Widened JSoup range for HTML services.

ObeoNetwork · Apr 21, 2023 · 27c3dec · 27c3dec
1 parent 5aceb77
commit 27c3dec
Show file tree

Hide file tree

Showing 8 changed files with 70 additions and 15 deletions.
diff --git a/plugins/org.obeonetwork.m2doc.html/META-INF/MANIFEST.MF b/plugins/org.obeonetwork.m2doc.html/META-INF/MANIFEST.MF
@@ -11,7 +11,7 @@ Require-Bundle: org.eclipse.acceleo.annotations;bundle-version="[5.0.0,8.0.0)",
  org.eclipse.emf.ecore,
  org.obeonetwork.m2doc;bundle-version="[3.0.0,4.0.0)",
  org.apache.poi;bundle-version="[5.2.3,5.2.4)",
- org.jsoup;bundle-version="[1.8.3,1.8.4)",
+ org.jsoup;bundle-version="[1.14.3,2.0.0)",
  org.eclipse.acceleo.query
 Export-Package: org.obeonetwork.m2doc.html.services
 Automatic-Module-Name: org.obeonetwork.m2doc.html
diff --git a/...s/org.obeonetwork.m2doc.html/src/org/obeonetwork/m2doc/html/services/M2DocHTMLParser.java b/...s/org.obeonetwork.m2doc.html/src/org/obeonetwork/m2doc/html/services/M2DocHTMLParser.java
@@ -13,6 +13,7 @@
 
 import java.awt.Color;
 import java.math.BigInteger;
+import java.nio.charset.StandardCharsets;
 import java.util.ArrayList;
 import java.util.HashMap;
 import java.util.List;
@@ -426,7 +427,7 @@ public List<MElement> parse(URI baseURI, String htmlString) {
 
         final Document document = Jsoup.parse(htmlString, baseURI.toString());
         document.outputSettings().syntax(org.jsoup.nodes.Document.OutputSettings.Syntax.xml);
-        document.outputSettings().charset("UTF-8");
+        document.outputSettings().charset(StandardCharsets.UTF_8);
 
         final MStyle defaultStyle = new MStyleImpl(null, -1, null, null, -1);
         if (document.body().hasAttr("bgcolor")) {
@@ -756,7 +757,7 @@ private void endElement(MList parent, MElement element) {
      *            tells if a new paragraph is needed
      */
     private void insertText(MList parent, final Context context, TextNode node, boolean needNewParagraph) {
-        final String text = node.text();
+        final String text = text(node);
         if (!text.trim().isEmpty()) {
             final String textToInsert;
             if (needNewParagraph) {
@@ -778,11 +779,11 @@ private void insertText(MList parent, final Context context, TextNode node, bool
     }
 
     /**
-     * Trims the begining of the given {@link String}.
+     * Trims the beginning of the given {@link String}.
      * 
      * @param text
      *            the {@link String}
-     * @return the trimed {@link String}
+     * @return the trimmed {@link String}
      */
     private String trimFirst(String text) {
         final String res;
@@ -801,6 +802,66 @@ private String trimFirst(String text) {
         return res;
     }
 
