Chez Phase 1 (increment 3f): quoted literals
Emit a :quote node by reconstructing the raw reader form as RT constructor calls: symbol -> jolt-symbol, list (array) -> jolt-list, vector (tuple) -> jolt-vector, map -> jolt-hash-map, set -> jolt-hash-set, scalars via emit-const. The runtime value of a quote is just that literal data (the interpreter returns the reader form verbatim). Quote exposed a latent seq.ss bug: empty map/filter results were jolt-nil, but Clojure's (map f []) is an empty seq, so (= () (map f [])) must be true. Return jolt-empty-list (which seqs back to nil, so it's still a valid lazy-tail terminator) instead — matching jolt-take/drop/rest/list. Prelude emit reach 334 -> 342/355. Subset probe 632 -> 664/664 compiled, 0 divergences (quote + the seq fix pull 32 corpus cases into the subset). emit-test 110/110 (added 16 quote cases). corpus.edn regenerated (the 3 malformed-catch spec rows). Full gate green.
This commit is contained in:
parent
d35783bf1b
commit
930800f9a6
6 changed files with 81 additions and 19 deletions
|
|
@ -138,6 +138,35 @@
|
|||
(string "(integer->char " (get v :ch) ")")
|
||||
(errorf "emit-const: unsupported literal %p" v)))
|
||||
|
||||
# Quoted literals (jolt-u8j7). A :quote node's :form is the RAW reader form (a
|
||||
# Janet value): scalars are Janet natives, a symbol is {:jolt/type :symbol …}, a
|
||||
# list is an array, a vector a tuple, a map a struct/phm, a set a tagged struct.
|
||||
# Reconstruct each as the matching Chez RT constructor — the runtime value of a
|
||||
# quote is just that literal data (the interpreter returns the reader form
|
||||
# verbatim; the Janet backend Janet-quotes it; here we rebuild it on the RT).
|
||||
(var emit-quoted nil)
|
||||
(defn- emit-quoted-map [m]
|
||||
(def flat @[])
|
||||
(eachp [k v] m (array/push flat (emit-quoted k)) (array/push flat (emit-quoted v)))
|
||||
(string "(jolt-hash-map " (string/join flat " ") ")"))
|
||||
(set emit-quoted (fn emit-quoted [form]
|
||||
(cond
|
||||
# scalars emit-const already lowers (nil/bool/number/string/keyword/char)
|
||||
(or (nil? form) (boolean? form) (number? form) (string? form) (keyword? form))
|
||||
(emit-const form)
|
||||
(and (struct? form) (= :symbol (get form :jolt/type)))
|
||||
(let [ns (get form :ns)]
|
||||
(string "(jolt-symbol " (if ns (string/format "%j" ns) "#f") " "
|
||||
(string/format "%j" (get form :name)) ")"))
|
||||
(and (struct? form) (= :jolt/char (get form :jolt/type))) (emit-const form)
|
||||
(and (struct? form) (= :jolt/set (get form :jolt/type)))
|
||||
(string "(jolt-hash-set " (string/join (map emit-quoted (get form :value)) " ") ")")
|
||||
(array? form) (string "(jolt-list " (string/join (map emit-quoted form) " ") ")")
|
||||
(tuple? form) (string "(jolt-vector " (string/join (map emit-quoted form) " ") ")")
|
||||
(phm/phm? form) (emit-quoted-map (phm/phm-to-struct form))
|
||||
(or (struct? form) (table? form)) (emit-quoted-map form)
|
||||
(errorf "emit-quoted: unsupported quoted form %p" form))))
|
||||
|
||||
(defn- emit-binding [b]
|
||||
(def b (vv b))
|
||||
(string "(" (munge (get b 0)) " " (emit (get b 1)) ")"))
|
||||
|
|
@ -331,6 +360,7 @@
|
|||
:recur (emit-recur node)
|
||||
:throw (string "(jolt-throw " (emit (get node :expr)) ")")
|
||||
:try (emit-try node)
|
||||
:quote (emit-quoted (get node :form))
|
||||
:fn (emit-fn node)
|
||||
:def (string "(def-var! " (string/format "%j" (get node :ns)) " "
|
||||
(string/format "%j" (get node :name)) " " (emit (get node :init)) ")")
|
||||
|
|
|
|||
|
|
@ -107,11 +107,15 @@
|
|||
;; map / filter / reduce / into / remove + range / take / concat / apply
|
||||
;; ============================================================================
|
||||
(define (any-nil? seqs) (and (pair? seqs) (or (jolt-nil? (car seqs)) (any-nil? (cdr seqs)))))
|
||||
;; An EMPTY seq result is () (jolt-empty-list), NOT nil — Clojure's (map f []) is
|
||||
;; an empty seq, so (= () (map f [])) is true and (nil? (map f [])) is false.
