[DrunkOnJava]

Summary

This PR contains #398's commit (f540e7e, merged into the branch) and must land after #398. Once #398 is on main its diff drops out of this one.

Re-measures the requirements of docs/release-1.0-criteria.md that a green main run can re-measure, and changes only what that run proves. One status flips: X1 (DXF) goes from not measured to failed. Every other status is unchanged. R-numbering and wording are otherwise intact.

Spec reference

None; no code. Refs #195 (the epic; it stays open) and #169, #381.

Evidence: CI run 40542129639, the ci.yml run on main at 61255e9, every job success, artifact corpus-measurement (corpus-results.json, oracle-summary.md, oracle.tsv, dxf-header-probe.log, dxf-tables-probe.log, dxf-entities-probe.log, dxf-blocks-probe.log, header-probe.log). The R3 counts come from docs/data/spec-coverage.tsv at that commit.

Test plan

• Each number below was recomputed from the artifact files by a script that reads them (corpus totals and per-release file counts from corpus-results.json; scored sections, disagreements and unscored records from oracle-summary.md; probe totals from the four dxf-*-probe.log files; state counts from the TSV), not copied from earlier prose.
• python3 tools/ci/spec_coverage.py --check: clean on this branch with docs(spec-coverage): move the rows the 2026-10-27 main run proves, and say what measured them (#140, #147, #148, #133, #127) #398 merged in (this PR adds no change to the page or the TSV beyond docs(spec-coverage): move the rows the 2026-10-27 main run proves, and say what measured them (#140, #147, #148, #133, #127) #398's).
• No code and no build.

What changed, row by row (before, after, evidence):

Row	Status	Before	After	Evidence
header	n/a	"Last measured 2026-10-21, b268e884"	2026-10-27, 61255e9, run linked	the run
R2	failed, unchanged	R14 to R2018 open and walk on at least four drawings per release	same, plus 52 drawings, 15,541 records decoded, 0 errored, with the run linked	corpus-results.json: per release 8, 8, 8, 8, 8, 8 and 4 drawings (R14 to R2018)
R3	failed, unchanged	0 of 138 complete; 34 missing, 31 partial, 6 decoder, 67 closes	0 of 138; 28 missing, 7 partial, 6 decoder, 97 closes	the TSV with #398 (138 rows excluding the n/a chapter 1). On main alone it is 29, 13, 6 and 90; #398's seven moves to closes and §22 to partial are proven by run 40532346172, the run #398 cites
R4	failed, unchanged	46 sections scored, 19 values disagree (#241-#244)	53 sections scored, 22 values disagree (#241-#244, #381)	oracle-summary.md: 53 spec sections with records; differences BLOCK 3, LTYPE 4, VIEW_CONTROL 6, APPID_CONTROL 4, VPORT 2, MULTILEADER 3. MULTILEADER was scored by #385; its 3 differences are #381.
R8	failed, unchanged	308 records error honestly (#226, #227, #228, #71)	0 records error, 15,541 decode, 533 are skipped for want of a decoder, each named by kind	corpus-results.json: errored 0, errors empty on every drawing, 28 skipped kinds. The status stays failed: the AppInfo misreading (#137) is still open and the run cannot show otherwise, so that clause is kept.
X1	not measured -> failed	"Written for R12 to R2018; checked only syntactically"	the HEADER, TABLES, ENTITIES and BLOCKS probes pair every written record with AutoCAD's export by handle: on 42 drawings 0 groups disagree and none is extra (HEADER 9,583 variables, 9,410 agree and 173 hold a stated relation; TABLES 1,239 records, 8,831 of 8,840 groups agree and 9 by relation; ENTITIES 803 records, 5,627 of 5,627 groups; BLOCKS 1,532 records, 12,577 of 12,583 groups and 6 by relation). Not equal yet: 342 ENTITIES records of 18 kinds are not written, and the OBJECTS section is not written on main	the four dxf-*-probe.log files. The bar is "equals AutoCAD's own DXF of the same drawing"; the doc defines failed as measured and not met, and not measured as no evidence either way, so the old status no longer fits.
S3	failed, unchanged	"no 32-bit job"	the library compiles for wasm32-unknown-unknown in the Clippy job (step "The library builds where usize is 32 bits"), but no job runs tests or a corpus walk on a 32-bit target	the run's Clippy job steps

Rows I did not change, because the run does not move them: R1 (STATUS.md's other numbers are not shown to come from a run), R5 (VISUALSTYLE 1,008 records, 58,104 equal, 264 different, the view styles and MATERIAL still unscored: the run's numbers equal the page's), R6 (42 nextgis and 10 ACadSharp drawings in pins.tsv, matches), R7, W1 to W3, X2 to X5, S1, S2, S4 and P1 (no evidence in a main run either way). "No recorded fuzz durations" in S3 is kept for the same reason.

R3 and #398. The R3 counts include #398 (the spec-coverage refresh): it moves seven rows from partial to closes and §22 from missing to partial. This branch has #398's commit merged in (f540e7e, merge commit 27bc149), so python3 tools/ci/spec_coverage.py --check passes here (spec-coverage: 139 rows, 135 required sections, 0 problems) and R3 matches the TSV the branch carries. It must land after #398; the other rows are measured from run 40542129639 and do not depend on it.

Source-provenance declaration

By submitting this PR I certify that my contribution does not incorporate executable code from any source whose license is incompatible with Apache-2.0. Specifically:

• No Autodesk DWG SDK source (RealDWG, ObjectARX, ObjectDBX) or NDA-protected Autodesk documentation was consulted or imported.
• No Open Design Alliance Teigha / Drawings SDK source was consulted or imported. The ODA's public Open Design Specification for .dwg files PDF is explicitly allowed (see CLEANROOM.md).
• No GPL-licensed DWG implementation source (LibreDWG, etc.) was consulted or imported. This matters because dwg-rs is Apache-2.0 and must remain compatible.

My contribution is based on:

• The ODA's freely-redistributable Open Design Specification for .dwg files (v5.4.1);
• Public reverse-engineering notes, research papers, or blog posts cited inline;
• Reading algorithm-description comments (not executable code) in permissively-licensed projects (MIT / Apache / BSD) to resolve spec ambiguities — disclose any such reading in this PR body so it can be recorded in CLEANROOM.md.

See CLEANROOM.md for the full allowed/forbidden source list and the honest scope of what "clean-room" means for this project.

Developer Certificate of Origin

By submitting this PR, I certify that the contribution complies with the
Developer Certificate of Origin 1.1.
Specifically, I affirm that the contribution is my own work (or rightfully
licensed to me) and that it can be distributed under the project's Apache-2.0
license.

I will sign off commits with git commit -s when asked, which adds a
Signed-off-by trailer to the commit message asserting the DCO.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	🔄 Running since 2026-10-28T16:52:31.982595Z	76bd093	New commits

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.