Update bonding v2 merged with main and tested.

[CristianMG]

This PR includes the following key changes:

Merged with main branch
Replaced the value of NOT_EXPORTED with EXPORTED.
Motivation: Although the Android developer documentation doesn't mention this explicitly, through my testing and by finding examples in other codebases, I realized that this is necessary for proper functionality.

For example, the Nordic Semiconductor BLE library includes this in their implementation: Example from Nordic Semiconductor BLE Library

Executing on the main thread: Broadcasting should be registered and executed on the main thread, as this is a requirement for correct operation.

Another thing, added waitForBonding to connect constructor function I think that's can be usefull because you can control bonding at the connect which in some cases is a must for all operations. Maybe including in the main call can be little aggresive to the api be free to indicate another place to put this functionality.

I will work on make a little piece of software to execute in raspberry or similar to be able to test by your part. I'll keep on touch with you through this pull request.

[twyatt]

Thanks for this! It is great to have a branch that is tested and known working.

I've also updated #494 with the changes that will be going into the upcoming Kable release. Although I have not been able to test my branch (I'm very excited for being able to test against custom Arduino or Raspberry Pi code though, when that becomes available).

[CristianMG]

Hi @twyatt ,

I’ve just finished developing a Python script for Raspberry and an example app in Jetpack Compose, both aimed at testing the pairing process using the library you created. You can find the script at this link: kable-ble-test.

In particular, in "Just Works" mode on devices running Android 13 and later, pairing is always requested due to a change in the Bluetooth stack. However, in Android 12 and earlier, the device connects without requiring authentication, and when a feature that requires bonding is executed, it initiates the authentication process. I’ve also observed that if bonding has already been performed on Android 12 and the keys are then deleted from the Android device, bonding is requested again during the next connection attempt.

This raises doubts for me about the waitBonding part in the connect method, as it seems there is no predictable behavior across different Android versions. Perhaps it would be better to remove this part and revert to the original implementation, which waits on the read or write methods.

I’ve been observing these behaviors for a while, as I have extensive experience working with Android and the BLE stack, which seems to have quite a few bugs in the latest versions due to the stack refactoring.

I’d love to hear if you’ve received similar reports from other users or if you have any recommendations on how to handle these variations. I truly appreciate your work and hope to contribute to the community with my findings.

Thank you!

Image of python script log:

Image of example application:

Bonding process:

[twyatt]

Wow, thanks so much! This is great!
I'll try to find time this weekend to spin up the Raspberry Pi script and try it out.

This is exactly what I needed to unblock progress on bonding support.

I'll be in touch if I run into any issues or have any questions.

Thanks again, this is so incredibly helpful!

[CristianMG]

I made some changes in requirements.txt, could you try again?

[twyatt]

I made some changes in requirements.txt, could you try again?

Thanks for the update, @CristianMG!

I was able to get everything setup per your instructions (and some additional dependency installations):

$ pip install virtualenv
$ virtualenv kable-env
$ source kable-env/bin/activate

# https://stackoverflow.com/a/62510708
$ sudo apt install build-essential libdbus-glib-1-dev libgirepository1.0-dev
$ pip install dbus-python

# https://stackoverflow.com/a/23649012
$ sudo apt-get install libcairo2-dev libjpeg-dev libgif-dev

$ pip install -r requirements.txt

Unfortunately, now it is failing with:

$ python cputemp.py

Traceback (most recent call last):
  File "/home/travis/kable-ble-test/cputemp.py", line 27, in <module>
    from ble_agent import Agent, register_agent
  File "/home/travis/kable-ble-test/ble_agent.py", line 4, in <module>
    from gi.repository import GLib
  File "/home/travis/kable-ble-test/kable-env/lib/python3.11/site-packages/gi/__init__.py", line 39
    print url
    ^^^^^^^^^
SyntaxError: Missing parentheses in call to 'print'. Did you mean print(...)?

Perhaps I'm using the wrong Python version?

$ python --version

Python 3.11.2

[CristianMG]

I hate that error I can't fix properly it's seems related with this libraries
PyGObject==3.50.0
gi==1.2
Sometimes I simply fix playing with the version of this libraries and re-installing after requirerements

pip install PyGObject==3.48.2
I know that´s weird as you say some people says it´s related with the interpreter version. If you can´t fix I will investigate further if this error is related with version interpreter

[twyatt]

Thanks for the tips! Got it working.

I ended up getting it running by deleting the gi=1.2 line from the requirements.txt file and re-spinning up the virtual environment.

I'm assuming it worked because I have a system version of the gi package installed that is Python3 compatible, whereas perhaps gi 1.2 is Python 2 only?

[CristianMG]

Maybe, but in my experience if you delete all gi references threw error I mean delete from pi, that´s the reason I think this error is weird. Whatever if this script become a standard to test the library would improve with more test cases and eventually I´ll fix this bug the reason that I have to involved in the improvement or fixing library errors is that the company I work use this library in some projects related with hardware.

[twyatt]

@CristianMG I'm going to continue the conversation in #111.