test: verify Code.tag() returns correct tags for all codecs

[sumanjeet0012]

Closes #42.

What was changed:

• Added a parametrized test test_code_tag_all_categories to tests/test_code.py that verifies the tags of 22 codecs covering all known tags.
• Added test_no_unknown_tags() to verify that no codec in CODECS returns "<unknown>" as its tag.
• Fixed multiple unmapped tags in multicodec/code.py (like provenance-log, aes-128, etc.) which caused test_no_unknown_tags to fail, ensuring all 603 codecs have correct recognized tags.

Why the change was needed:

• The existing tests only checked 4 tags out of 23. This PR increases tag coverage to all categories.
• Ensures all registered codecs have correct tag metadata.

How it was verified:

• Verified by running make pr which invokes pytest. All tests pass successfully, confirming that Code.tag() properly maps all codecs in multicodec/constants.py.

[acul71]

AI PR Review: #50 — test: verify Code.tag() returns correct tags for all codecs

Reviewer: acul71 (maintainer)
PR: #50
Author: sumanjeet0012
Date: 2026-09-07

---

1. Summary of Changes

This PR addresses Closes #42: tag tests previously covered only 4 codecs / 4 tags. It adds broader category coverage and ensures no registered codec returns "<unknown>".

Changes:

• Expands _get_tag() mappings in multicodec/code.py (same six tags as #49, plus many additional name matches across key/libp2p/multiaddr/multiformat/multihash/namespace/serialization/transport)
• Adds inline assertions for the six previously missing tags in test_code_tag
• Adds parametrized test_code_tag_all_categories (22 cases covering known tag categories)
• Adds test_no_unknown_tags() iterating CODECS

Modules/files: multicodec/code.py, tests/test_code.py
Breaking changes / deprecations: None (behavior change: fewer "<unknown>" tags — desirable)

---

2. Branch Sync Status and Merge Conflicts

Branch Sync Status

• Status: Ahead of origin/master
• Details: 0 1 → 0 behind, 1 ahead

Merge Conflict Analysis

✅ No merge conflicts detected. The PR branch can be merged cleanly into origin/master.

Overlap with PR #49

Both PRs modify the same files. PR #50 fully subsumes PR #49’s six-tag additions (multikey, multisig, nonce, shelter, softhash, vlad) and goes further with additional mappings required for test_no_unknown_tags to pass.

Recommendation: Prefer merging #50 (after newsfragment) and closing or superseding #49, rather than merging both sequentially. If both must land, merge #49 first then rebase #50 (expect conflicts in code.py / test_code.py).

---

3. Strengths

• Implements the issue’s proposed parametrized coverage + no-unknown invariant
• Fixes the root cause of "<unknown>" tags beyond the six categories from #37/#49
• Correct multisig ordering before multihash
• Local validation: lint, typecheck, 712 tests passed, docs build OK; GitHub CI green

---

4. Issues Found

Critical

• File: newsfragments/ (missing)
• Issue: Missing mandatory newsfragment for issue #42
• Suggestion: Add newsfragments/42.misc.rst or 42.bugfix.rst describing that Code.tag() coverage/tests ensure all registered codecs map to recognized tags

Major

• File: tests/test_code.py

• Line(s): TAG_TEST_CASES / test_code_tag

• Issue: Duplication: the six new tags are tested both inline in test_code_tag and again via parametrize (plus #49 overlap)

• Suggestion: Keep parametrized + test_no_unknown_tags as the source of truth; drop redundant inline asserts to reduce maintenance cost

• File: multicodec/code.py

• Line(s): Multiformat list includes "multisig" / "multikey" while later branches also return those tags for other names

• Issue: Easy to misread; the multiformat entries are for codecs literally named multisig/multikey, while chacha20-poly1305 / *-msig use dedicated branches — this matches Go-style dual use but deserves a short comment

• Suggestion: Add a one-line comment clarifying codec-name vs category-tag matching

Minor

• File: tests/test_code.py
• Issue: Parametrized list still samples one codec per category rather than “all 603 codecs” as the issue title suggests; test_no_unknown_tags closes that gap for unknown detection but does not assert the expected tag string per codec against an external oracle
• Suggestion: Acceptable for #42’s proposed solution; fuller oracle checks belong with #51’s table.csv work if tags are present upstream

---

5. Security Review

No security concerns. Classification-only changes over static codec tables.

---

6. Documentation and Examples

No API docs required beyond newsfragment for user-visible tag corrections.

---

7. Newsfragment Requirement

⚠️ BLOCKER

• Severity: CRITICAL / BLOCKER
• Issue: No newsfragments/42.*.rst; PR correctly references Closes #42
• Suggestion: Add newsfragments/42.bugfix.rst (or .misc.rst) with user-facing note that Code.tag() now returns recognized tags for all registered codecs / expanded categories
• Action Required: Newsfragment before approval

---

8. Tests and Validation

Linting (make lint)

• Exit 0 — all hooks passed

Type Checking (make typecheck)

• Exit 0 — mypy passed

Test Execution (make test)

• Exit 0 — 712 passed (up from 689 on master/#49), 0 failed

Documentation Build (make docs-ci)

• Exit 0 — Sphinx -W succeeded

GitHub CI

• All tox/windows checks SUCCESS

---

9. Recommendations for Improvement

1. Add newsfragment 42.*.rst (blocker)
2. Coordinate with #49: recommend supersede #49 with this PR (or close #49 after #50 merges)
3. Deduplicate inline vs parametrized tag assertions
4. Optional comment on multiformat vs category multikey/multisig names

---

10. Questions for the Author

1. Should #49 be closed as superseded by this PR?
2. Preferred newsfragment type: bugfix vs misc?
3. Were the additional name mappings (aes-*, provenance-log, silverpine, etc.) cross-checked against go-multicodec Tag()?

---

11. Overall Assessment

• Quality Rating: Good
• Security Impact: None
• Merge Readiness: Needs fixes (newsfragment; clarify relationship to #49)
• Confidence: High

Verdict: REQUEST_CHANGES — strong improvement that subsumes #49 and addresses #42, blocked only on newsfragment (and process clarity vs #49).