Closed rootulp closed 5 months ago
Note to reviewers: we could add a unit test to this PR that includes a bad_block.json
from a failing TestMaxBlockSize test case
We could merge this as-is but I want @cmwaters 's review. I'm also nervous about regressions in this repo because this fix was added without a test that would fail if we ever revert this behavior 😞 I'll work on adding a test now.
Closing in favor of https://github.com/celestiaorg/go-square/pull/49
Closes https://github.com/celestiaorg/go-square/issues/47
This restores the behavior from celestia-app v1.x where square size upper bound is used to calculate worstCaseShareIndexes.
Testing
In celestia-app, TestMaxBlockSize flake is no longer observed