|
||||
;; jolt-empty-list seqs back to nil, so it stays a valid lazy-tail terminator for
|
||||
;; the non-empty case (printing / seq= / reduce all walk via jolt-seq).
|
||||
(define (map-seq f s)
|
||||
(if (jolt-nil? s) jolt-nil
|
||||
(if (jolt-nil? s) jolt-empty-list
|
||||
(cseq-lazy (jolt-invoke f (seq-first s)) (lambda () (map-seq f (jolt-seq (seq-more s)))))))
|
||||
(define (map-seq* f seqs) ; multi-collection map; stops at the shortest
|
||||
(if (any-nil? seqs) jolt-nil
|
||||
(if (any-nil? seqs) jolt-empty-list
|
||||
(cseq-lazy (apply jolt-invoke f (map seq-first seqs))
|
||||
(lambda () (map-seq* f (map (lambda (s) (jolt-seq (seq-more s))) seqs))))))
|
||||
(define (jolt-map f . colls)
|
||||
|
|
@ -121,7 +125,7 @@
|
|||
|
||||
(define (filter-seq pred s keep)
|
||||
(let loop ((s s))
|
||||
(cond ((jolt-nil? s) jolt-nil)
|
||||
(cond ((jolt-nil? s) jolt-empty-list) ; empty result is () (see map-seq)
|
||||
((eq? keep (jolt-truthy? (jolt-invoke pred (seq-first s))))
|
||||
(cseq-lazy (seq-first s) (lambda () (filter-seq pred (jolt-seq (seq-more s)) keep))))
|
||||
(else (loop (jolt-seq (seq-more s)))))))
|
||||
|
|
|
|||
|
|
@ -48,12 +48,13 @@ compile-time signal) and are counted "out of subset", not as divergences.
|
|||
|
||||
JOLT_CHEZ_CORPUS=1 janet test/chez/run-corpus-chez.janet
|
||||
|
||||
Baseline after inc 3e (throw/try + ex-info): **632/632 compiled cases pass**, 0
|
||||
divergences; 2023/2655 out of subset (await clojure.core on Chez). Inc 3c
|
||||
(multi-arity + variadic fns) brought this to 619/619 — `emit-fn` lowers
|
||||
multi-arity fns to a Scheme `case-lambda` and variadic fns to a rest-arg lambda
|
||||
(the Scheme rest list is coerced to a jolt seq, nil when empty); inc 3e's
|
||||
throw/try/ex-info support pulled 13 more corpus cases into the subset.
|
||||
Baseline after inc 3f (quoted literals): **664/664 compiled cases pass**, 0
|
||||
divergences; 1994/2658 out of subset (await clojure.core on Chez). Inc 3e
|
||||
(throw/try + ex-info) was 632/632; inc 3f's quote support + a seq.ss fix (empty
|
||||
`map`/`filter` results are `()` not nil, matching Clojure) pulled 32 more corpus
|
||||
cases into the subset. `emit-fn` lowers multi-arity fns to a Scheme `case-lambda`
|
||||
and variadic fns to a rest-arg lambda (rest list coerced to a jolt seq, nil when
|
||||
empty).
|
||||
|
||||
## Phase 1 — clojure.core prelude emission (inc 3d, jolt-ocvi)
|
||||
The `-e`-capable jolt-chez path: emit the clojure.core tiers
|
||||
|
|
@ -70,15 +71,16 @@ catalogs the gaps; macros are skipped (analyze-time only, not a runtime value):
|
|||
|
||||
JOLT_CHEZ_PRELUDE=1 janet test/chez/core-prelude-probe.janet
|
||||
|
||||
Baseline after inc 3e (throw/try + ex-info): **334/355 non-macro core forms emit**
|
||||
to Scheme (was 303 at inc 3d). `:throw` lowers to `jolt-throw` (Scheme `raise` of
|
||||
the raw jolt value, like the Janet compiled back end); `:try` lowers to `guard`
|
||||
(catch, class dropped → catch-all) + `dynamic-wind` (finally); `ex-info` is a
|
||||
native-op building the tagged jolt map, so the ex-data/ex-message/ex-cause tier
|
||||
fns read it over `get` for free. Remaining 21 gaps: `:quote` (8, needs RT
|
||||
symbols/lists), Java host interop `.write`/`.isDirectory` (6, io tier), `letfn`
|
||||
(4), `declare`/def-no-init (2), one edge (`parse-uuid`). Each is a clean
|
||||
next-increment target. The probe has a regression floor (raise it as gaps close).
