[Git][java-team/nekohtml][upstream] Import Upstream version 1.9.22.noko2
Emmanuel Bourg (@ebourg)
gitlab at salsa.debian.org
Fri Sep 29 01:23:05 BST 2023
Emmanuel Bourg pushed to branch upstream at Debian Java Maintainers / nekohtml
Commits:
0c2a35f8 by Emmanuel Bourg at 2023-09-29T02:20:53+02:00
Import Upstream version 1.9.22.noko2
- - - - -
9 changed files:
- build.xml
- data/meta/test-meta-encoding3.html.canonical
- pom.xml
- src/org/cyberneko/html/HTMLScanner.java
- src/org/cyberneko/html/HTMLTagBalancingListener.java
- src/org/cyberneko/html/LostText.java
- test/java/org/cyberneko/html/HTMLScannerTest.java
- test/java/org/cyberneko/html/HTMLTagBalancingListenerTest.java
- test/java/org/cyberneko/html/LocatorEncodingTest.java
Changes:
=====================================
build.xml
=====================================
@@ -4,7 +4,7 @@
<!-- PROPERTIES -->
<property file='build-custom.properties' />
- <property name='version' value='1.9.22'/>
+ <property name='version' value='1.9.22.noko2'/>
<property name='name' value='nekohtml'/>
<property name='fullname' value='${name}-${version}'/>
<property name='Title' value='NekoHTML'/>
@@ -14,8 +14,8 @@
<property name='copyright' value='(C) Copyright 2002-2014, ${author}. All rights reserved.'/>
<property name='URL' value='http://nekohtml.sourceforge.net/index.html'/>
- <property name='compile.source' value='1.3' />
- <property name='compile.target' value='1.3' />
+ <property name='compile.source' value='1.7' />
+ <property name='compile.target' value='1.7' />
<property name='src.dir' value='src' />
<property name='src.test.dir' value='test/java' />
@@ -51,7 +51,7 @@
<!-- DEPENDENCIES -->
<available property='bcel.available' file='${lib.dir}/bcel-5.2.jar' />
- <property name="xerces.version" value="2.9.1"/>
+ <property name="xerces.version" value="2.12.2"/>
<!-- TARGETS -->
<target name='all' depends='zip,tgz'/>
@@ -103,13 +103,7 @@
unless="compile.skip">
<mkdir dir="${build.classes.dir}"/>
- <compileWith xercesVersion="2.0.2" bridge="2_0"/>
- <compileWith xercesVersion="2.1.0" bridge="2_1"/>
- <compileWith xercesVersion="2.2.1" bridge="2_2"/>
- <compileWith xercesVersion="2.3.0" bridge="2_3"/>
- <compileWith xercesVersion="2.8.1" bridge="2_3"/>
- <compileWith xercesVersion="2.9.1" bridge="2_3"/>
- <compileWith xercesVersion="2.10.0" bridge="2_3"/>
+ <compileWith xercesVersion="${xerces.version}" bridge="2_3"/>
<copy todir="${build.classes.dir}">
<fileset dir='${src.dir}' includes='**' excludes='META-INF/**, **/*.java*' />
@@ -178,8 +172,11 @@
author='true' version='true' use='true'
windowtitle="${Name} Implementation"
doctitle="${Name}" encoding="ISO-8859-1"
- bottom="${copyright}"
- />
+ bottom="${copyright}">
+ <classpath>
+ <fileset dir='${lib.dir}/xerces-${xerces.version}' includes='*.jar' />
+ </classpath>
+ </javadoc>
</target>
<target name='doc-init'>
@@ -268,7 +265,8 @@ public class Version {
source='${compile.source}' target='${compile.target}'
includeAntRuntime='true'>
<classpath>
- <fileset dir='${lib.dir}' includes='xml-apis.jar,xerces*.jar, junit*.jar' />
+ <fileset dir='${lib.dir}' includes='junit*.jar' />
+ <fileset dir='${lib.dir}/xerces-${xerces.version}' includes='xml-apis*.jar,xerces*.jar' />
<pathelement location='${build.classes.dir}' />
</classpath>
<src path='${src.test.dir}' />
@@ -308,17 +306,7 @@ public class Version {
<delete dir="${build.dir}/junit"/>
<mkdir dir="${build.dir}/junit"/>
- <testWith xercesVersion="2.11.0"/>
- <testWith xercesVersion="2.10.0"/>
- <testWith xercesVersion="2.9.1"/>
- <testWith xercesVersion="2.8.1"/>
- <testWith xercesVersion="2.3.0"/>
- <testWith xercesVersion="2.2.1"/>
- <testWith xercesVersion="minimal"/>
- <!--
- <testWith xercesVersion="2.1.0"/>
- <testWith xercesVersion="2.0.2"/>
- -->
+ <testWith xercesVersion="${xerces.version}"/>
<junitreport todir="${build.dir}/junit">
<fileset dir="${build.dir}/junit">
=====================================
data/meta/test-meta-encoding3.html.canonical
=====================================
@@ -8,6 +8,6 @@ Ahttp-equiv Content-Type
"\n
)HEAD
(BODY
-"\n宋体\n\n新\n\n細明體\n\n宋体\n\n浠茇忘�\n\n？\n\n
+"\n宋体\n\n新\n\n細明體\n\n宋体\n\n浠茇忘�\n\n？\n
)BODY
)HTML
=====================================
pom.xml
=====================================
@@ -4,7 +4,7 @@
<artifactId>nekohtml</artifactId>
<name>Neko HTML</name>
<description>An HTML parser and tag balancer.</description>
- <version>1.9.22</version>
+ <version>1.9.22.noko2</version>
<url>http://nekohtml.sourceforge.net/</url>
<licenses>
<license>
@@ -20,7 +20,7 @@
<dependency>
<groupId>xerces</groupId>
<artifactId>xercesImpl</artifactId>
- <version>2.11.0</version>
+ <version>2.12.2</version>
</dependency>
</dependencies>
<developers>
