Destructurer dispatcher v2: admission-keyed buckets, two levels deep

The v1 dispatcher tried arms in declaration order, so the late pratt op forms
paid ~20 failed unifications per expression node. v2 derives each arm's
FIRST-CHILD admission keys from its step plan (node rule / leaf tokenType /
literal first charCode, with nullable-first arms admitting everything) and
dispatches through nested switches: node child -> rule bucket, leaf ->
tokenType bucket ($keyword/$punct sub-switched on the first character). Big
node-rule buckets (a pratt rule's self bucket holds every led + op form)
sub-dispatch one level deeper on c[1] - the connector position. Buckets are
SUPERSET filters (every arm fn still verifies exactly) and preserve
declaration order internally, so tie semantics are unchanged.

Destructure-EVERY-node overhead over the bare CST pass: +64% (v1) -> +39%
(one level) -> +31% (two levels); totality unchanged (32,336 nodes, 0 misses
across all seven grammars + the TS corpus sample) and the tsc-structural
oracle stays node-identical (32/32).

Author: johnsoncodehk

html.cst-match.ts

@@ -158,10 +158,25 @@ function _Element$lt4(c: readonly CstChild[], src: string): ElementMatch | null
 
 export function matchElement(n: ElementNode, src: string): ElementMatch {
   const c = n.children;
-  { const m = _Element$lt(c, src); if (m !== null) return m; }
-  { const m = _Element$lt2(c, src); if (m !== null) return m; }
-  { const m = _Element$lt3(c, src); if (m !== null) return m; }
-  { const m = _Element$lt4(c, src); if (m !== null) return m; }
+  const k0 = c[0] as (CstChild & { tokenType?: string; rule?: string }) | undefined;
+  if (k0 === undefined) {
+  } else if (k0.tokenType === undefined) {
+    switch (k0.rule) {
+    }
+  } else if (k0.tokenType === '$keyword' || k0.tokenType === '$punct') {
+    switch (src.charCodeAt(k0.offset)) {
+      case 60: {
+        { const m = _Element$lt(c, src); if (m !== null) return m; }
+        { const m = _Element$lt2(c, src); if (m !== null) return m; }
+        { const m = _Element$lt3(c, src); if (m !== null) return m; }
+        { const m = _Element$lt4(c, src); if (m !== null) return m; }
+        break;
+      }
+    }
+  } else {
+    switch (k0.tokenType) {
+    }
+  }
   throw new Error("matchElement: no arm matches" + ' @' + n.offset);
 }
 
@@ -222,7 +237,22 @@ function _Attr$name(c: readonly CstChild[], src: string): AttrMatch | null {
 
 export function matchAttr(n: AttrNode, src: string): AttrMatch {
   const c = n.children;
-  { const m = _Attr$name(c, src); if (m !== null) return m; }
+  const k0 = c[0] as (CstChild & { tokenType?: string; rule?: string }) | undefined;
+  if (k0 === undefined) {
+  } else if (k0.tokenType === undefined) {
+    switch (k0.rule) {
+    }
+  } else if (k0.tokenType === '$keyword' || k0.tokenType === '$punct') {
+    switch (src.charCodeAt(k0.offset)) {
+    }
+  } else {
+    switch (k0.tokenType) {
+      case "Name": {
+        { const m = _Attr$name(c, src); if (m !== null) return m; }
+        break;
+      }
+    }
+  }
   throw new Error("matchAttr: no arm matches" + ' @' + n.offset);
 }
 
@@ -274,10 +304,34 @@ function _Node$text(c: readonly CstChild[], src: string): NodeMatch | null {
 
 export function matchNode(n: NodeNode, src: string): NodeMatch {
   const c = n.children;
-  { const m = _Node$element(c, src); if (m !== null) return m; }
-  { const m = _Node$comment(c, src); if (m !== null) return m; }
-  { const m = _Node$rawText(c, src); if (m !== null) return m; }
-  { const m = _Node$text(c, src); if (m !== null) return m; }
+  const k0 = c[0] as (CstChild & { tokenType?: string; rule?: string }) | undefined;
+  if (k0 === undefined) {
+  } else if (k0.tokenType === undefined) {
+    switch (k0.rule) {
+      case "Element": {
+        { const m = _Node$element(c, src); if (m !== null) return m; }
+        break;
+      }
+    }
+  } else if (k0.tokenType === '$keyword' || k0.tokenType === '$punct') {
+    switch (src.charCodeAt(k0.offset)) {
+    }
+  } else {
+    switch (k0.tokenType) {
+      case "Comment": {
+        { const m = _Node$comment(c, src); if (m !== null) return m; }
+        break;
+      }
+      case "RawText": {
+        { const m = _Node$rawText(c, src); if (m !== null) return m; }
+        break;
+      }
+      case "Text": {
+        { const m = _Node$text(c, src); if (m !== null) return m; }
+        break;
+      }
+    }
+  }
   throw new Error("matchNode: no arm matches" + ' @' + n.offset);
 }
 
@@ -333,7 +387,31 @@ function _Document$element(c: readonly CstChild[], src: string): DocumentMatch |
 
 export function matchDocument(n: DocumentNode, src: string): DocumentMatch {
   const c = n.children;
-  { const m = _Document$element(c, src); if (m !== null) return m; }
+  const k0 = c[0] as (CstChild & { tokenType?: string; rule?: string }) | undefined;
+  if (k0 === undefined) {
+    { const m = _Document$element(c, src); if (m !== null) return m; }
+  } else if (k0.tokenType === undefined) {
+    switch (k0.rule) {
+      case "Element": {
+        { const m = _Document$element(c, src); if (m !== null) return m; }
+        break;
+      }
+    }
+  } else if (k0.tokenType === '$keyword' || k0.tokenType === '$punct') {
+    switch (src.charCodeAt(k0.offset)) {
+    }
+  } else {
+    switch (k0.tokenType) {
+      case "Comment": {
+        { const m = _Document$element(c, src); if (m !== null) return m; }
+        break;
+      }
+      case "Text": {
+        { const m = _Document$element(c, src); if (m !== null) return m; }
+        break;
+      }
+    }
+  }
   throw new Error("matchDocument: no arm matches" + ' @' + n.offset);
 }