chore(derive): [Holocene] Drain previous channel in one iteration

[clabby]

Overview

Drains the entire previous channel at one time, rather than over multiple iterations, if the following case is met:

If a first frame is decoded while the previous frame isn't a last frame (i.e., is_last is false), all previous frames for the same channel are dropped and this new first frame remains in the queue.

Also fixes a small issue w/ the DERIVED_FRAMES_COUNT metric's accounting, where it was counting dropped frames.

Tests

Added a new test for this functionality, when there is a valid, closed channel earlier on in the frame queue @ test_holocene_replace_channel.

[clabby]

• chore(derive): [Holocene] Drain previous channel in one iteration #583 👈
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.

Join @clabby and the rest of your teammates on Graphite

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 79.3%. Comparing base (d5e7038) to head (f826bfd).
Report is 1 commits behind head on main.

✅ All tests successful. No failed tests found.

Additional details and impacted files

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

[refcell]

The logic makes sense to me and I like the short variable rename to prev_frame and next_frame - much clearer.

I would like to verify that it passes tests that cover the sad path ruleset where frames with differing channels are interleaved in #585.

Maybe we can merge #585 and then rebase this on-top?