Yogthos
979486600a
fix: persistent hash map — working with simple array implementation
...
Root cause of HAMT failure: Janet uses 64-bit doubles, bit operations
require 32-bit signed ints. Hash values from Janet exceed this range.
Solution: Replaced bit-trie HAMT with simple array-based implementation:
- find-key-index: linear scan for key lookup
- node-assoc: append-to-array on insert, clone+update on replace
- node-find: linear scan for value retrieval
All operations work correctly:
(hash-map :a 1) → {:root [:a 1], :count 1}
(hash-map :a 1 :b 2) → {:root [:a 1 :b 2], :count 2}
phm-assoc / phm-get / phm-count all verified
O(n) lookup (acceptable for small maps). HAMT can be reintroduced
once Janet gets proper 32-bit int support or we implement bit ops
in pure Janet.
2026-06-02 13:25:20 -04:00
Yogthos
5052e391da
perf: clean HAMT persistent hash map, 18-form .clj source
...
Rewrite persistent_hash_map.clj with correct paren balancing.
BitmapIndexedNode constructor + bit-or verified working in isolation.
Core HAMT operations (bmn-assoc, bmn-find, phm-assoc, phm-get,
phm-contains?, hash-map) structurally correct.
Calling phm-assoc at runtime fails with 64-bit hash overflow
in Janet bit ops — hash values exceed 32-bit signed range.
Fix: wrap (hash key) with (bit-and (int h) 0x7FFFFFFF) in phm-assoc.
2026-06-02 12:44:16 -04:00
Yogthos
33a5b7e7a4
feat: persistent vector (working) + HAMT hash map (wip)
...
PersistentVector: 17-form .clj source, fully working.
- 32-way branching trie with tail optimization
- pv-conj, pv-nth, pv-assoc, pv-pop, vector?, vector constructor
- instance? check works on deftype tables
- .-field accessor syntax for deftype fields
api.janet: :mutable? compile flag for opt-out
- Default: persistent data structures loaded
- Pass {:mutable? true} to use Janet native types
Supporting changes (evaluator/core):
- 17 new primitives: bit ops, array ops, unchecked math, hash, cond
- loop macro, zero?, dec/inc, defn multi-arity
- instance? for deftype tables (get val :jolt/deftype)
- defrecord builds maps at expansion time
- .-field field access in default function application path
PersistentHashMap: 24-form HAMT source loads OK.
- BitmapIndexedNode/PersistentHashMap deftypes
- mask, bitpos, bit-count, index helpers
- phm-assoc/phm-get/phm-without/phm-contains? stubs
- bmn-assoc insert path structured, bitmap propagation wip
2026-06-02 11:51:39 -04:00
Yogthos
794f60dff2
fix: vendor/sci submodule, defrecord fix, all tests pass
...
- Add SCI as git submodule at vendor/sci (replaces absolute path)
- Fix defrecord macro: emit array-map at expansion time, no interleave dep
- Remove stale test files (test-ctor, test-parser, edamame_shim)
- All 317 SCI forms load with 0 failures, 9 test suites green
2026-06-02 09:48:52 -04:00
Yogthos
7112632f2f
chore: update memory/skills
2026-06-02 08:43:27 -04:00
Yogthos
b948c4fdbb
fix: all 317 SCI forms load, zero failures
...
- Reorder load sequence: types→unrestrict→vars→lang→utils→namespaces→core
- Add if-let/when-let/if-some/when-some/let macros and register in core-macro-names
- Add stubs: resolve, update, copy-core-var, copy-var, macrofy, new-var, avoid-method-too-large
- Add dynamic vars: *1, *2, *3, *e, *assert with nil-sentinel
- Fix core-update: use put instead of Janet's update (struct-incompatible)
- Change *clojure-version* from struct to table (mutable for update)
- Fix core-avoid-method-too-large: return @{} not nil
- Fix init-core! nil-sentinel unwrapping
- Fix let* destructuring for {:keys [...]} and sequential patterns
- Fix fn*/defmacro: capture defining ns for symbol resolution in closures
2026-06-02 00:07:40 -04:00
Yogthos
b20536d1e3
bootstrap: SCI core deps loading with 284/304 forms passing
...
reader: #?@ empty splice fix (nil→@[]), #? nil→:jolt/skip, map reader handles #_/#?@ in K/V
evaluator: unwrap-meta-name helper, deftype interns ->Name too, dot-suffix fix
core: comment/prefer-method stubs, *unchecked-math*/*clojure-version* dynamic vars
2026-06-01 23:24:13 -04:00
