@es3d/es3d 1.0.0 → 1.0.1

package/src/format/xml-parser.js

@@ -1,136 +1,136 @@
-// Parser for the ES3D markup syntax: tokens -> AST (Document of Elements).
-
-import { Attr, Comment, Document, Element } from './ast.js';
-import { Diagnostic, ES3DError, LineIndex } from './errors.js';
-import { Tok, XmlTokenizer, decodeEntities } from './xml-tokenizer.js';
-
-function isBlank(s) {
-  for (let i = 0; i < s.length; i++) {
-    const c = s.charCodeAt(i);
-    if (c !== 32 && c !== 10 && c !== 9 && c !== 13) return false;
-  }
-  return true;
-}
-
-export function parseXml(text, lineIndex = new LineIndex(text)) {
-  const tz = new XmlTokenizer(text, lineIndex);
-  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 ${p.column}`;
-  };
-
-  const stack = [];
-  let root = null;
-  const leading = [];
-  const trailing = [];
-
-  for (;;) {
-    const t = tz.next();
-    switch (t.type) {
-      case Tok.EOF: {
-        if (stack.length) {
-          const open = stack[stack.length - 1];
-          fail('E_UNCLOSED_ELEMENT', `Unclosed element <${open.name}> (opened at ${where(open.start)}).`, open.start, open.nameEnd - open.start, {
-            hint: `Add </${open.name}> or write <${open.name} ... /> for an element without children.`,
-          });
-        }
-        if (!root) fail('E_EMPTY_DOCUMENT', 'Empty document: expected an <es3d> root element.', t.start, 0);
-        const doc = new Document({ root, syntax: 'xml', source: text, leading, trailing });
-        doc.lineIndex = lineIndex;
-        return doc;
-      }
-      case Tok.PI:
-        if (root || stack.length) fail('E_SYNTAX', 'Processing instructions (<?...?>) are only allowed before the root element.', t.start, 2);
-        break;
-      case Tok.DOCTYPE:
-        if (root || stack.length) fail('E_SYNTAX', 'A DOCTYPE is only allowed before the root element.', t.start, 2);
-        break;
-      case Tok.COMMENT: {
-        const c = new Comment(t.text, t.start, t.end);
-        if (stack.length) stack[stack.length - 1].appendChild(c);
-        else if (root) trailing.push(c);
-        else leading.push(c);
-        break;
-      }
-      case Tok.TEXT:
-      case Tok.CDATA: {
-        const raw = t.type === Tok.TEXT ? t.raw : t.text;
-        if (isBlank(raw)) break;
-        if (!stack.length) {
-          let k = t.start;
-          while (k < t.end && /\s/.test(text[k])) k++;
-          fail('E_TEXT_OUTSIDE_ROOT', 'Text is not allowed outside the root element.', k, 1);
-        }
-        const el = stack[stack.length - 1];
-        const value = t.type === Tok.TEXT ? decodeEntities(raw, t.start, (code, msg, off, len) => fail(code, msg, off, len)) : raw;
-        if (el.text === null) {
-          el.text = value;
-          el.textStart = t.start;
-        } else {
-          el.text += value;
-        }
-        el.textEnd = t.end;
-        break;
-      }
-      case Tok.OPEN: {
-        if (!stack.length && root) {
-          fail('E_MULTIPLE_ROOTS', `Only one root element is allowed; found a second root <${t.name}>.`, t.start, t.nameEnd - t.start);
-        }
-        const el = new Element(t.name, t.start);
-        el.nameStart = t.nameStart;
-        el.nameEnd = t.nameEnd;
-        let a = tz.next();
-        while (a.type === Tok.ATTR) {
-          for (const prev of el.attrs) {
-            if (prev.name === a.name) {
-              fail('E_DUPLICATE_ATTRIBUTE', `Duplicate attribute '${a.name}' on <${el.name}>.`, a.nameStart, a.nameEnd - a.nameStart);
-            }
-          }
-          const attr = new Attr(a.name, a.value, a.nameStart, a.nameEnd, a.valueStart, a.valueEnd, a.quote);
-          attr.decoded = a.decoded;
-          el.attrs.push(attr);
-          a = tz.next();
-        }
-        // a is OPEN_END (the tokenizer throws on anything else)
-        el.openEnd = a.end;
-        if (stack.length) stack[stack.length - 1].appendChild(el);
-        else root = el;
-        if (a.selfClosing) {
-          el.selfClosing = true;
-          el.end = a.end;
-        } else {
-          el.selfClosing = false;
-          stack.push(el);
-        }
-        break;
-      }
-      case Tok.CLOSE: {
-        if (!stack.length) {
-          fail('E_UNEXPECTED_CLOSE', `Unexpected closing tag </${t.name}>: no element is open.`, t.start, t.end - t.start);
-        }
-        const open = stack[stack.length - 1];
-        if (open.name !== t.name) {
-          const deeper = stack.findIndex((e) => e.name === t.name);
-          fail(
-            'E_MISMATCHED_TAG',
-            `Mismatched closing tag: expected </${open.name}> (opened at ${where(open.start)}) but found </${t.name}>.`,
-            t.start,
-            t.end - t.start,
-            deeper >= 0 ? { hint: `<${open.name}> is not closed before </${t.name}>.` } : {},
-          );
-        }
-        stack.pop();
-        open.closeStart = t.start;
-        open.end = t.end;
-        break;
-      }
-      default:
-        fail('E_SYNTAX', 'Unexpected token.', t.start, 1);
-    }
-  }
-}
+// Parser for the ES3D markup syntax: tokens -> AST (Document of Elements).