=====================================
src/org/cyberneko/html/HTMLScanner.java
=====================================
@@ -2588,7 +2588,7 @@ public class HTMLScanner
if (c == '?' || c == '/') {
char c0 = (char)c;
c = fCurrentEntity.read();
- if (c == '>') {
+ if (c == '>' || c == -1) {
break;
}
fStringBuffer.append(c0);
@@ -2888,18 +2888,18 @@ public class HTMLScanner
else if (c == '>') {
return false;
}
- else if (c == '<') {
- fCurrentEntity.rewind();
- return false;
+ else if(c == '<') {
+ fCurrentEntity.rewind();
+ if(fReportErrors) {
+ fErrorReporter.reportError("HTML2001", null);
+ }
+ return false;
}
fCurrentEntity.rewind();
String aname = scanName(false);
- if (aname == null) {
- if (fReportErrors) {
- fErrorReporter.reportError("HTML1011", null);
- }
- empty[0] = skipMarkup(false);
- return false;
+ if(aname == null) {
+ empty[0] = skipMarkup(false);
+ return false;
}
if (!skippedSpaces && fReportErrors) {
fErrorReporter.reportError("HTML1013", new Object[] { aname });
=====================================
src/org/cyberneko/html/HTMLTagBalancingListener.java
=====================================
@@ -28,7 +28,7 @@ import org.apache.xerces.xni.XMLAttributes;
* <li>activate the tag balancing feature</li>
* </ul>
* @author Marc Guillemot
- * @version $Id: HTMLTagBalancingListener.java 260 2009-09-02 08:26:01Z mguillem $
+ * @version $Id$
*/
public interface HTMLTagBalancingListener
{
=====================================
src/org/cyberneko/html/LostText.java
=====================================
@@ -28,7 +28,7 @@ import org.apache.xerces.xni.XMLString;
* in <body>
* @author Marc Guillemot
*
- * @version $Id: LostText.java 226 2009-02-09 20:48:44Z mguillem $
+ * @version $Id$
*/
class LostText
{
=====================================
test/java/org/cyberneko/html/HTMLScannerTest.java
=====================================
@@ -16,6 +16,7 @@ import org.apache.xerces.xni.XMLAttributes;
import org.apache.xerces.xni.XNIException;
import org.apache.xerces.xni.parser.XMLDocumentFilter;
import org.apache.xerces.xni.parser.XMLInputSource;
+import org.apache.xerces.xni.parser.XMLParseException;
import org.apache.xerces.xni.parser.XMLParserConfiguration;
import org.cyberneko.html.filters.DefaultFilter;
@@ -40,6 +41,27 @@ public class HTMLScannerTest extends TestCase {
assertFalse(scanner.isEncodingCompatible("UTF-16","Cp1252"));
}
+ public void testInvalidInput() throws Exception {
+ HTMLConfiguration parser = new HTMLConfiguration();
+ parser.setFeature("http://cyberneko.org/html/features/report-errors", true);
+ final ArrayList errors = new ArrayList();
+ parser.setProperty("http://cyberneko.org/html/properties/error-reporter", new HTMLErrorReporter() {
+ public String formatMessage(String key, Object[] args) {
+ return null;
+ }
+
+ public void reportWarning(String key, Object[] args) throws XMLParseException {
+ }
+
+ public void reportError(String key, Object[] args) throws XMLParseException {
+ errors.add(key);
+ }
+ });
+ XMLInputSource source = new XMLInputSource(null, "myTest", null, new StringReader("<div </div>"), "UTF-8");
+ parser.parse(source);
+ assertEquals("Expected to receive an error", 1, errors.size());
+ }
+
public void testEvaluateInputSource() throws Exception {
String string = "<html><head><title>foo</title></head>"
+ "<body>"
=====================================
test/java/org/cyberneko/html/HTMLTagBalancingListenerTest.java
=====================================
@@ -17,7 +17,7 @@ import org.apache.xerces.xni.parser.XMLInputSource;
/**
* Unit tests for {@link HTMLTagBalancingListener}.
* @author Marc Guillemot
- * @version $Id: HTMLTagBalancingListenerTest.java 320 2013-01-29 07:46:36Z mguillem $
+ * @version $Id$
*/
public class HTMLTagBalancingListenerTest extends TestCase {
=====================================
test/java/org/cyberneko/html/LocatorEncodingTest.java
=====================================
@@ -18,7 +18,7 @@ import org.xml.sax.ext.Locator2;
/**
* Regression test for <a href="http://sourceforge.net/tracker/?func=detail&atid=952178&aid=3381270&group_id=195122">Bug 3381270</a>.
* @author Marc Guillemot
- * @version $Revision: 291 $
+ * @version $Revision$
*/
public class LocatorEncodingTest extends TestCase {
View it on GitLab: https://salsa.debian.org/java-team/nekohtml/-/commit/0c2a35f8a0025c92eb15c5a760cc29f91379c85f
--
View it on GitLab: https://salsa.debian.org/java-team/nekohtml/-/commit/0c2a35f8a0025c92eb15c5a760cc29f91379c85f
You're receiving this email because of your account on salsa.debian.org.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://alioth-lists.debian.net/pipermail/pkg-java-commits/attachments/20230929/a349be6e/attachment.htm>
More information about the pkg-java-commits
mailing list