lua: minimal Lua pattern engine for string.find (classes/anchors/quantifiers)

lib/lua/runtime.sx

@@ -696,6 +696,163 @@
 (dict-set! coroutine "wrap" lua-coroutine-wrap)
 
 ;; ── string library ────────────────────────────────────────────
+
+
+;; ── Lua patterns (minimal) ────────────────────────────────────
+;; Predicates for single-char classes
+(define
+  lua-pat-class-match
+  (fn (class c)
+    (cond
+      ((= class "d") (and (>= c "0") (<= c "9")))
+      ((= class "D") (not (and (>= c "0") (<= c "9"))))
+      ((= class "a") (or (and (>= c "a") (<= c "z")) (and (>= c "A") (<= c "Z"))))
+      ((= class "A") (not (or (and (>= c "a") (<= c "z")) (and (>= c "A") (<= c "Z")))))
+      ((= class "s") (or (= c " ") (= c "\t") (= c "\n") (= c "\r")))
+      ((= class "S") (not (or (= c " ") (= c "\t") (= c "\n") (= c "\r"))))
+      ((= class "w")
+        (or (and (>= c "a") (<= c "z")) (and (>= c "A") (<= c "Z")) (and (>= c "0") (<= c "9"))))
+      ((= class "W")
+        (not (or (and (>= c "a") (<= c "z")) (and (>= c "A") (<= c "Z")) (and (>= c "0") (<= c "9")))))
+      ((= class "l") (and (>= c "a") (<= c "z")))
+      ((= class "L") (not (and (>= c "a") (<= c "z"))))
+      ((= class "u") (and (>= c "A") (<= c "Z")))
+      ((= class "U") (not (and (>= c "A") (<= c "Z"))))
+      ((= class "p") (or (and (>= c "!") (<= c "/")) (and (>= c ":") (<= c "@")) (and (>= c "[") (<= c "`")) (and (>= c "{") (<= c "~"))))
+      ((= class "P") (not (or (and (>= c "!") (<= c "/")) (and (>= c ":") (<= c "@")) (and (>= c "[") (<= c "`")) (and (>= c "{") (<= c "~")))))
+      ((= class "c") (or (< c " ") (= c "\127")))
+      ((= class "C") (not (or (< c " ") (= c "\127"))))
+      ((= class "x")
+        (or (and (>= c "0") (<= c "9")) (and (>= c "a") (<= c "f")) (and (>= c "A") (<= c "F"))))
+      ((= class "X")
+        (not (or (and (>= c "0") (<= c "9")) (and (>= c "a") (<= c "f")) (and (>= c "A") (<= c "F")))))
+      (else (= c class)))))
+
+;; Match a single "pattern atom" (char or class) at src position — returns true or false.
+;; pat-pos is position of the atom start. Returns the atom's length in pat (1 or 2).
+(define
+  lua-pat-atom-len
+  (fn (pat pat-pos)
+    (cond
+      ((>= pat-pos (len pat)) 0)
+      ((= (char-at pat pat-pos) "%") 2)
+      (else 1))))
+
+(define
+  lua-pat-atom-match
+  (fn (pat pat-pos s s-pos)
+    (cond
+      ((>= s-pos (len s)) false)
+      ((>= pat-pos (len pat)) false)
+      (else
+        (let ((pc (char-at pat pat-pos)) (sc (char-at s s-pos)))
+          (cond
+            ((= pc ".") true)
+            ((= pc "%")
+              (cond
+                ((>= (+ pat-pos 1) (len pat)) false)
+                (else (lua-pat-class-match (char-at pat (+ pat-pos 1)) sc))))
+            (else (= pc sc))))))))
+
+;; Match pattern against string starting at s-pos and pat-pos. Returns end-pos (inclusive+1) or -1 on no-match.
+(define
+  lua-pat-match
+  (fn (pat pat-pos s s-pos)
+    (cond
+      ((>= pat-pos (len pat)) s-pos)
+      ((and (= (char-at pat pat-pos) "$") (= (+ pat-pos 1) (len pat)))
+        (if (= s-pos (len s)) s-pos -1))
+      (else
+        (let
+          ((alen (lua-pat-atom-len pat pat-pos)))
+          (let
+            ((next-pat (+ pat-pos alen)))
+            (let
+              ((qc (if (< next-pat (len pat)) (char-at pat next-pat) "")))
+              (cond
+                ((= qc "*") (lua-pat-match-greedy pat pat-pos alen s s-pos (+ next-pat 1) 0))
+                ((= qc "+") (lua-pat-match-greedy pat pat-pos alen s s-pos (+ next-pat 1) 1))
+                ((= qc "-") (lua-pat-match-lazy pat pat-pos alen s s-pos (+ next-pat 1)))
+                ((= qc "?")
+                  (let ((tried (if (lua-pat-atom-match pat pat-pos s s-pos)
+                                   (lua-pat-match pat (+ next-pat 1) s (+ s-pos 1))
+                                   -1)))
+                    (if (>= tried 0) tried (lua-pat-match pat (+ next-pat 1) s s-pos))))
+                (else
+                  (cond
+                    ((lua-pat-atom-match pat pat-pos s s-pos)
+                      (lua-pat-match pat next-pat s (+ s-pos 1)))
+                    (else -1)))))))))))
+
+(define
+  lua-pat-match-greedy
+  (fn (pat atom-pos atom-len s s-pos rest-pat-pos min-count)
+    (let ((count 0) (i s-pos))
+      (begin
+        (define
+          count-loop
+          (fn ()
+            (when (lua-pat-atom-match pat atom-pos s i)
+              (begin (set! i (+ i 1)) (set! count (+ count 1)) (count-loop)))))
+        (count-loop)
+        (let ((best -1))
+          (begin
+            (define
+              try-loop
+              (fn (k)
+                (when (and (< best 0) (>= k min-count))
+                  (let ((r (lua-pat-match pat rest-pat-pos s (+ s-pos k))))
+                    (cond
+                      ((>= r 0) (set! best r))
+                      (else (try-loop (- k 1))))))))
+            (try-loop count)
+            best))))))
+
+(define
+  lua-pat-match-lazy
+  (fn (pat atom-pos atom-len s s-pos rest-pat-pos)
+    (let ((best -1) (i s-pos))
+      (begin
+        (define
+          try-loop
+          (fn ()
+            (when (< best 0)
+              (let ((r (lua-pat-match pat rest-pat-pos s i)))
+                (cond
+                  ((>= r 0) (set! best r))
+                  ((and (< i (len s)) (lua-pat-atom-match pat atom-pos s i))
+                    (begin (set! i (+ i 1)) (try-loop)))
+                  (else (set! best -2)))))))
+        (try-loop)
+        (if (= best -2) -1 best)))))
+
+;; Top-level find: return (start-index-0based . end-index) or nil on no match.
+;; If pat starts with ^, anchor to init. Otherwise scan.
+(define
+  lua-pat-find
+  (fn (pat s init)
+    (let
+      ((anchored (and (> (len pat) 0) (= (char-at pat 0) "^"))))
+      (let
+        ((start-pat (if anchored 1 0)))
+        (cond
+          (anchored
+            (let ((end (lua-pat-match pat start-pat s init)))
+              (if (>= end 0) (list init end) nil)))
+          (else
+            (let ((i init) (result nil))
+              (begin
+                (define
+                  scan
+                  (fn ()
+                    (when (and (= result nil) (<= i (len s)))
+                      (let ((end (lua-pat-match pat 0 s i)))
+                        (cond
+                          ((>= end 0) (set! result (list i end)))
+                          (else (begin (set! i (+ i 1)) (scan))))))))
+                (scan)
+                result))))))))
+
 (define string {})
 