+
+import { Attr, Comment, Document, Element } from './ast.js';
+import { Diagnostic, ES3DError, LineIndex } from './errors.js';
+import { Tok, XmlTokenizer, decodeEntities } from './xml-tokenizer.js';
+
+function isBlank(s) {
+  for (let i = 0; i < s.length; i++) {
+    const c = s.charCodeAt(i);
+    if (c !== 32 && c !== 10 && c !== 9 && c !== 13) return false;
+  }
+  return true;
+}
+
+export function parseXml(text, lineIndex = new LineIndex(text)) {
+  const tz = new XmlTokenizer(text, lineIndex);
+  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 ${p.column}`;
+  };
+
+  const stack = [];
+  let root = null;
+  const leading = [];
+  const trailing = [];
+
+  for (;;) {
+    const t = tz.next();
+    switch (t.type) {
+      case Tok.EOF: {
+        if (stack.length) {
+          const open = stack[stack.length - 1];
+          fail('E_UNCLOSED_ELEMENT', `Unclosed element <${open.name}> (opened at ${where(open.start)}).`, open.start, open.nameEnd - open.start, {
+            hint: `Add </${open.name}> or write <${open.name} ... /> for an element without children.`,
+          });
+        }
+        if (!root) fail('E_EMPTY_DOCUMENT', 'Empty document: expected an <es3d> root element.', t.start, 0);
+        const doc = new Document({ root, syntax: 'xml', source: text, leading, trailing });
+        doc.lineIndex = lineIndex;
+        return doc;
+      }
+      case Tok.PI:
+        if (root || stack.length) fail('E_SYNTAX', 'Processing instructions (<?...?>) are only allowed before the root element.', t.start, 2);
+        break;
+      case Tok.DOCTYPE:
+        if (root || stack.length) fail('E_SYNTAX', 'A DOCTYPE is only allowed before the root element.', t.start, 2);
+        break;
+      case Tok.COMMENT: {
+        const c = new Comment(t.text, t.start, t.end);
+        if (stack.length) stack[stack.length - 1].appendChild(c);
+        else if (root) trailing.push(c);
+        else leading.push(c);
+        break;
+      }
+      case Tok.TEXT:
+      case Tok.CDATA: {
+        const raw = t.type === Tok.TEXT ? t.raw : t.text;
+        if (isBlank(raw)) break;
+        if (!stack.length) {
+          let k = t.start;
+          while (k < t.end && /\s/.test(text[k])) k++;
+          fail('E_TEXT_OUTSIDE_ROOT', 'Text is not allowed outside the root element.', k, 1);
+        }
+        const el = stack[stack.length - 1];
+        const value = t.type === Tok.TEXT ? decodeEntities(raw, t.start, (code, msg, off, len) => fail(code, msg, off, len)) : raw;
+        if (el.text === null) {
+          el.text = value;
+          el.textStart = t.start;
+        } else {
+          el.text += value;
+        }
+        el.textEnd = t.end;
+        break;
+      }
+      case Tok.OPEN: {
+        if (!stack.length && root) {
+          fail('E_MULTIPLE_ROOTS', `Only one root element is allowed; found a second root <${t.name}>.`, t.start, t.nameEnd - t.start);
+        }
+        const el = new Element(t.name, t.start);
+        el.nameStart = t.nameStart;
+        el.nameEnd = t.nameEnd;
+        let a = tz.next();
+        while (a.type === Tok.ATTR) {
+          for (const prev of el.attrs) {
+            if (prev.name === a.name) {
+              fail('E_DUPLICATE_ATTRIBUTE', `Duplicate attribute '${a.name}' on <${el.name}>.`, a.nameStart, a.nameEnd - a.nameStart);
+            }
+          }
+          const attr = new Attr(a.name, a.value, a.nameStart, a.nameEnd, a.valueStart, a.valueEnd, a.quote);
+          attr.decoded = a.decoded;
+          el.attrs.push(attr);
+          a = tz.next();
+        }
+        // a is OPEN_END (the tokenizer throws on anything else)
+        el.openEnd = a.end;
+        if (stack.length) stack[stack.length - 1].appendChild(el);
+        else root = el;
+        if (a.selfClosing) {
+          el.selfClosing = true;
+          el.end = a.end;
+        } else {
+          el.selfClosing = false;
+          stack.push(el);
+        }
+        break;
+      }
+      case Tok.CLOSE: {
+        if (!stack.length) {
+          fail('E_UNEXPECTED_CLOSE', `Unexpected closing tag </${t.name}>: no element is open.`, t.start, t.end - t.start);
+        }
+        const open = stack[stack.length - 1];
+        if (open.name !== t.name) {
+          const deeper = stack.findIndex((e) => e.name === t.name);
+          fail(
+            'E_MISMATCHED_TAG',
+            `Mismatched closing tag: expected </${open.name}> (opened at ${where(open.start)}) but found </${t.name}>.`,
+            t.start,
+            t.end - t.start,
+            deeper >= 0 ? { hint: `<${open.name}> is not closed before </${t.name}>.` } : {},
+          );
+        }
+        stack.pop();
+        open.closeStart = t.start;
+        open.end = t.end;
+        break;
+      }
+      default:
+        fail('E_SYNTAX', 'Unexpected token.', t.start, 1);
+    }
+  }
+}