Fixing the occasional errors in tests

[evalott100]

A couple of tests fail occasionally:

• test_create_softioc_system sometimes cagets SEQ.TABLE fields part way through the panda-side update of SEQ.TABLE. To fix this, the caget has been moved to the start of the test and more time has been added before SEQ.TABLE is updated in the mocked panda.
• test_table_column_info would fail because
  • the test process was too fast: pva would try to fetch values before panda introspection
  • the mocked panda process was too fast: by the time the pva context had been updated the mocked panda would have altered its table
  • To fix the former we used a pva monitor instead of a caget, timing out at TIMEOUT, to fix the latter we used a different mocked panda which doesn't update its tables. We've also changed to the asyncio p4p client.
• During testing for the above two tests, another test failed: test_non_defined_seq_table_can_be_added_to_panda_side this was because very rarely the panda subprocess was too fast and had already changed SEQ3 to have values.
  • To fix this we'll give the panda 50 updates before it changes the value of SEQ3 rather than 10.

Tested locally with

for i in {1..50}; do echo $i; pytest | grep failed; done

to ensure no failures,

tested on the CI 10 times.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (eba34d5) 86.29% compared to head (532d964) 86.29%.

Additional details and impacted files

@@           Coverage Diff           @@
##             main      #63   +/-   ##
=======================================
  Coverage   86.29%   86.29%           
=======================================
  Files           7        7           
  Lines        1138     1138           
  Branches      187      187           
=======================================
  Hits          982      982           
  Misses        121      121           
  Partials       35       35           

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

[AlexanderWells-diamond]

I think there's better ways to do these fixes

[evalott100]

Strange, somehow the table packing and unpacking have found their way back into this repo... I will look into this.

[evalott100]

For some reason TablePacking had been reintroduced. My last push removed it again (we moved them to the client pandablocks.utils which is where all the ioc code currently gets table_to_words and words_to_table from. I'm unsure of where they were reintroduced from.

[AlexanderWells-diamond]

As best as I can work out, TablePacking was never deleted in any PR that was merged.

We stopped using it in #15 , which was merged.

However, there also exists both #14 and #17 , both of which did delete the code but neither of them were merged. There's no notes to say why that may be.

It would be clearer if we don't delete that code in this PR - it's an entirely unrelated change to fixing the tests.
(And while we're on the subject of neatness, can you please rewrite your commits to conform to the standard 50-character header line and 72 character body line. There's a decent commit message guide here)

[evalott100]

@AlexanderWells-diamond

I've cherry-picked the TablePacking to another branch and formatted the git commits.

Docs is failing due to the warning

WARNING: The default value for `navigation_with_keys` will change to `False` in the next release. If you wish to preserve the old behavior for your site, set `navigation_with_keys=True` in the `html_theme_options` dict in your `conf.py` file.Be aware that `navigation_with_keys = True` has negative accessibility implications:https://github.com/pydata/pydata-sphinx-theme/issues/1492

this is occurring on main too. We should fix this on another branch (pr here).

[AlexanderWells-diamond]

Looks good to me.