Repository navigation
FIX: Run the known-issues cases with twins, and classify same-call twin flips as divergent - #31
Merged
Merged
Conversation
Every known-issues case now runs its five 1-ULP twins, so the study can tell XFOIL's own ill-conditioning from a translation difference. Publish run 2026-10-23T19-39-42Z and correct the parting section of the known-issues index, which said the runs had no twins.
…l as divergent The study classifier counted a twin's branch flip only from an earlier call. In naca64a010 N60 TYPE 2 at alpha 0 the twins flip ITRAN at iterations 11 and 12 of call 1 and then stop at 14 or 15, so iteration 16 has no recorded spread and the run's parting there read as a mismatch. A parting at or after a twin's flip in the same call is now divergent. Republish the known-issues and branch-coverage runs, the only rows it changes.
…reference The known-issues fixtures and their twins are regenerated on the reference of main (its fingerprint changed with the host-libm stage), and the study rerun from them reproduces every table and figure byte for byte; only the run's provenance changes, now a clean commit.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
group = "known-issues"now run their five seeded 1-ULP twins (twins = true), so the study can tell XFOIL's own ill-conditioning from a translation difference. Updated the notes that list which groups carry twins: CLAUDE.md,testing.mdand the fixtures README. Republished from the rerun, and corrected the known-issues index, which said these runs have no twins.studies/study-support/records.rs). A twin's branch flip used to count only if it was in an earlier call. Innaca64a010N60 TYPE 2 at α 0, the twins flip ITRAN at iterations 11 and 12 of call 1 and then stop at iteration 14 or 15. Iteration 16 therefore had no recorded spread, and the parting there read as a mismatch. A parting at or after a twin's flip in the same call is now divergent.Results
naca64a010_n60_type2_a0_re1e6, call 1, iteration 16, goes from MISMATCH to divergent. No other row changes. Its non-finite table picks up the correcteddocs/validation/branch-coverage/path.Evidence the reclassified point is not a translation difference. yFoil and XFOIL agree through iteration 11. The twins flip ITRAN at 11 and 12, and yFoil's RMSBL moves by 1.35e-4 at 12, inside the twins' spread of 0.25. From then on, the relative RMSBL difference stays constant at about 2e-3 through iterations 14, 15 and 16: the iteration-12 difference decaying with RMSBL. Nothing new appears at iteration 16. Only the twins' record ends there.
Test plan
cargo fmt --checkandcargo clippy --all-targets --all-features -- -D warningsscripts/docs.sh build --clean, docs check clean