Skip to content

release: 0.5.2 — infer marker-fallback fix - #30

Merged
scottvr merged 1 commit into
mainfrom
release/0.5.2
Aug 24, 2026
Merged

release: 0.5.2 — infer marker-fallback fix#30
scottvr merged 1 commit into
mainfrom
release/0.5.2

Conversation

@scottvr

@scottvr scottvr commented Aug 24, 2026

Copy link
Copy Markdown
Owner

Patch release for a real infer bug that shipped in 0.5.0 and 0.5.1.

The bug: a spurious byte-autocorrelation peak (arithmetic payloads can autocorrelate strongly at a wrong lag) could shadow a sparse-marker record structure, so infer picked the wrong stride, found no constant columns, and gave up as "variable-length text" — never trying the sparse-marker detector.

The fix (already on main via #29): when the autocorrelation stride has no constant columns, fall back to the marker detector before concluding it is text.

  • __version__ 0.5.1 → 0.5.2; CHANGELOG [0.5.2]. PATCH (bug fix).
  • Preflight: vizbin --version → 0.5.2; wheel/sdist metadata → 0.5.2; twine check PASSED; ruff + mypy clean; 164 tests green.

After merge: git tag v0.5.2 && git push origin v0.5.2 → gated publish → approval.

🤖 Generated with Claude Code

Patch release for the infer stride-detection bug present in 0.5.0/0.5.1: a
spurious byte-autocorrelation peak could shadow a sparse-marker record structure,
making infer give up as 'variable-length text'. Now it falls back to the marker
detector. Bump __version__ to 0.5.2; roll CHANGELOG [Unreleased] into [0.5.2].
@scottvr
scottvr merged commit d25045f into main Aug 24, 2026
6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant