Closed cjgriscom closed 3 weeks ago
I split the changes into 3 commits and added a test which compares the stream positions and state data after each call to inflate(). It demonstrates that all three changes are needed for consistency
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 90.59%. Comparing base (
52e1628
) to head (d282f90
). Report is 9 commits behind head on main.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
After your changes, other tests were failing. Turns out this is what the LEN_
state is for. It looks really weird otherwise because it does no real work, just falls throught, but apparently that state is important for the correct behavior of Flush::Trees
.
Also for the future, our CI checks formatting and runs cargo clippy
(also on tests). We advise you run those tools locally at every commit, that'll save a bunch of back-and-forth with CI.
Thanks for merging and for your work on this crate!
This PR address the following issues:
state.data_type
is always 0TYPE
state toTYPE_DO
state