@es3d/es3d 1.0.0 → 1.0.1

package/src/format/compact-parser.js

@@ -1,171 +1,171 @@
-// Parser for the ES3D compact syntax: tokens -> the same AST as the markup syntax.
-//
-// Grammar (newlines are whitespace):
-//   document  := comment* root
-//   root      := 'es3d' header ( '{' statement* '}' | statement* )
-//   statement := element | comment | ';'
-//   element   := NAME header ( '{' statement* '}' )?
-//   header    := ('#' ID)? ('@' REF)? STRING? (NAME '=' VALUE)*
-
-import { Attr, Comment, Document, Element } from './ast.js';
-import { Diagnostic, ES3DError, LineIndex } from './errors.js';
-import { CTok, CompactTokenizer } from './compact-tokenizer.js';
-
-const ROOT_NAMES = new Set(['es3d', 'v3d']);
-
-export function parseCompact(text, lineIndex = new LineIndex(text)) {
-  const tz = new CompactTokenizer(text, lineIndex);
-  const buf = [];
-  const peek = (k = 0) => {
-    while (buf.length <= k) buf.push(tz.next());
-    return buf[k];
-  };
-  const take = () => (buf.length ? buf.shift() : tz.next());
-  const fail = (code, message, offset, length = 1, extra = {}) => {
-    const d = new Diagnostic({ code, message, offset, length, ...extra });
-    d.locate(lineIndex);
-    throw new ES3DError([d], lineIndex);
-  };
-  const where = (offset) => {
-    const p = lineIndex.position(offset);
-    return `line ${p.line}, column ${p.column}`;
-  };
-
-  function parseHeader(el) {
-    let last = el.nameEnd;
-    for (;;) {
-      const t = peek();
-      if (t.type === CTok.ID || t.type === CTok.REF) {
-        take();
-        const name = t.type === CTok.ID ? 'id' : 'ref';
-        if (el.attrNode(name)) fail('E_DUPLICATE_ATTRIBUTE', `Duplicate attribute '${name}' on ${el.name}.`, t.start, t.end - t.start);
-        el.attrs.push(new Attr(name, t.value, t.start, t.start + 1, t.start + 1, t.end, t.type === CTok.ID ? '#' : '@'));
-        last = t.end;
-      } else if (t.type === CTok.STRING) {
-        take();
-        if (el.text !== null) fail('E_SYNTAX', `${el.name} already has text content.`, t.start, t.end - t.start);
-        el.text = t.value;
-        el.textStart = t.start;
-        el.textEnd = t.end;
-        last = t.end;
-      } else if (t.type === CTok.NAME && peek(1).type === CTok.EQ) {
-        take();
-        take();
-        const v = take();
-        if (v.type !== CTok.VALUE) fail('E_MISSING_VALUE', `Expected a value for attribute '${t.value}'.`, v.start, 1);
-        for (const prev of el.attrs) {
-          if (prev.name === t.value) fail('E_DUPLICATE_ATTRIBUTE', `Duplicate attribute '${t.value}' on ${el.name}.`, t.start, t.end - t.start);
-        }
-        const a = new Attr(t.value, v.value, t.start, t.end, v.valueStart, v.valueEnd, v.quote);
-        a.decoded = !!v.escaped;
-        el.attrs.push(a);
-        last = v.end;
-      } else {
-        break;
-      }
-    }
-    el.openEnd = last;
-  }
-
-  function parseStatements(parent, closeTok) {
-    for (;;) {
-      const t = peek();
-      if (t.type === CTok.EOF) {
-        if (closeTok) {
-          fail('E_UNCLOSED_ELEMENT', `Unclosed block of '${parent.name}' (opened at ${where(closeTok.start)}): missing '}'.`, closeTok.start, 1);
-        }
-        return;
-      }
-      if (t.type === CTok.RBRACE) {
-        if (!closeTok) fail('E_UNEXPECTED_CLOSE', "Unexpected '}': no block is open.", t.start, 1);
-        take();
-        parent.closeStart = t.start;
-        parent.end = t.end;
-        return;
-      }
-      if (t.type === CTok.SEMI) {
-        take();
-        continue;
-      }
-      if (t.type === CTok.COMMENT) {
-        take();
-        parent.appendChild(new Comment(t.text, t.start, t.end));
-        continue;
-      }
-      if (t.type === CTok.NAME) {
-        if (peek(1).type === CTok.EQ) {
-          fail('E_SYNTAX', `Attribute '${t.value}' must follow an element name.`, t.start, t.end - t.start, {
-            hint: 'Attributes belong to the element written before them on the same statement.',
-          });
-        }
-        parent.appendChild(parseElement());
-        continue;
-      }
-      if (t.type === CTok.LBRACE) fail('E_SYNTAX', "Unexpected '{': a block must follow an element name.", t.start, 1);
-      if (t.type === CTok.STRING || t.type === CTok.ID || t.type === CTok.REF) {
-        fail('E_SYNTAX', 'Expected an element name.', t.start, t.end - t.start);
