Updated the matrix per the review: Linux now runs 3.10, 3.11, 3.12, 3.13, and 3.14; Windows and macOS run only 3.10 and 3.13.

Verification run locally:

/tmp/actionlint/actionlint .github/workflows/build.yaml
git diff --check
python - <<'PY'
import yaml
from pathlib import Path
workflow = yaml.safe_load(Path('.github/workflows/build.yaml').read_text())
include = workflow['jobs']['build']['strategy']['matrix']['include']
actual = [(item['os'], item['python_version'], item['tox_env']) for item in include]
expected = [
    ('ubuntu-latest', '3.10', 'py310'),
    ('ubuntu-latest', '3.11', 'py311'),
    ('ubuntu-latest', '3.12', 'py312'),
    ('ubuntu-latest', '3.13', 'py313'),
    ('ubuntu-latest', '3.14', 'py314'),
    ('windows-latest', '3.10', 'py310'),
    ('windows-latest', '3.13', 'py313'),
    ('macos-latest', '3.10', 'py310'),
    ('macos-latest', '3.13', 'py313'),
]
if actual != expected:
    raise SystemExit(f'matrix mismatch: {actual!r}')
print('matrix OK:', actual)
PY

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 81.46%. Comparing base (bffb749) to head (232f8c4).
⚠️ Report is 116 commits behind head on main.

@@            Coverage Diff             @@
##             main     #999      +/-   ##
==========================================
- Coverage   86.63%   81.46%   -5.17%     
==========================================
  Files          84      139      +55     
  Lines        4473    13525    +9052     
==========================================
+ Hits         3875    11018    +7143     
- Misses        598     2507    +1909     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

Updated this PR after the Windows matrix exposed a cross-platform test assertion issue.

What changed:

• Rebased the branch onto current dapr/python-sdk:main to clear the stale/behind Codecov comparison.
• Made the sync/async test_wait_timeout assertions check for ConnectionRefusedError instead of the Linux-specific message text Connection refused, which does not appear in Windows' socket error text.

Local verification:

• coverage run -m unittest discover -q ./tests ✅ (Ran 975 tests in 65.016s)
• python -m unittest -v tests.clients.test_dapr_grpc_client.DaprGrpcClientTests.test_wait_timeout tests.clients.test_dapr_grpc_client_secure.DaprSecureGrpcClientTests.test_wait_timeout tests.clients.test_dapr_grpc_client_async.DaprGrpcClientAsyncTests.test_wait_timeout tests.clients.test_dapr_grpc_client_async_secure.DaprSecureGrpcClientAsyncTests.test_wait_timeout ✅
• ruff check tests/clients/test_dapr_grpc_client.py tests/clients/test_dapr_grpc_client_async.py ✅
• /tmp/actionlint/actionlint .github/workflows/build.yaml ✅
• git diff --check ✅

The new GitHub Actions runs for commit 2d13963639cd8fe729e42477b4bbabd287ad4a56 are currently action_required, so they need maintainer approval before they can execute.

Updated this PR to address the Windows failures from the latest CI run.

What changed:

• The sync/async test_wait_timeout assertions now accept either ConnectionRefusedError or TimeoutError. The Windows runners were raising TimeoutError('timed out') for the same unused-port wait path, while Linux raises connection-refused.

Local verification:

• python -m pip install -q -r dev-requirements.txt -e . ✅
• python -m unittest -v tests.clients.test_dapr_grpc_client.DaprGrpcClientTests.test_wait_timeout tests.clients.test_dapr_grpc_client_secure.DaprSecureGrpcClientTests.test_wait_timeout tests.clients.test_dapr_grpc_client_async.DaprGrpcClientAsyncTests.test_wait_timeout tests.clients.test_dapr_grpc_client_async_secure.DaprSecureGrpcClientAsyncTests.test_wait_timeout ✅
• ruff check tests/clients/test_dapr_grpc_client.py tests/clients/test_dapr_grpc_client_async.py ✅
• git diff --check ✅

Note: the earlier validate (3.11) failure was in tests/examples/test_pubsub_streaming_async.py missing the final Closing subscription... output after Dapr shut down; that looks separate from this matrix/test assertion change.