real chunked seqs for vector seqs

A vector's seq is now a real chunked-seq (chunked-seq? true), matching Clojure/
CLJS. Each vector-seq cell carries its backing vector + element index as two
cseq fields (cvec/ci, no extra allocation vs the old lazy cell), so:
  - chunk-first hands out a 32-element block (a pvec slice), chunk-rest is the
    seq at the next block boundary — the ChunkedSeq contract (chunk-first ++
    chunk-rest == the seq);
  - reduce/transduce take a fast path that walks the backing vector by index in
    a tight loop with no per-element seq cells (reduce over a 1M-vector ~0.4s).
The seq cell stays a cseq, so first/rest/count/printing and the ~26 cseq?
dispatch sites are untouched. The eager chunk-buffer model (chunk-buffer/chunk/
chunk-cons) is preserved for the round-trip case. No seed change (runtime only).
This commit is contained in:
Yogthos 2026-06-22 00:21:05 -04:00
parent 0e4ccc97e0
commit 3721423a64
4 changed files with 56 additions and 14 deletions

View file

@ -94,3 +94,6 @@
(def-var! "clojure.core" "rational?" jolt-rational?)
(def-var! "clojure.core" "decimal?" jolt-decimal?)
(def-var! "clojure.core" "==" jolt-num-equiv)
;; chunked-seq? is true for a vector's seq (a real chunked-seq); the overlay's
;; always-false stub loaded over the host fn, so re-assert it (jolt-hs5q).
(def-var! "clojure.core" "chunked-seq?" na-chunked-seq?)