-      }
-      fail('E_SYNTAX', 'Unexpected token.', t.start, Math.max(1, t.end - t.start));
-    }
-  }
-
-  function parseElement() {
-    const t = take();
-    const el = new Element(t.value, t.start);
-    el.nameStart = t.start;
-    el.nameEnd = t.end;
-    parseHeader(el);
-    el.end = el.openEnd;
-    if (peek().type === CTok.LBRACE) {
-      const open = take();
-      el.selfClosing = false;
-      parseStatements(el, open);
-    } else {
-      el.selfClosing = true;
-    }
-    return el;
-  }
-
-  const leading = [];
-  while (peek().type === CTok.COMMENT || peek().type === CTok.SEMI) {
-    const c = take();
-    if (c.type === CTok.COMMENT) leading.push(new Comment(c.text, c.start, c.end));
-  }
-  const first = peek();
-  if (first.type === CTok.EOF) fail('E_EMPTY_DOCUMENT', "Empty document: expected 'es3d version=1'.", first.start, 0);
-  if (first.type !== CTok.NAME || !ROOT_NAMES.has(first.value)) {
-    fail('E_COMPACT_ROOT', "An ES3D compact document must start with the root element 'es3d'.", first.start, Math.max(1, first.end - first.start), {
-      hint: "Begin the file with: es3d version=1",
-    });
-  }
-  const rt = take();
-  const root = new Element(rt.value, rt.start);
-  root.nameStart = rt.start;
-  root.nameEnd = rt.end;
-  parseHeader(root);
-  root.end = root.openEnd;
-  const trailing = [];
-  if (peek().type === CTok.LBRACE) {
-    const open = take();
-    root.selfClosing = false;
-    parseStatements(root, open);
-    for (;;) {
-      const t = peek();
-      if (t.type === CTok.EOF) break;
-      if (t.type === CTok.COMMENT) {
-        take();
-        trailing.push(new Comment(t.text, t.start, t.end));
-      } else if (t.type === CTok.SEMI) take();
-      else fail('E_MULTIPLE_ROOTS', 'Content after the end of the root block.', t.start, Math.max(1, t.end - t.start));
-    }
-  } else {
-    // Header form: everything after the root header belongs to the root.
-    root.headerForm = true;
-    parseStatements(root, null);
-    root.selfClosing = root.children.length === 0;
-    root.end = text.length;
-  }
-  const doc = new Document({ root, syntax: 'compact', source: text, leading, trailing });
-  doc.lineIndex = lineIndex;
-  return doc;
-}
+// Parser for the ES3D compact syntax: tokens -> the same AST as the markup syntax.
+//
+// Grammar (newlines are whitespace):
+//   document  := comment* root
+//   root      := 'es3d' header ( '{' statement* '}' | statement* )
+//   statement := element | comment | ';'
+//   element   := NAME header ( '{' statement* '}' )?
+//   header    := ('#' ID)? ('@' REF)? STRING? (NAME '=' VALUE)*
+
+import { Attr, Comment, Document, Element } from './ast.js';
+import { Diagnostic, ES3DError, LineIndex } from './errors.js';
+import { CTok, CompactTokenizer } from './compact-tokenizer.js';
+
+const ROOT_NAMES = new Set(['es3d', 'v3d']);
+
+export function parseCompact(text, lineIndex = new LineIndex(text)) {
+  const tz = new CompactTokenizer(text, lineIndex);
+  const buf = [];
+  const peek = (k = 0) => {
+    while (buf.length <= k) buf.push(tz.next());
+    return buf[k];
+  };
+  const take = () => (buf.length ? buf.shift() : tz.next());
+  const fail = (code, message, offset, length = 1, extra = {}) => {
+    const d = new Diagnostic({ code, message, offset, length, ...extra });
+    d.locate(lineIndex);
+    throw new ES3DError([d], lineIndex);
+  };
+  const where = (offset) => {
+    const p = lineIndex.position(offset);
+    return `line ${p.line}, column ${p.column}`;
+  };
+
+  function parseHeader(el) {
+    let last = el.nameEnd;
+    for (;;) {
+      const t = peek();
+      if (t.type === CTok.ID || t.type === CTok.REF) {
+        take();
+        const name = t.type === CTok.ID ? 'id' : 'ref';
+        if (el.attrNode(name)) fail('E_DUPLICATE_ATTRIBUTE', `Duplicate attribute '${name}' on ${el.name}.`, t.start, t.end - t.start);
+        el.attrs.push(new Attr(name, t.value, t.start, t.start + 1, t.start + 1, t.end, t.type === CTok.ID ? '#' : '@'));
+        last = t.end;
+      } else if (t.type === CTok.STRING) {
+        take();
+        if (el.text !== null) fail('E_SYNTAX', `${el.name} already has text content.`, t.start, t.end - t.start);
+        el.text = t.value;
+        el.textStart = t.start;
+        el.textEnd = t.end;
+        last = t.end;
+      } else if (t.type === CTok.NAME && peek(1).type === CTok.EQ) {
