Made panda use pva signal backend instead of pvi_get

[rosesyrett]

Make use of PVASignalBackend standard Backend

[DiamondJoseph]

Still need, but do not need to get the initial state, just get the pvi.

[DiamondJoseph]

@coretl: just checking that I am going in the right direction with this change? I've gotten to the point that the new code looks correct as far as I understand, but the pva fixture isn't playing nicely (timing out tests, not letting the test suite complete). It looks like INCOMPLETE_BLOCK_RECORD, INCOMPLETE_RECORD ,EXTRA_BLOCKS_RECORD don't actually exist in tests/panda/db?

[DiamondJoseph]

Also: is there a standard for "FOO" + ":BAR" vs. "FOO:" + "BAR:" ? The latter seems to be more common in dodal, so I've been using it, assuming the trailing colon won't cause issue.

I see in Rose's documentation change she also enforced that prefix should end with : if it isn't already
https://github.com/bluesky/ophyd-async/pull/107/files#diff-fbf444aec05ace8f7baf32ccb7fe37cbb7bb988607c665bb235f89d06dfa1c49

[coretl]

Also: is there a standard for "FOO" + ":BAR" vs. "FOO:" + "BAR:" ? The latter seems to be more common in dodal, so I've been using it, assuming the trailing colon won't cause issue.

I see in Rose's documentation change she also enforced that prefix should end with : if it isn't already https://github.com/bluesky/ophyd-async/pull/107/files#diff-fbf444aec05ace8f7baf32ccb7fe37cbb7bb988607c665bb235f89d06dfa1c49

DLS convention used to be no colon in the prefix, then colon added in the database. This only work because DLS PVs look like BLxxI-MY-DEVICE-01:RECORD, it doesn't work for all facilities that might use a different separator. Much better is to include the colon in the prefix which is what Rose's change did

[DiamondJoseph]

Test for getting the PVI info as a PVADict currently times out. I don't know enough about Epics DB Records to diagnose

[DiamondJoseph]

Tom should Approve