+    /**
+     * Gets the text of the given {@link TextNode}.
+     * Taken from JSoup see https://github.com/jhy/jsoup/issues/1063
+     * 
+     * @param textNode
+     *            the {@link TextNode}
+     * @return the text of the given {@link TextNode}
+     */
+    private String text(TextNode textNode) {
+        final String string = textNode.getWholeText();
+        StringBuilder sb = new StringBuilder(string.length());
+        appendNormalisedWhitespace(sb, string, false);
+        return sb.toString();
+    }
+
+    /**
+     * Tests if a code point is "whitespace" as defined in the HTML spec.
+     * Taken from JSoup see https://github.com/jhy/jsoup/issues/1063
+     * 
+     * @param c
+     *            code point to test
+     * @return true if code point is whitespace, false otherwise
+     */
+    private boolean isWhitespace(int c) {
+        return c == ' ' || c == '\t' || c == '\n' || c == '\f' || c == '\r';
+    }
+
+    /**
+     * After normalizing the whitespace within a string, appends it to a string builder.
+     * Taken from JSoup see https://github.com/jhy/jsoup/issues/1063
+     * 
+     * @param accum
+     *            builder to append to
+     * @param string
+     *            string to normalize whitespace within
+     * @param stripLeading
+     *            set to true if you wish to remove any leading whitespace
+     */
+    public void appendNormalisedWhitespace(StringBuilder accum, String string, boolean stripLeading) {
+        boolean lastWasWhite = false;
+        boolean reachedNonWhite = false;
+
+        int len = string.length();
+        int c;
+        for (int i = 0; i < len; i += Character.charCount(c)) {
+            c = string.codePointAt(i);
+            if (isWhitespace(c)) {
+                if ((stripLeading && !reachedNonWhite) || lastWasWhite) {
+                    continue;
+                }
+                accum.append(' ');
+                lastWasWhite = true;
+            } else {
+                accum.appendCodePoint(c);
+                lastWasWhite = false;
+                reachedNonWhite = true;
+            }
+        }
+    }
+
     /**
      * Starts the given {@link Element}.
      * 
@@ -822,7 +883,7 @@ private MList startElement(MList parent, Context context, Element element) {
         } else if (BLOCKQUOTE_TAG.equals(nodeName)) {
             if (element.childNodeSize() > 0 && element.childNode(0) instanceof TextNode) {
                 TextNode textNode = (TextNode) element.childNode(0);
-                String newText = trimFirst(textNode.text());
+                String newText = trimFirst(text(textNode));
                 textNode.text(newText);
                 if (!newText.isEmpty()) {
                     res = createMParagraph(context, parent, element, null, null);

diff --git a/releng/org.obeonetwork.m2doc.targetplatforms/2022-12/m2doc-2022-12.target b/releng/org.obeonetwork.m2doc.targetplatforms/2022-12/m2doc-2022-12.target
@@ -1,7 +1,7 @@
 <?xml version="1.0" encoding="UTF-8" standalone="no"?>
 <?pde?>
 <!-- generated with https://github.com/eclipse-cbi/targetplatform-dsl -->
-<target name="M2Doc 2022-12 Target Platform" sequenceNumber="1680171126">
+<target name="M2Doc 2022-12 Target Platform" sequenceNumber="1682072983">
   <locations>
     <location includeMode="planner" includeAllPlatforms="false" includeSource="true" includeConfigurePhase="false" type="InstallableUnit">
       <unit id="org.eclipse.license.feature.group" version="0.0.0"/>
@@ -71,7 +71,6 @@
     </location>
     <location includeMode="planner" includeAllPlatforms="false" includeSource="true" includeConfigurePhase="false" type="InstallableUnit">
       <unit id="org.kohsuke.args4j" version="2.0.21.v201301150030"/>
-      <unit id="org.jsoup" version="1.8.3.v20181012-1713"/>
       <unit id="javax.servlet" version="3.1.0.v201410161800"/>
       <repository location="https://download.eclipse.org/tools/orbit/downloads/drops/R20190827152740/repository/"/>
     </location>

diff --git a/releng/org.obeonetwork.m2doc.targetplatforms/2022-12/m2doc-2022-12.tpd b/releng/org.obeonetwork.m2doc.targetplatforms/2022-12/m2doc-2022-12.tpd
@@ -80,7 +80,6 @@ location Orbit-202212 "https://download.eclipse.org/tools/orbit/downloads/drops/
 
 location "https://download.eclipse.org/tools/orbit/downloads/drops/R20190827152740/repository/" {
 	org.kohsuke.args4j [2.0.21,2.1.0)
-	org.jsoup [1.8.3,1.8.4)
 	javax.servlet [3.1.0,3.2.0)
 }
 

