From 10ae9fbdc70d225888cd847fd4cdd51338d78ed4 Mon Sep 17 00:00:00 2001
From: tallison <tallison@apache.org>
Date: Sat, 10 Oct 2026 09:01:41 -0400
Subject: [PATCH 1/3] TIKA-4966: apply font formatting styles in docx

---
 CHANGES.txt                                   |   6 +
 .../parser/microsoft/OfficeParserConfig.java  |  18 +++
 .../tika/parser/microsoft/WordExtractor.java  |  12 +-
 .../microsoft/ooxml/FormattingTagManager.java |  35 ++++-
 .../ooxml/OOXMLTikaBodyPartHandler.java       |  30 ++++-
 .../OOXMLWordAndPowerPointTextHandler.java    |  12 +-
 .../parser/microsoft/ooxml/RunProperties.java |  45 +++++++
 .../microsoft/ooxml/xwpf/XWPFStylesShim.java  | 123 +++++++++++++++---
 .../microsoft/ooxml/OOXMLDocxSAXTest.java     |  34 +++++
 .../testWORD_characterStyles.docx             | Bin 0 -> 1887 bytes
 10 files changed, 280 insertions(+), 35 deletions(-)
 create mode 100644 tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/resources/test-documents/testWORD_characterStyles.docx

diff --git a/CHANGES.txt b/CHANGES.txt
index 48dc600091e..d1b2a3ef41c 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -1,5 +1,11 @@
 Release 4.2.0 - unreleased
 
+   * The docx parser now applies bold and italics set by a run's character style
+     (e.g. Word's built-in Strong and Emphasis), following basedOn chains; previously
+     only direct formatting was emitted. The new OfficeParserConfig option
+     includeCharacterStyleClasses wraps such runs in <span class="..."> named after
+     the style (TIKA-4966).
+
    * Config properties whose value reaches the host (paths, executables, class names,
      URLs, credentials, operator-sized limits) are marked @OperatorOnly and refused in
      per-request JSON before any setter runs. Newly locked: pdf-parser's
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/OfficeParserConfig.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/OfficeParserConfig.java
index 5b174caaaa4..6e8708baffa 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/OfficeParserConfig.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/OfficeParserConfig.java
@@ -46,6 +46,7 @@ public class OfficeParserConfig implements Serializable {
     private int rtfEmbeddedMaxBytesInKb = 2 * 1024 * 1024; // 2 GB
 
     private boolean includeGlossary = true;
+    private boolean includeCharacterStyleClasses = false;
     private String dateOverrideFormat = null;
     private int maxOverride = 0;//ignore
 
@@ -203,6 +204,23 @@ public void setIncludeGlossary(boolean includeGlossary) {
         this.includeGlossary = includeGlossary;
     }
 
+    public boolean isIncludeCharacterStyleClasses() {
+        return includeCharacterStyleClasses;
+    }
+
+    /**
+     * Whether to wrap docx runs that carry a character style in
+     * <code>&lt;span class="..."&gt;</code>, named after the style the same way
+     * paragraph styles are.
+     * <p/>
+     * Default: <code>false</code>
+     *
+     * @param includeCharacterStyleClasses whether to emit character style classes
+     */
+    public void setIncludeCharacterStyleClasses(boolean includeCharacterStyleClasses) {
+        this.includeCharacterStyleClasses = includeCharacterStyleClasses;
+    }
+
     public boolean isIncludeMissingRows() {
         return includeMissingRows;
     }
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java
index 346de85dcc5..9af6233f229 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java
@@ -170,14 +170,20 @@ public static TagAndStyle buildParagraphTagAndStyle(String styleName, boolean is
             // Turn it into a H1 - H6 (H7+ isn't valid!)
             tag = "h" + Math.min(num, 6);
         } else {
-            styleClass = styleName.replace(' ', '_');
-            styleClass =
-                    styleClass.substring(0, 1).toLowerCase(Locale.ROOT) + styleClass.substring(1);
+            styleClass = toStyleClass(styleName);
         }
 
         return new TagAndStyle(tag, styleClass);
     }
 
+    /**
+     * @return the XHTML class for a style name, e.g. "List Bullet" -&gt; "list_Bullet"
+     */
+    public static String toStyleClass(String styleName) {
+        String styleClass = styleName.replace(' ', '_');
+        return styleClass.substring(0, 1).toLowerCase(Locale.ROOT) + styleClass.substring(1);
+    }
+
     protected void parse(POIFSFileSystem filesystem, XHTMLContentHandler xhtml)
             throws IOException, SAXException, TikaException {
         parse(filesystem.getRoot(), xhtml);
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/FormattingTagManager.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/FormattingTagManager.java
index 3e8ae2f0de8..8717b524aa0 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/FormattingTagManager.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/FormattingTagManager.java
@@ -44,8 +44,9 @@ class FormattingTagManager {
 
     private final XHTMLContentHandler xhtml;
 
