fix: implement attribute normalization

Section 3.3.3 of the XML specification. Fixes #24
lddubeau · Oct 11, 2019 · be51114 · be51114
1 parent bed38a8
commit be51114
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 7 deletions.
diff --git a/lib/saxes.js b/lib/saxes.js
@@ -1907,8 +1907,10 @@ class SaxesParser {
         this.state = S_ENTITY;
         this.entityReturnState = S_ATTRIB_VALUE_QUOTED;
         return;
+      case NL:
       case NL_LIKE:
-        this.text += `${chunk.slice(start, this.prevI)}\n`;
+      case TAB:
+        this.text += `${chunk.slice(start, this.prevI)} `;
         start = this.i;
         break;
       case LESS:
@@ -1947,6 +1949,11 @@ class SaxesParser {
 
   /** @private */
   sAttribValueUnquoted() {
+    // We don't do anything regarding EOL or space handling for unquoted
+    // attributes. We already have failed by the time we get here, and the
+    // contract that saxes upholds states that upon failure, it is not safe to
+    // rely on the data passed to event handlers (other than
+    // ``onerror``). Passing "bad" data is not a problem.
     const c = this.captureTo(ATTRIB_VALUE_UNQUOTED_TERMINATOR);
     switch (c) {
     case AMP:

diff --git a/test/eol-handling.js b/test/eol-handling.js
@@ -25,15 +25,15 @@ describe("eol handling", () => {
       ["opentag", {
         name: "moo",
         attributes: {
-          a: "12\n        3",
+          a: "12         3",
         },
         isSelfClosing: false,
       }],
       ["text", "\n  abc\n  def\r\n  ghi\n\n  xx\nxx\n"],
       ["closetag", {
         name: "moo",
         attributes: {
-          a: "12\n        3",
+          a: "12         3",
         },
         isSelfClosing: false,
       }],
@@ -145,8 +145,8 @@ SYSTEM
       ["opentag", {
         name: "moo",
         attributes: {
-          a: "12\n        3",
-          b: "\nz\n",
+          a: "12         3",
+          b: " z ",
         },
         isSelfClosing: false,
       }],
@@ -178,8 +178,8 @@ abc
       ["closetag", {
         name: "moo",
         attributes: {
-          a: "12\n        3",
-          b: "\nz\n",
+          a: "12         3",
+          b: " z ",
         },
         isSelfClosing: false,
       }],