 (define __ascii-32-126 " !\"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~")
@@ -784,20 +941,28 @@
   (fn (&rest args)
     (let ((s (first args))
           (pat (nth args 1))
-          (init (if (> (len args) 2) (nth args 2) 1)))
+          (init (if (> (len args) 2) (nth args 2) 1))
-      (let ((start-i (cond
+          (plain (if (> (len args) 3) (nth args 3) false)))
-                       ((< init 0) (+ (len s) init 1))
+      (let ((start-i0 (cond
-                       ((= init 0) 1)
+                        ((< init 0) (let ((v (+ (len s) init))) (if (< v 0) 0 v)))
-                       (else init))))
+                        ((= init 0) 0)
-        (let ((sub (if (<= start-i 1) s (substring s (- start-i 1) (len s)))))
+                        (else (- init 1)))))
-          (let ((idx (index-of sub pat)))
+        (cond
-            (cond
+          ((lua-truthy? plain)
-              ((< idx 0) nil)
+            (let ((sub (if (<= start-i0 0) s (substring s start-i0 (len s)))))
-              (else
+              (let ((idx (index-of sub pat)))
-                (list
+                (cond
-                  (quote lua-multi)
+                  ((< idx 0) nil)
-                  (+ start-i idx)
+                  (else
-                  (+ start-i idx (len pat) -1))))))))))
+                    (list
+                      (quote lua-multi)
+                      (+ start-i0 idx 1)
+                      (+ start-i0 idx (len pat))))))))
+          (else
+            (let ((r (lua-pat-find pat s start-i0)))
+              (cond
+                ((= r nil) nil)
+                (else (list (quote lua-multi) (+ (first r) 1) (nth r 1)))))))))))
 