+        take();
+        take();
+        const v = take();
+        if (v.type !== CTok.VALUE) fail('E_MISSING_VALUE', `Expected a value for attribute '${t.value}'.`, v.start, 1);
+        for (const prev of el.attrs) {
+          if (prev.name === t.value) fail('E_DUPLICATE_ATTRIBUTE', `Duplicate attribute '${t.value}' on ${el.name}.`, t.start, t.end - t.start);
+        }
+        const a = new Attr(t.value, v.value, t.start, t.end, v.valueStart, v.valueEnd, v.quote);
+        a.decoded = !!v.escaped;
+        el.attrs.push(a);
+        last = v.end;
+      } else {
+        break;
+      }
+    }
+    el.openEnd = last;
+  }
+
+  function parseStatements(parent, closeTok) {
+    for (;;) {
+      const t = peek();
+      if (t.type === CTok.EOF) {
+        if (closeTok) {
+          fail('E_UNCLOSED_ELEMENT', `Unclosed block of '${parent.name}' (opened at ${where(closeTok.start)}): missing '}'.`, closeTok.start, 1);
+        }
+        return;
+      }
+      if (t.type === CTok.RBRACE) {
+        if (!closeTok) fail('E_UNEXPECTED_CLOSE', "Unexpected '}': no block is open.", t.start, 1);
+        take();
+        parent.closeStart = t.start;
+        parent.end = t.end;
+        return;
+      }
+      if (t.type === CTok.SEMI) {
+        take();
+        continue;
+      }
+      if (t.type === CTok.COMMENT) {
+        take();
+        parent.appendChild(new Comment(t.text, t.start, t.end));
+        continue;
+      }
+      if (t.type === CTok.NAME) {
+        if (peek(1).type === CTok.EQ) {
+          fail('E_SYNTAX', `Attribute '${t.value}' must follow an element name.`, t.start, t.end - t.start, {
+            hint: 'Attributes belong to the element written before them on the same statement.',
+          });
+        }
+        parent.appendChild(parseElement());
+        continue;
+      }
+      if (t.type === CTok.LBRACE) fail('E_SYNTAX', "Unexpected '{': a block must follow an element name.", t.start, 1);
+      if (t.type === CTok.STRING || t.type === CTok.ID || t.type === CTok.REF) {
+        fail('E_SYNTAX', 'Expected an element name.', t.start, t.end - t.start);
+      }
+      fail('E_SYNTAX', 'Unexpected token.', t.start, Math.max(1, t.end - t.start));
+    }
+  }
+
+  function parseElement() {
+    const t = take();
+    const el = new Element(t.value, t.start);
+    el.nameStart = t.start;
+    el.nameEnd = t.end;
+    parseHeader(el);
+    el.end = el.openEnd;
+    if (peek().type === CTok.LBRACE) {
+      const open = take();
+      el.selfClosing = false;
+      parseStatements(el, open);
+    } else {
+      el.selfClosing = true;
+    }
+    return el;
+  }
+
+  const leading = [];
+  while (peek().type === CTok.COMMENT || peek().type === CTok.SEMI) {
+    const c = take();
+    if (c.type === CTok.COMMENT) leading.push(new Comment(c.text, c.start, c.end));
+  }
+  const first = peek();
+  if (first.type === CTok.EOF) fail('E_EMPTY_DOCUMENT', "Empty document: expected 'es3d version=1'.", first.start, 0);
+  if (first.type !== CTok.NAME || !ROOT_NAMES.has(first.value)) {
+    fail('E_COMPACT_ROOT', "An ES3D compact document must start with the root element 'es3d'.", first.start, Math.max(1, first.end - first.start), {
+      hint: "Begin the file with: es3d version=1",
+    });
+  }
+  const rt = take();
+  const root = new Element(rt.value, rt.start);
+  root.nameStart = rt.start;
+  root.nameEnd = rt.end;
+  parseHeader(root);
+  root.end = root.openEnd;
+  const trailing = [];
+  if (peek().type === CTok.LBRACE) {
+    const open = take();
+    root.selfClosing = false;
+    parseStatements(root, open);
+    for (;;) {
+      const t = peek();
+      if (t.type === CTok.EOF) break;
+      if (t.type === CTok.COMMENT) {
+        take();
+        trailing.push(new Comment(t.text, t.start, t.end));
+      } else if (t.type === CTok.SEMI) take();
+      else fail('E_MULTIPLE_ROOTS', 'Content after the end of the root block.', t.start, Math.max(1, t.end - t.start));
+    }
+  } else {
+    // Header form: everything after the root header belongs to the root.
+    root.headerForm = true;
+    parseStatements(root, null);
+    root.selfClosing = root.children.length === 0;
+    root.end = text.length;
+  }
+  const doc = new Document({ root, syntax: 'compact', source: text, leading, trailing });
+  doc.lineIndex = lineIndex;
+  return doc;
+}