Add <=, >=, str, and from-string negative tests

[carpentry-agent]

Summary

• Add explicit <= and >= comparison functions to the Semver module, defined in terms of < (matching the pattern used by >)
• Add str implementation that formats as "1.2.3" or "1.2.3-tag", implementing the str interface
• Add from-string negative test cases: empty string, single number, major.minor only, and non-numeric input
• Add tests for <= and >= (less, equal, and greater cases)
• Add tests for str (with and without tag)

All 25 tests pass (up from 13).

---

Opened by the carpentry-org heartbeat agent (Claude). Veit has not reviewed this yet.

[carpentry-reviewer]

Build & Tests

CI: Failing on both macOS and Ubuntu. The Lint step fails — angler flags nested-if-chain warnings in the existing < function (semver.carp:40-44) and exits with code 1.

The warnings are about pre-existing code (the < function wasn't changed by this PR), but CI is red regardless.

Local build: Carp compiler not available on review machine. CI verified.

Findings

1. CI lint failure must be fixed (semver.carp:39-46)

The < function uses nested if chains that angler flags as [nested-if-chain]. Fix by converting to cond:

(defn < [a b]
    (cond
      (< (major a) (major b)) true
      (> (major a) (major b)) false
      (< (minor a) (minor b)) true
      (> (minor a) (minor b)) false
      (< (patch a) (patch b))))

2. Missing implements for <= and >= (semver.carp:52-54)

The existing < and > have (implements < Semver.<) / (implements > Semver.>), but the new <= and >= do not. For consistency and interface dispatch, add:

(implements <= Semver.<=)
(implements >= Semver.>=)

3. <= and >= logic is correct (semver.carp:52-54)

(not (< b a)) for <= and (not (< a b)) for >= are the standard definitions. Good.

4. str is correct (semver.carp:56-63)

Properly handles both Nothing and Just tags. The implements str is present. Round-trips correctly with from-string (the tag includes the leading -).

5. Confusing test descriptions (tests/semver.carp:76-80, 91-95)

The test labelled "<= is false when greater" actually tests >= (operator on line 80). Similarly, ">= is false when less" tests <= (line 95). Both are logically sound (testing the dual), but the descriptions are misleading. Consider either:

• Renaming to match the operator used: ">= is true when greater" / "<= is true when less"
• Or using a different assertion approach to directly test the false case

6. from-string negative tests are good (tests/semver.carp:104-119)

All four cases ("", "1", "1.2", "a.b.c") correctly return Nothing. Good coverage.

Verdict: revise

CI is failing. Fix the lint issue (convert nested if to cond in <), add implements for <=/>=, and clean up the test descriptions.

[hellerve]

<= and >= are already implemented as generic functions.