diff --git a/releng/org.obeonetwork.m2doc.targetplatforms/capella-6.1.0/m2doc-capella-6.1.0.target b/releng/org.obeonetwork.m2doc.targetplatforms/capella-6.1.0/m2doc-capella-6.1.0.target
@@ -1,7 +1,7 @@
 <?xml version="1.0" encoding="UTF-8" standalone="no"?>
 <?pde?>
 <!-- generated with https://github.com/eclipse-cbi/targetplatform-dsl -->
-<target name="M2Doc Capella 6.1.0 Target Platform" sequenceNumber="1681983032">
+<target name="M2Doc Capella 6.1.0 Target Platform" sequenceNumber="1682072995">
   <locations>
     <location includeMode="planner" includeAllPlatforms="false" includeSource="true" includeConfigurePhase="false" type="InstallableUnit">
       <unit id="org.eclipse.gmf.runtime.notation.sdk.feature.group" version="0.0.0"/>
@@ -171,7 +171,6 @@
     </location>
     <location includeMode="planner" includeAllPlatforms="false" includeSource="true" includeConfigurePhase="false" type="InstallableUnit">
       <unit id="org.kohsuke.args4j" version="2.0.21.v201301150030"/>
-      <unit id="org.jsoup" version="1.8.3.v20181012-1713"/>
       <unit id="javax.servlet" version="3.1.0.v201410161800"/>
       <repository location="https://download.eclipse.org/tools/orbit/downloads/drops/R20190827152740/repository/"/>
     </location>

diff --git a/releng/org.obeonetwork.m2doc.targetplatforms/capella-6.1.0/m2doc-capella-6.1.0.tpd b/releng/org.obeonetwork.m2doc.targetplatforms/capella-6.1.0/m2doc-capella-6.1.0.tpd
@@ -181,7 +181,6 @@ location site_Capella "https://download.eclipse.org/capella/core/updates/release
 
 location "https://download.eclipse.org/tools/orbit/downloads/drops/R20190827152740/repository/" {
 	org.kohsuke.args4j [2.0.21,2.1.0)
-	org.jsoup [1.8.3,1.8.4)
 	javax.servlet [3.1.0,3.2.0)
 }
 

diff --git a/releng/org.obeonetwork.m2doc.targetplatforms/sirius-7.1.0/m2doc-sirius-7.1.0.target b/releng/org.obeonetwork.m2doc.targetplatforms/sirius-7.1.0/m2doc-sirius-7.1.0.target
@@ -1,7 +1,7 @@
 <?xml version="1.0" encoding="UTF-8" standalone="no"?>
 <?pde?>
 <!-- generated with https://github.com/eclipse-cbi/targetplatform-dsl -->
-<target name="M2Doc Sirius 7.1.0 Target Platform" sequenceNumber="1680173290">
+<target name="M2Doc Sirius 7.1.0 Target Platform" sequenceNumber="1682072991">
   <locations>
     <location includeMode="planner" includeAllPlatforms="false" includeSource="true" includeConfigurePhase="false" type="InstallableUnit">
       <unit id="org.eclipse.license.feature.group" version="0.0.0"/>
@@ -74,7 +74,6 @@
     </location>
     <location includeMode="planner" includeAllPlatforms="false" includeSource="true" includeConfigurePhase="false" type="InstallableUnit">
       <unit id="org.kohsuke.args4j" version="2.0.21.v201301150030"/>
-      <unit id="org.jsoup" version="1.8.3.v20181012-1713"/>
       <unit id="javax.servlet" version="3.1.0.v201410161800"/>
       <repository location="https://download.eclipse.org/tools/orbit/downloads/drops/R20190827152740/repository/"/>
     </location>

diff --git a/releng/org.obeonetwork.m2doc.targetplatforms/sirius-7.1.0/m2doc-sirius-7.1.0.tpd b/releng/org.obeonetwork.m2doc.targetplatforms/sirius-7.1.0/m2doc-sirius-7.1.0.tpd
@@ -84,7 +84,6 @@ location Orbit-202212 "https://download.eclipse.org/tools/orbit/downloads/drops/
 
 location "https://download.eclipse.org/tools/orbit/downloads/drops/R20190827152740/repository/" {
 	org.kohsuke.args4j [2.0.21,2.1.0)
-	org.jsoup [1.8.3,1.8.4)
 	javax.servlet [3.1.0,3.2.0)
 }