 ;; Literal-only string.match: returns matched substring or nil (no captures since no pattern).
 (define

lib/lua/scoreboard.json

@@ -51,7 +51,7 @@
       "name": "attrib.lua",
       "status": "fail",
       "reason": "other: Unhandled exception: \\\"Unhandled exception: \\\\\\\"lua: module 'C' not found\\\\\\\"\\",
-      "ms": 6342
+      "ms": 6112
     },
     {
       "name": "big.lua",
@@ -63,7 +63,7 @@
       "name": "calls.lua",
       "status": "fail",
       "reason": "undefined symbol: fat\\",
-      "ms": 5134
+      "ms": 4800
     },
     {
       "name": "checktable.lua",
@@ -75,7 +75,7 @@
       "name": "closure.lua",
       "status": "timeout",
       "reason": "per-test timeout",
-      "ms": 8007
+      "ms": 8008
     },
     {
       "name": "code.lua",
@@ -87,7 +87,7 @@
       "name": "constructs.lua",
       "status": "fail",
       "reason": "other: Unhandled exception: \\\"Unhandled exception: \\\\\\\"lua: attempt to compare incompat",
-      "ms": 7182
+      "ms": 6618
     },
     {
       "name": "db.lua",
@@ -99,13 +99,13 @@
       "name": "errors.lua",
       "status": "fail",
       "reason": "other: Unhandled exception: \\\"Unhandled exception: \\\\\\\"assertion failed!\\\\\\\"\\",
-      "ms": 3464
+      "ms": 3080
     },
     {
       "name": "events.lua",
       "status": "fail",
       "reason": "other: Unhandled exception: \\\"Unhandled exception: \\\\\\\"assertion failed!\\\\\\\"\\",
-      "ms": 7726
+      "ms": 7192
     },
     {
       "name": "files.lua",
@@ -123,13 +123,13 @@
       "name": "literals.lua",
       "status": "fail",
       "reason": "other: Unhandled exception: \\\"Unhandled exception: \\\\\\\"assertion failed!\\\\\\\"\\",
-      "ms": 1940
+      "ms": 1777
     },
     {
       "name": "locals.lua",
       "status": "fail",
       "reason": "other: Unhandled exception: \\\"Unhandled exception: \\\\\\\"lua: attempt to call non-functio",
-      "ms": 1759
+      "ms": 1644
     },
     {
       "name": "main.lua",
@@ -141,43 +141,43 @@
       "name": "math.lua",
       "status": "fail",
       "reason": "other: Unhandled exception: \\\"Unhandled exception: \\\\\\\"assertion failed!\\\\\\\"\\",
-      "ms": 4487
+      "ms": 3999
     },
     {
       "name": "nextvar.lua",
       "status": "fail",
       "reason": "other: Unhandled exception: \\\"Unhandled exception: \\\\\\\"assertion failed!\\\\\\\"\\",
-      "ms": 7441
+      "ms": 6948
     },
     {
       "name": "pm.lua",
       "status": "fail",
       "reason": "other: Unhandled exception: \\\"Unhandled exception: \\\\\\\"assertion failed!\\\\\\\"\\",
-      "ms": 6440
+      "ms": 6326
     },
     {
       "name": "sort.lua",
       "status": "timeout",
       "reason": "per-test timeout",
-      "ms": 8007
+      "ms": 8008
     },
     {
       "name": "strings.lua",
       "status": "fail",
       "reason": "other: Unhandled exception: \\\"Unhandled exception: \\\\\\\"assertion failed!\\\\\\\"\\",
-      "ms": 4243
+      "ms": 3975
     },
     {
       "name": "vararg.lua",
       "status": "fail",
       "reason": "other: Unhandled exception: \\\"Unhandled exception: \\\\\\\"assertion failed!\\\\\\\"\\",
-      "ms": 2425
+      "ms": 2152
     },
     {
       "name": "verybig.lua",
       "status": "pass",
       "reason": "",
-      "ms": 622
+      "ms": 562
     }
   ]
 }

