2018-01-22 15:24:56 +10:00
Benjamin Tissoires 3e5ef6c939 ble: remove the duplicate events filter
So we get more chances of capturing the advertisement from the device
if it is rather static.
2018-01-22 15:24:56 +10:00
Benjamin Tissoires 380518d0c7 ble: keep track of Start/StopDiscovery
discovery-stopped may call stop_discovery() if there
is no client listening, resulting in an infinite loop
2018-01-22 15:24:26 +10:00
Benjamin Tissoires f4e35df67d tuhi: merge _on_bluez_device_added with _on_bluez_device_updated
This allows to create the device even if the conditions were not met
while starting up tuhi (if bluez considers the device still in
pairing mode).

This will also allow us to retrieve the button presses from the devices,
as they are triggering a BLE advertisement.
2018-01-22 15:23:05 +10:00
Benjamin Tissoires 36b9b1cbb7 README: add note about the sensor orientation
The coordinates are provided from the top-left position, but the sensor
is effectively turned. So on the Spark and the Slate, X goes down, and
Y is reversed and goes left.
2018-01-20 12:37:50 +10:00
Benjamin Tissoires f6d09d7086 dbus: fix timeout error if the index is invalid
If the index is not valid, a python exception was raised, and the dbus
message was left without and answer.

Coincidentally, this matches the documentation
2018-01-20 12:37:50 +10:00
Benjamin Tissoires 1b73de68b2 config: make sure we create the config file properly
I guess this code was tested with an existing config file only, because
I have the following:

DEBUG: tuhi.config: E0:61:C6:BF:14:4E: adding new config, uuid 8bbc6be4347a
Traceback (most recent call last):
  File "./tuhi.py", line 153, in _on_uuid_updated
    self.config.new_device(bluez_device.address, wacom_device.uuid)
  File "tuhi/tuhi/config.py", line 89, in new_device
    config['Device']['Address'] = address
  File "/usr/lib64/python3.6/configparser.py", line 959, in __getitem__
    raise KeyError(key)
KeyError: 'Device'
2018-01-20 12:37:50 +10:00
Benjamin Tissoires fda5d7f132 update the paired property on uuid change
When the uuid changes, it means the device just has been paired.
We should therefore update the paired property or a subsequent connect
to retrieve the data will result in an attempt to pair to the device.
2018-01-20 12:37:50 +10:00
Benjamin Tissoires c79ffbd12c do not create more than one DBus object per bluetooth device
When the device is in pairing mode, it will send several properties
updates, and tuhi will attempt to create a second DBus object.

This is not good.
2018-01-20 12:37:50 +10:00
Peter Hutterer 4ee07a18b6 Hook up UUID changes to be written in the config file
Make the uuid property a GObject.Property so we can listen to it. Notify about
the uuid change at the end of the pairing process, then write that value into
the device's config file.
2018-01-19 16:17:44 +10:00
Peter Hutterer 0deb5f8e49 Add an extra log message when it's time to call Pair()
Otherwise it looks like we just sit here and do nothing
2018-01-19 16:17:44 +10:00
Peter Hutterer 8e50421f2e wacom: generate the uuid on pairing
Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2018-01-19 16:17:44 +10:00