apache-commons-io/XmlStreamReader-can-t-parse-XML-document-with-multi-.patch

From 17f8b44d50372f4b540059232ed0ffa189eceb62 Mon Sep 17 00:00:00 2001
From: Gary Gregory <garydgregory@gmail.com>
Date: Tue, 2 Jan 2024 09:08:58 -0500
Subject: [PATCH] XmlStreamReader can't parse XML document with multi-line
 prolog #550

- Apply PR #550, not merged or would have caused the build to fail.
- Implement fix

Origin:
https://github.com/apache/commons-io/commit/17f8b44d50372f4b540059232ed0ffa189eceb62
---
 .../apache/commons/io/input/XmlStreamReader.java | 16 +++++++++++-----
 .../commons/io/input/XmlStreamReaderTest.java    | 10 ++++++++++
 2 files changed, 21 insertions(+), 5 deletions(-)

diff --git a/src/main/java/org/apache/commons/io/input/XmlStreamReader.java b/src/main/java/org/apache/commons/io/input/XmlStreamReader.java
index 2b9b379..ff16987 100644
--- a/src/main/java/org/apache/commons/io/input/XmlStreamReader.java
+++ b/src/main/java/org/apache/commons/io/input/XmlStreamReader.java
@@ -214,6 +214,16 @@ public class XmlStreamReader extends Reader {
      * <p>
      * See also the <a href="https://www.w3.org/TR/2008/REC-xml-20081126/#NT-EncName">XML specification</a>.
      * </p>
+     * <p>
+     * Note the documented pattern is:
+     * </p>
+     * <pre>
+     * EncName   ::=   [A-Za-z] ([A-Za-z0-9._] | '-')*
+     * </pre>
+     * <p>
+     * However this does not match all the aliases that are supported by Java.
+     * For example, '437', 'ISO_8859-1:1987' and 'ebcdic-de-273+euro'.
+     * </p>
      */
     public static final Pattern ENCODING_PATTERN = Pattern.compile(
     // @formatter:off
@@ -223,10 +233,6 @@ public class XmlStreamReader extends Reader {
             + "((?:\"[A-Za-z0-9][A-Za-z0-9._+:-]*\")"  // double-quoted
             +  "|(?:'[A-Za-z0-9][A-Za-z0-9._+:-]*'))", // single-quoted
             Pattern.MULTILINE);
-    // N.B. the documented pattern is
-    // EncName   ::=   [A-Za-z] ([A-Za-z0-9._] | '-')*
-    // However this does not match all the aliases that are supported by Java.
-    // e.g.  '437', 'ISO_8859-1:1987' and 'ebcdic-de-273+euro'
     // @formatter:on
 
     private static final String RAW_EX_1 = "Illegal encoding, BOM [{0}] XML guess [{1}] XML prolog [{2}] encoding mismatch";
@@ -325,7 +331,7 @@ public class XmlStreamReader extends Reader {
                 inputStream.reset();
                 final BufferedReader bReader = new BufferedReader(new StringReader(xmlProlog.substring(0, firstGT + 1)));
                 final StringBuilder prolog = new StringBuilder();
-                IOConsumer.forEach(bReader.lines(), prolog::append);
+                IOConsumer.forEach(bReader.lines(), l -> prolog.append(l).append(' '));
                 final Matcher m = ENCODING_PATTERN.matcher(prolog);
                 if (m.find()) {
                     encoding = m.group(1).toUpperCase(Locale.ROOT);
diff --git a/src/test/java/org/apache/commons/io/input/XmlStreamReaderTest.java b/src/test/java/org/apache/commons/io/input/XmlStreamReaderTest.java
index 63d587a..de986c9 100644
--- a/src/test/java/org/apache/commons/io/input/XmlStreamReaderTest.java
+++ b/src/test/java/org/apache/commons/io/input/XmlStreamReaderTest.java
@@ -60,6 +60,8 @@ public class XmlStreamReaderTest {
     private static final String UTF_32LE = "UTF-32LE";
     private static final String UTF_32BE = "UTF-32BE";
     private static final String UTF_8 = StandardCharsets.UTF_8.name();
+
+    private static final String XML6 = "xml-prolog-encoding-new-line";
     private static final String XML5 = "xml-prolog-encoding-spaced-single-quotes";
     private static final String XML4 = "xml-prolog-encoding-single-quotes";
     private static final String XML3 = "xml-prolog-encoding-double-quotes";
@@ -102,6 +104,8 @@ public class XmlStreamReaderTest {
 
     private static final MessageFormat XML_WITH_PROLOG = new MessageFormat(
             "<?xml version=\"1.0\"?>\n<root>{2}</root>");
+    private static final MessageFormat XML_WITH_PROLOG_AND_ENCODING_NEW_LINES = new MessageFormat(
+            "<?xml\nversion\n=\n\"1.0\"\nencoding\n=\n\"{1}\"\n?>\n<root>{2}</root>");
 
     private static final MessageFormat XML_WITH_PROLOG_AND_ENCODING_DOUBLE_QUOTES = new MessageFormat(
             "<?xml version=\"1.0\" encoding=\"{1}\"?>\n<root>{2}</root>");
@@ -123,6 +127,7 @@ public class XmlStreamReaderTest {
         XMLs.put(XML3, XML_WITH_PROLOG_AND_ENCODING_DOUBLE_QUOTES);
         XMLs.put(XML4, XML_WITH_PROLOG_AND_ENCODING_SINGLE_QUOTES);
         XMLs.put(XML5, XML_WITH_PROLOG_AND_ENCODING_SPACED_SINGLE_QUOTES);
+        XMLs.put(XML6, XML_WITH_PROLOG_AND_ENCODING_NEW_LINES);
     }
 
     /**
@@ -624,5 +629,10 @@ public class XmlStreamReaderTest {
         xmlReader = new XmlStreamReader(is);
         assertEquals(xmlReader.getEncoding(), encoding);
         xmlReader.close();
+
+        is = getXmlInputStream("no-bom", XML6, encoding, encoding);
+        xmlReader = new XmlStreamReader(is);
+        assertEquals(xmlReader.getEncoding(), encoding);
+        xmlReader.close();
     }
 }
-- 
2.47.0