lib/lua/scoreboard.md

@@ -18,25 +18,25 @@ fail=12 timeout=3 skip=8 total=24
 |---|---|---|---:|
 | all.lua | skip | driver uses dofile to chain other tests | 0 |
 | api.lua | skip | requires testC (C debug library) | 0 |
-| attrib.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"lua: module 'C' not found\\\"\ | 6342 |
+| attrib.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"lua: module 'C' not found\\\"\ | 6112 |
 | big.lua | timeout | per-test timeout | 8008 |
-| calls.lua | fail | undefined symbol: fat\ | 5134 |
+| calls.lua | fail | undefined symbol: fat\ | 4800 |
 | checktable.lua | skip | internal debug helpers | 0 |
-| closure.lua | timeout | per-test timeout | 8007 |
+| closure.lua | timeout | per-test timeout | 8008 |
 | code.lua | skip | bytecode inspection via debug library | 0 |
-| constructs.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"lua: attempt to compare incompat | 7182 |
+| constructs.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"lua: attempt to compare incompat | 6618 |
 | db.lua | skip | debug library | 0 |
-| errors.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 3464 |
+| errors.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 3080 |
-| events.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 7726 |
+| events.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 7192 |
 | files.lua | skip | io library | 0 |
 | gc.lua | skip | collectgarbage / finalisers | 0 |
-| literals.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 1940 |
+| literals.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 1777 |
-| locals.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"lua: attempt to call non-functio | 1759 |
+| locals.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"lua: attempt to call non-functio | 1644 |
 | main.lua | skip | standalone interpreter driver | 0 |
-| math.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 4487 |
+| math.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 3999 |
-| nextvar.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 7441 |
+| nextvar.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 6948 |
-| pm.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 6440 |
+| pm.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 6326 |
-| sort.lua | timeout | per-test timeout | 8007 |
+| sort.lua | timeout | per-test timeout | 8008 |
-| strings.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 4243 |
+| strings.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 3975 |
-| vararg.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 2425 |
+| vararg.lua | fail | other: Unhandled exception: \"Unhandled exception: \\\"assertion failed!\\\"\ | 2152 |
-| verybig.lua | pass | - | 622 |
+| verybig.lua | pass | - | 562 |

plans/lua-on-sx.md

@@ -82,6 +82,7 @@ Each item: implement → tests → tick box → update progress log.
 
 _Newest first. Agent appends on every commit._
 
+- 2026-04-24: lua: scoreboard iteration — **Lua pattern engine (minimal)** for `string.find`. Supports character classes (`%d`/`%a`/`%s`/`%w`/`%p`/`%l`/`%u`/`%c`/`%x` + complements), `.` any, `^`/`$` anchors, quantifiers `*`/`+`/`-`/`?`, literal chars, `%%`. Added `plain` arg pathway. match/gmatch/gsub still literal. Scoreboard unchanged (pattern-using tests still hit other issues downstream).
 - 2026-04-24: lua: scoreboard iteration — `package.cpath`/`config`/`loaders`/`searchers`/`searchpath` stubs. attrib.lua moves from #9 (checking `package.cpath` is a string) to "module 'C' not found" — test requires filesystem-based module loading, not tractable. Most remaining failures need Lua pattern matching (pm.lua/strings.lua), env tracking (locals.lua/events.lua), or filesystem (attrib.lua).
 - 2026-04-24: lua: scoreboard iteration — **parenthesized expressions truncate multi-return** (Lua spec: `(f())` forces single value even if `f` returns multi). Parser wraps `(expr)` in a new `lua-paren` AST node; transpile emits `(lua-first inner)`. Fixes `constructs.lua`@30 (`a,b,c = (f())` expects `a=1, b=nil, c=nil`) and `math.lua`@13. 375/375 green (+2 paren tests). Scoreboard: 8× asserts (was 10).
 - 2026-04-24: lua: scoreboard iteration — stripped `(else (raise e))` from `lua-tx-loop-guard`. SX `guard` with `(else (raise e))` hangs in a loop (re-enters the same guard). Since unmatched sentinels fall through to the enclosing guard naturally, the else is unnecessary. Diagnosed `calls.lua` undefined-`fat`: `function fat(x)` defined at Lua top-level is scoped inside the SX top-level guard's scope; loadstring-captured closures don't see it via lexical env. Fix would require either dropping the top-level guard (breaking top-level `return`) or dynamic env access — deferred.