[PATCH BlueZ 0/3] Promote the bearer interfaces to stable

[Kevin Dewald <kevin@simpleble.org>]

As offered in the Role series, this marks org.bluez.Bearer.LE1 and
org.bluez.Bearer.BREDR1 stable, now that test/functional/test_bearer.py
covers them.

The bearer interfaces are the only way to connect a given bearer of a
dual-mode device. Device1.Connect picks one, by default the last used,
so an application that only uses LE, such as a GATT library, can end
up with only BR/EDR connected, and Device1.Connected does not say which
bearer is up. Today such an application has to ask its users to run
bluetoothd with --experimental.

The series drops the [experimental] tags from the documentation, the
experimental flags from src/bearer.c, and Experimental from the bearer
tests. The Disconnected signal was only marked experimental in the
documentation.

The bearer tests pass without Experimental in three consecutive runs,
and the rest of test-functional still passes.

test_bearer_le_connect can fail in CI when
test_bass_past_transport_acquire runs just before it on the same VMs.
That is a kernel bug, unrelated to the bearer interfaces, fixed by:
https://lore.kernel.org/linux-bluetooth/20261023191437.87024-1-kevin@simpleble.org/

Kevin Dewald (3):
  org.bluez.Bearer: Promote to stable
  bearer: Promote to stable
  test: functional: run bearer tests without experimental

 doc/functional-bearer.rst      |  7 +++----
 doc/org.bluez.Bearer.BREDR.rst | 32 +++++++++++++++---------------
 doc/org.bluez.Bearer.LE.rst    | 36 +++++++++++++++++-----------------
 src/bearer.c                   | 23 ++++++++--------------
 test/functional/test_bearer.py |  8 +++-----
 5 files changed, 48 insertions(+), 58 deletions(-)


base-commit: 7bc9684dc9e45358661301da4a0ede433fdf0a30
-- 
2.54.0 (Apple Git-157)