Repository navigation
Fix Windows connection issues - #56
Conversation
|
@odedtap Thanks for digging into this — I agree there is a real Windows regression after the Bleak 3 upgrade, but I think we can preserve the old behavior more elegantly and avoid relying on Bleak's private backend. Before Bleak 3, the Windows path bypassed Suggested shape:
Could you confirm the hardware repro for both cases before we slim this down?
If the first case still fails quickly with |
Mostly related to the order of scanning/retrieving devices.
Docs check
|
This will properly release GATT objects on script exit.
|
Workflow run: https://github.com/TapWithUs/tap-python-sdk/actions/runs/35230943339 Docs check
|
|
Workflow run: https://github.com/TapWithUs/tap-python-sdk/actions/runs/35231779903 Docs check
|
|
Workflow run: https://github.com/TapWithUs/tap-python-sdk/actions/runs/35232556995 Docs check
|
|
Workflow run: https://github.com/TapWithUs/tap-python-sdk/actions/runs/36563353049 Docs check
|
Keep both Unreleased notes. Fix flake8 E203 in the Windows address parser. Document the v2 NUS API and the async context manager. Co-authored-by: Liron Ilouz <ilouzl@users.noreply.github.com>
NUS communication implementation to be done using inheritence with a private SDK class.
|
Workflow run: https://github.com/TapWithUs/tap-python-sdk/actions/runs/38714497727 Docs check
|
Commit db1cd9e deleted write_nus_command and register_nus_raw_data_events. The reference pages, release notes, AGENTS.md, and the raw-sensors skill still described them. Co-authored-by: Liron Ilouz <ilouzl@users.noreply.github.com>
Summary
Test plan