From 13ae6d493a1129b6958b10be6ab1e557485c0bbc Mon Sep 17 00:00:00 2001
From: Calixte Denizet <calixte.denizet@gmail.com>
Date: Sun, 7 Nov 2021 21:41:37 +0100
Subject: [PATCH] XFA - Encode tag names in UTF-8 when saving (fix #14249)

---
 src/core/xfa/xfa_object.js           | 10 ++++++----
 test/unit/xfa_serialize_data_spec.js |  4 +++-
 2 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/src/core/xfa/xfa_object.js b/src/core/xfa/xfa_object.js
index b1808fb0f..e22c4435c 100644
--- a/src/core/xfa/xfa_object.js
+++ b/src/core/xfa/xfa_object.js
@@ -14,7 +14,7 @@
  */
 
 import { getInteger, getKeyword, HTMLResult } from "./utils.js";
-import { shadow, warn } from "../../shared/util.js";
+import { shadow, utf8StringToString, warn } from "../../shared/util.js";
 import { encodeToXmlString } from "../core_utils.js";
 import { NamespaceIds } from "./namespaces.js";
 import { searchNode } from "./som.js";
@@ -819,10 +819,12 @@ class XmlObject extends XFAObject {
       buf.push(encodeToXmlString(this[$content]));
       return;
     }
+    const utf8TagName = utf8StringToString(tagName);
     const prefix = this[$namespaceId] === NS_DATASETS ? "xfa:" : "";
-    buf.push(`<${prefix}${tagName}`);
+    buf.push(`<${prefix}${utf8TagName}`);
     for (const [name, value] of this[_attributes].entries()) {
-      buf.push(` ${name}="${encodeToXmlString(value[$content])}"`);
+      const utf8Name = utf8StringToString(name);
+      buf.push(` ${utf8Name}="${encodeToXmlString(value[$content])}"`);
     }
     if (this[_dataValue] !== null) {
       if (this[_dataValue]) {
@@ -848,7 +850,7 @@ class XmlObject extends XFAObject {
         child[$toString](buf);
       }
     }
-    buf.push(`</${prefix}${tagName}>`);
+    buf.push(`</${prefix}${utf8TagName}>`);
   }
 
   [$onChild](child) {
diff --git a/test/unit/xfa_serialize_data_spec.js b/test/unit/xfa_serialize_data_spec.js
index 6e29cd006..faeaf4c84 100644
--- a/test/unit/xfa_serialize_data_spec.js
+++ b/test/unit/xfa_serialize_data_spec.js
@@ -33,6 +33,7 @@ describe("Data serializer", function () {
           <Units>1</Units>
           <Unit_Price>250.00</Unit_Price>
           <Total_Price>250.00</Total_Price>
+          <àé></àé>
         </Detail>
         <Page>2</Page>
         <Detail PartNo="RRB-LB">
@@ -59,6 +60,7 @@ describe("Data serializer", function () {
       ["Receipt.Detail[0].Units", "12&3"],
       ["Receipt.Detail[0].Unit_Price", "456>"],
       ["Receipt.Detail[0].Total_Price", "789"],
+      ["Receipt.Detail[0].àé", "1011"],
       ["Receipt.Detail[1].PartNo", "foo-bar😀"],
       ["Receipt.Detail[1].Description", "hello world"],
     ]) {
@@ -66,7 +68,7 @@ describe("Data serializer", function () {
     }
 
     const serialized = dataHandler.serialize(storage);
-    const expected = `<xfa:datasets xmlns:xfa="http://www.xfa.org/schema/xfa-data/1.0/"><foo>bar</foo><bar>foo</bar><xfa:data><Receipt><Page>1</Page><Detail PartNo="GS001"><Description>Giant Slingshot</Description><Units>12&amp;3</Units><Unit_Price>456&gt;</Unit_Price><Total_Price>789</Total_Price></Detail><Page>2</Page><Detail PartNo="foo-bar&#x1F600;"><Description>hello world</Description><Units>5</Units><Unit_Price>12.00</Unit_Price><Total_Price>60.00</Total_Price></Detail><Sub_Total>310.00</Sub_Total><Tax>24.80</Tax><Total_Price>334.80</Total_Price></Receipt></xfa:data></xfa:datasets>`;
+    const expected = `<xfa:datasets xmlns:xfa="http://www.xfa.org/schema/xfa-data/1.0/"><foo>bar</foo><bar>foo</bar><xfa:data><Receipt><Page>1</Page><Detail PartNo="GS001"><Description>Giant Slingshot</Description><Units>12&amp;3</Units><Unit_Price>456&gt;</Unit_Price><Total_Price>789</Total_Price><\xC3\xA0\xC3\xA9>1011</\xC3\xA0\xC3\xA9></Detail><Page>2</Page><Detail PartNo="foo-bar&#x1F600;"><Description>hello world</Description><Units>5</Units><Unit_Price>12.00</Unit_Price><Total_Price>60.00</Total_Price></Detail><Sub_Total>310.00</Sub_Total><Tax>24.80</Tax><Total_Price>334.80</Total_Price></Receipt></xfa:data></xfa:datasets>`;
 
     expect(serialized).toEqual(expected);
   });