@nodable/flexible-xml-parser 1.4.0 → 1.6.0

package/src/XmlPartReader.js

@@ -48,6 +48,10 @@ export function readClosingTagName(source) {
  */
 export function readTagExp(parser) {
   parser.source.markTokenStart(1);
+  // Absolute document offset where `exp` (tag name onward, right after '<')
+  // begins — captured before any reads so buildTagExpObj can compute each
+  // attribute's absolute document position from its offset within attrsExp.
+  const expStart = parser.source.startIndex;
   let inSingleQuotes = false;
   let inDoubleQuotes = false;
   let i;
@@ -77,7 +81,7 @@ export function readTagExp(parser) {
 
   const exp = parser.source.readStr(i);
   parser.source.updateBufferBoundary(i + 1);
-  return buildTagExpObj(exp, parser);
+  return buildTagExpObj(exp, parser, expStart);
 }
 
 /**
@@ -90,6 +94,7 @@ export function readTagExp(parser) {
  */
 export function readPiExp(parser) {
   parser.source.markTokenStart(1);
+  const expStart = parser.source.startIndex;
   let inSingleQuotes = false;
   let inDoubleQuotes = false;
   let i;
@@ -127,7 +132,7 @@ export function readPiExp(parser) {
 
   const exp = parser.source.readStr(i);
   parser.source.updateBufferBoundary(i + 2);
-  return buildTagExpObj(exp, parser);
+  return buildTagExpObj(exp, parser, expStart, true);
 }
 
 // ─── Internal helpers ─────────────────────────────────────────────────────────
@@ -135,16 +140,22 @@ export function readPiExp(parser) {
 /**
  * Parse a raw tag expression string into a structured tag descriptor.
  *
- * @param {string} exp    - everything between '<' and '>' (exclusive)
+ * @param {string} exp      - everything between '<' and '>' (exclusive)
  * @param {object} parser
- * @returns {{ tagName, selfClosing, rawAttributes, _attrsExp }}
+ * @param {number} [expStart] - absolute document offset where `exp` begins
+ *   (i.e. right after '<' or '<?'). Used to compute each attribute's absolute
+ *   document position (tagExp._attrsExpStart) for addAttribute()'s meta arg.
+ *   Optional so callers that don't have/need it can omit it — attribute
+ *   position metadata is simply unavailable in that case, not an error.
+ * @returns {{ tagName, selfClosing, rawAttributes, _attrsExp, _attrsExpStart }}
  */
-function buildTagExpObj(exp, parser) {
+function buildTagExpObj(exp, parser, expStart, forceToReadAttrs = false) {
   const tagExp = {
     tagName: "",
     selfClosing: false,
     rawAttributes: Object.create(null),
     _attrsExp: "", // stored for two-pass attribute flushing in readOpeningTag
+    _attrsExpStart: undefined, // absolute document offset of _attrsExp's first char
   };
 
   const expLen = exp.length;
@@ -163,6 +174,7 @@ function buildTagExpObj(exp, parser) {
     if (isSpace(c)) {
       tagExp.tagName = exp.substring(0, i);
       attrsExp = exp.substring(i + 1);
+      if (expStart !== undefined) tagExp._attrsExpStart = expStart + i + 1;
       break;
     }
   }
@@ -171,13 +183,13 @@ function buildTagExpObj(exp, parser) {
   tagExp.tagName = tagExp.tagName.trimEnd();
   tagExp._attrsExp = attrsExp;
 
-  if (!isQName(tagExp.tagName, parser.xmlVersion)) {
+  if (!isQName(tagExp.tagName, parser.xmlDec.version)) {
     throw new ParseError("Invalid tag name", ErrorCode.INVALID_TAG_NAME);
   }
 
   // Pass 1: collect raw attribute values for matcher.updateCurrent().
   // Pass 2 (flushAttributes) runs later in readOpeningTag, after updateCurrent().
-  if (!parser.options.skip.attributes && attrsExp.length > 0) {
+  if (forceToReadAttrs || !parser.options.skip.attributes && attrsExp.length > 0) {
     collectRawAttributes(attrsExp, parser, tagExp);
   }
   // console.log(tagExp)

package/src/XmlSpecialTagsReader.js

@@ -46,9 +46,7 @@ export function readPiTag(parser) {
     // Read version from the declaration and store it on the parser for validators.
     const version = tagExp.rawAttributes?.version;
     if (version === '1.1') {
-      parser.xmlVersion = 1.1;
-    } else {
-      parser.xmlVersion = 1.0; // default
+      parser.xmlDec.version = 1.1;
     }
   }
 
@@ -57,15 +55,15 @@ export function readPiTag(parser) {
   // does for regular tags. PI tags are not pushed onto the matcher, so no
   // updateCurrent() call is needed here.
   if (!skipOptions.attributes) {
-    flushAttributes(tagExp._attrsExp, parser);
+    flushAttributes(tagExp._attrsExp, parser, tagExp._attrsExpStart);
   }
 
   if (tagExp.tagName === "xml") {
     //TODO: verify it is very first tag else error
-    if (!skipOptions.declaration) {
-      parser.outputBuilder.addDeclaration("?xml");
+    if (!skipOptions.declaration) { //TODO: unnecessary. builder can ommit it from response if not needed
+      parser.outputBuilder.addDeclaration("?xml", parser.xmlDec);
     }
-  } else if (!skipOptions.pi) {
+  } else if (!skipOptions.pi) { //TODO: unnecessary. builder can ommit it from response if not needed
     parser.outputBuilder.addInstruction("?" + tagExp.tagName);
   }
 }