fdfd086df6
jolt no longer satisfies :clj in reader conditionals. The shortcut was a measured net liability: :clj branches carry JVM interop and JVM-specific test expectations jolt fails, and they shadowed :default branches jolt passes. A/B over the suite: clj,default = 4967 assertions / 4324 pass / 119 errors; jolt,default = 5069 / 4470 / 81 (+146 pass, -38 errors, +8 clean files). Baselines raised to 4470/86. Matching is now by CLAUSE order like Clojure — the first clause whose key is in the feature set wins (#?(:default 5 :clj 6) is 5 everywhere); the old code scanned for :clj first, then :default, regardless of position. Foreign clj-targeted libraries are a property of the LOADING CONTEXT, not the platform: reader-features-set! opts a load into a compatibility set, and the SCI bootstrap/runtime tests load SCI under ["jolt" "clj" "default"] (its .cljc selects implementations via :clj with no :jolt branches). JOLT_FEATURES remains the process-wide override. RFC 0002 records the decision with the measured data; spec 02-reader S18 is now normative (clause order, documented feature set, per-context override). Reader tests updated to the portable set + an opt-in round-trip. |
||
|---|---|---|
| .. | ||
| 00-front-matter.md | ||
| 02-reader.md | ||
| 03-special-forms.md | ||
| 09-core-library.md | ||
| coverage.md | ||
| README.md | ||
The Clojure Language Specification (Draft)
A normative, implementation-independent specification of the Clojure
language, developed alongside jolt's self-hosted compiler and validated by
its executable conformance suites. Why: Clojure has no spec — every
alternative implementation re-derives semantics from the reference
implementation and folklore. See the RFC for motivation, scope, evidence
sources, and process: ../rfc/0001-language-specification.md.
Documents
| Doc | Content | Status |
|---|---|---|
00-front-matter.md |
conformance terms, entry format, host classification | drafted |
01-evaluation.md … 08-macros.md |
see chapter plan in front matter | planned |
03-special-forms.md |
special-form catalog + normative exemplars (if, let*) |
exemplars |
09-core-library.md |
per-var entry format + exemplars (first, reduce, parse-uuid) |
exemplars |
coverage.md |
generated dashboard over the 694-var surface | generated |
Regenerate the dashboard after surface changes:
python3 tools/spec_coverage.py (requires clojuredocs-export.json in the
repo root and a working jolt checkout).
Current numbers (2026-06-10)
Of the 694 clojure.core vars in the ClojureDocs inventory:
- 380 implemented in jolt and exercised by the behavioral suites
- 154 implemented but not directly tested — each gets a test with its spec entry
- 35 portable but missing from jolt (
parse-long/parse-double/parse-boolean,update-keys/update-vals,macroexpand,time,partitionv/partitionv-all/splitv-at,with-redefs,with-open, reader fns, ns-introspection stragglers, …) — tracked as implementation gaps - 22 resolvable in code but invisible to ns introspection
(
resolve/ns-publicscan't see seed-fallback names likecompare,gensym,type) — a conformance finding in its own right - the rest classified host/JVM/concurrency (see dashboard)
How this connects to the test suites
test/integration/conformance-test.janet— 302 assertions, each run through three independent execution paths (interpreter, bootstrap compiler, self-hosted compiler) that must agree. Spec entries cite these.test/spec/*.janet— ~1,500 behavioral cases organized by topic.vendor/clojure-test-suite— the cross-dialect suite (≥4081 assertions passing); dialect splits there are classification evidence.- jank's per-construct corpus (
~/src/jank/compiler+runtime/test/jank) is the granularity model for §2/§3 conformance.
The invariant: every numbered normative statement names its conformance test, or is marked UNVERIFIED. The spec cannot drift from the implementations that check it.