|
||||
Baseline after inc 3f (quoted literals): **342/355 non-macro core forms emit** to
|
||||
Scheme (was 334 at inc 3e, 303 at inc 3d). A `:quote` node reconstructs the raw
|
||||
reader form as RT constructor calls — symbol → `jolt-symbol`, list → `jolt-list`,
|
||||
vector → `jolt-vector`, map/set → `jolt-hash-map`/`jolt-hash-set`, scalars via
|
||||
`emit-const`. Prior: `:throw` → `jolt-throw` (Scheme `raise` of the raw jolt
|
||||
value); `:try` → `guard` (catch, class dropped → catch-all) + `dynamic-wind`
|
||||
(finally); `ex-info` native-op building the tagged jolt map, so the tier
|
||||
ex-data/ex-message/ex-cause read it over `get` for free. Remaining 13 gaps: Java
|
||||
host interop `.write`/`.isDirectory` (6, io tier), `letfn` (4), `declare`/
|
||||
def-no-init (2), one edge (`parse-uuid`). The probe has a regression floor.
|
||||
|
||||
Prior, inc 3b (seq tier + dynamic IFn, jolt-5pso): 595/595 compiled, 0 divergences,
|
||||
2060/2655 out of subset. The seq tier brought up a list/lazy-seq type with
|
||||
|
|
|
|||
|
|
@ -98,7 +98,7 @@
|
|||
|
||||
# Regression floor (raise it as new IR ops / RT shims land, like the suite
|
||||
# baseline). Fails if prelude emit reach drops below the recorded baseline.
|
||||
(def reach-floor 334)
|
||||
(def reach-floor 342)
|
||||
(when (< compiled reach-floor)
|
||||
(printf "REGRESSION: prelude emit reach %d < floor %d" compiled reach-floor)
|
||||
(os/exit 1))
|
||||
|
|
|
|||
|
|
@ -152,6 +152,9 @@
|
|||
{:suite "exceptions / try-catch" :label "finally runs on ok" :expected "2" :actual "(let [a (atom 0)] (try 2 (finally (reset! a 9))) )"}
|
||||
{:suite "exceptions / try-catch" :label "finally runs on throw" :expected "9" :actual "(let [a (atom 0)] (try (throw (ex-info \"x\" {})) (catch :default e nil) (finally (reset! a 9))) @a)"}
|
||||
{:suite "exceptions / try-catch" :label "catch value of body" :expected "5" :actual "(try (+ 2 3) (catch :default e 0))"}
|
||||
{:suite "exceptions / try-catch" :label "malformed catch: non-symbol binding" :expected :throws :actual "(try 1 (catch e (* e 10)))"}
|
||||
{:suite "exceptions / try-catch" :label "malformed catch: literal binding" :expected :throws :actual "(try 1 (catch e 5))"}
|
||||
{:suite "exceptions / try-catch" :label "malformed catch: too short" :expected :throws :actual "(try 1 (catch Exception))"}
|
||||
{:suite "exceptions / assert" :label "assert true -> ok" :expected ":ok" :actual "(do (assert true) :ok)"}
|
||||
{:suite "exceptions / assert" :label "assert expr -> ok" :expected ":ok" :actual "(do (assert (= 1 1)) :ok)"}
|
||||
{:suite "exceptions / assert" :label "assert false throws" :expected :throws :actual "(assert false)"}
|
||||
|
|
|
|||
|
|
@ -236,6 +236,29 @@
|
|||
(ok "throw: uncaught exits non-zero" (not= code 0)
|
||||
(string "code=" code " out=" out)))
|
||||
|
||||
# 3j) quoted literals (inc 3f): a :quote node reconstructs the reader form as RT
|
||||
# values — symbols, lists, vectors, maps, sets, nested. Value parity vs the CLI.
|
||||
(each src ["'foo"
|
||||
"'foo/bar"
|
||||
"':kw"
|
||||
"'(1 2 3)"
|
||||
"'[1 2 3]"
|
||||
"'(a b c)"
|
||||
"'{:a 1}"
|
||||
"'(1 (2 3) 4)"
|
||||
"(first '(10 20 30))"
|
||||
"(count '[1 2 3])"
|
||||
"(rest '(1 2 3))"
|
||||
"(= 'foo 'foo)"
|
||||
"(= 'a 'b)"
|
||||
"(map inc '(1 2 3))"
|
||||
"(conj '[1 2] 3)"
|
||||
"(get '{:a 7} :a)"]
|
||||
(let [[code out err] (d/run-on-chez ctx src)
|
||||
want (cli-oracle src)]
|
||||
(ok (string "quote: " src) (and (= code 0) (= out want))
|
||||
(string "chez=" out " janet=" want " | " err))))
|
||||
|
||||
# 3h) prelude mode (inc 3d): emitting clojure.core ITSELF, a core->core ref must
|
||||
# lower to a runtime var-deref instead of being rejected as "out of subset".
|
||||
# `frequencies` is a core fn but not a native-op, so it exercises the switch.
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue