This allows defining session independent locks with the DEVICE_LOCKS …

[MarcusZuber]

…(list of strings) variable

In a session file you can add e.g. DEVICE_LOCKS = ['my_camera1', 'reco_1_server'] to prevent the usage of devices that should be accessed only once and adding the same entry in another session file.

This exends the MULTIINSTANCE implementation. I can now specify sets of locks in session files.
The use case is, that we often use the same devices (which are somethimes not happy to be accessed multiple times) from diffrent sessions.
I can now do stuff like this:

# session 1 that uses a specific camera and a remote-online-reco
DEVICE_LOCKS = ['camera1', 'reco_server1']

# session 2 that uses 'camera1' that does not like to be shared
DEVICE_LOCKS=['camera1']

TODO:

• Add documentation
• Add tests (also for the MULTIINSTANCE implementation)

[codecov]

Codecov Report

❌ Patch coverage is 45.34606% with 229 lines in your changes missing coverage. Please review.
✅ Project coverage is 82.59%. Comparing base (f30a00b) to head (e983ad0).

Files with missing lines	Patch %	Lines
concert/session/utils.py	30.69%	228 Missing ⚠️
concert/tests/integration/test_session_locks.py	98.70%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #552      +/-   ##
==========================================
- Coverage   83.69%   82.59%   -1.10%     
==========================================
  Files         133      134       +1     
  Lines       10973    11390     +417     
==========================================
+ Hits         9184     9408     +224     
- Misses       1789     1982     +193     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[MarcusZuber]

@tfarago : Do you have an idea how we could test the session-locks elegantly without system calls?

[tfarago]

Looks useful.

[tfarago]

@tfarago : Do you have an idea how we could test the session-locks elegantly without system calls?

Write a mockup StartCommand or make a mockup of IPython.

[MarcusZuber]

Some explanation about e983ad0 :

concert._aimoirt.register() could only be called once before this commit (the paths were stored in a global variable that is not reset between tests and it does nothing if there are aready patht sepcified). The testing of the session-start triggered this the first time in the pytest run and then the register() in the aimport-test did nothing.