Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 59.26%. Comparing base (656523e) to head (c26090b).
Report is 82 commits behind head on master.

@@            Coverage Diff             @@
##           master    #5863      +/-   ##
==========================================
- Coverage   59.27%   59.26%   -0.01%     
==========================================
  Files         353      357       +4     
  Lines       29694    29771      +77     
==========================================
+ Hits        17601    17645      +44     
- Misses      11113    11153      +40     
+ Partials      980      973       -7     

I could cleanup the scripts to always run with an experimental flag or we could keep them there in case we want to have a split between experimental and normal runs? Instead of removing the non-experimental target from the GitHub workflow, I kept it instead but commented out.

For the daemon, I think we won't be adding new features gated behind experimental - we largely came to the conclusion that a single "experimental" flags was too much of a big-hammer, and instead we now started to put individual options behind options in features (in daemon.json) where we don't want them to be enabled by default. In all cases (including existing ones), experimental options were adding to existing options, so they would always be a superset of what's tested without experimental. From that perspective, I think it's fine if we just have a single option, enable experimental by default in our tests, and remove distinction between "experimental" and "non-experimental".

For tests that test specific experimental features, we should keep the t.Skip though, but those skips should be based on what the daemon returns (i.e., if the daemon indicates it has experimental disabled, then skip the test).

I think I got all the occurrences of experimental-e2e. Please let me know if you know of more places that I might have missed 🙏