Make sure initial values are respected in CA transport

[jsouter]

Closes #202

Proof of concept, it would be nicer if we could avoid the need for a system test/use of cagets as these can be flaky.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.32%. Comparing base (9990065) to head (8167dfd).
⚠️ Report is 9 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #257      +/-   ##
==========================================
+ Coverage   89.03%   89.32%   +0.29%     
==========================================
  Files          46       46              
  Lines        2280     2268      -12     
==========================================
- Hits         2030     2026       -4     
+ Misses        250      242       -8     

☔ 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.

[GDYendell]

I think this is good and testing through caget is reasonable. See if running fastcs.serve as a task (like here) works and makes it less flaky.

[jsouter]

I think this is good and testing through caget is reasonable. See if running fastcs.serve as a task (like here) works and makes it less flaky.

Looks like the problem in doing this with a CA transport is that we get the error

AssertionError: Record database has already been loaded

[jsouter]

I've changed the test to look directly at the record wrapper instead of doing cagets to test initial values. To create_task on fastcs.serve we need to use pytest-forked to run the test in a new process circumvent the issue with the softioc.builder being invalidated in later tests, which would restrict that test to running only on linux.

[GDYendell]

This looks really neat! I don't mind it spying record set vs doing caget, but it looks like the way you have done it is probably easier than caget?

[jsouter]

Yeah I think it's easier this way. Will try and fix up the pyright issues