-    // Outermost to innermost: hyperlink > bold > italic > strike > underline
+    // Outermost to innermost: hyperlink > span > bold > italic > strike > underline
     private String currentHyperlink = null;
+    private String currentSpanClass = null;
     private boolean wrapperHyperlinkActive = false;
     private boolean isBold = false;
     private boolean isItalics = false;
@@ -65,7 +66,7 @@ void openHyperlink(String url) throws SAXException {
         if (url == null) {
             return;
         }
-        closeFormattingTags();
+        closeInlineTags();
         if (currentHyperlink != null) {
             xhtml.endElement("a");
         }
@@ -79,7 +80,7 @@ void openHyperlink(String url) throws SAXException {
      */
     void closeHyperlink() throws SAXException {
         if (currentHyperlink != null && wrapperHyperlinkActive) {
-            closeFormattingTags();
+            closeInlineTags();
             xhtml.endElement("a");
             currentHyperlink = null;
             wrapperHyperlinkActive = false;
@@ -98,11 +99,19 @@ boolean isHyperlinkActive() {
      * opening and closing XHTML tags as needed to maintain proper nesting.
      */
     void applyFormatting(RunProperties runProperties) throws SAXException {
+        applyFormatting(runProperties, null);
+    }
+
+    /**
+     * As {@link #applyFormatting(RunProperties)}, additionally wrapping the run in
+     * {@code <span class="spanClass">} when spanClass is non-null.
+     */
+    void applyFormatting(RunProperties runProperties, String spanClass) throws SAXException {
         // Run-property hyperlinks only when no wrapper is active
         if (!wrapperHyperlinkActive) {
             String newHyperlink = runProperties.getHlinkClickUrl();
             if (!Objects.equals(newHyperlink, currentHyperlink)) {
-                closeFormattingTags();
+                closeInlineTags();
                 if (currentHyperlink != null) {
                     xhtml.endElement("a");
                 }
@@ -113,6 +122,14 @@ void applyFormatting(RunProperties runProperties) throws SAXException {
             }
         }
 
+        if (!Objects.equals(spanClass, currentSpanClass)) {
+            closeInlineTags();
+            if (spanClass != null) {
+                xhtml.startElement("span", "class", spanClass);
+            }
+            currentSpanClass = spanClass;
+        }
+
         if (runProperties.isBold() != isBold) {
             // Close inner tags before flipping <b>. Nesting is <b><i><s><u>
             // (outermost to innermost), so close innermost first: u, s, i.
@@ -182,7 +199,7 @@ void applyFormatting(RunProperties runProperties) throws SAXException {
      * Closes all currently open tags in proper nesting order.
      */
     void closeAll() throws SAXException {
-        closeFormattingTags();
+        closeInlineTags();
         if (currentHyperlink != null) {
             xhtml.endElement("a");
             currentHyperlink = null;
@@ -190,6 +207,14 @@ void closeAll() throws SAXException {
         }
     }
 
+    private void closeInlineTags() throws SAXException {
+        closeFormattingTags();
+        if (currentSpanClass != null) {
+            xhtml.endElement("span");
+            currentSpanClass = null;
+        }
+    }
+
     private void closeFormattingTags() throws SAXException {
         if (isUnderline) {
             xhtml.endElement("u");
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/OOXMLTikaBodyPartHandler.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/OOXMLTikaBodyPartHandler.java
index 0366ab96700..df823c853d0 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/OOXMLTikaBodyPartHandler.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/OOXMLTikaBodyPartHandler.java
@@ -49,6 +49,8 @@ public class OOXMLTikaBodyPartHandler
     private final XWPFListManager listManager;
     private final boolean includeDeletedText;
     private final boolean includeMoveFromText;
+    private final boolean includeCharacterStyleClasses;
+    private final RunProperties effectiveRunProperties = new RunProperties();
     private final XWPFStylesShim styles;
     private final Metadata metadata;
 
@@ -95,6 +97,7 @@ public OOXMLTikaBodyPartHandler(XHTMLContentHandler xhtml, Metadata metadata) {
         this.listManager = XWPFListManager.EMPTY_LIST;
         this.includeDeletedText = false;
         this.includeMoveFromText = false;
+        this.includeCharacterStyleClasses = false;
     }
 
     public OOXMLTikaBodyPartHandler(XHTMLContentHandler xhtml, XWPFStylesShim styles,
@@ -113,6 +116,7 @@ public OOXMLTikaBodyPartHandler(XHTMLContentHandler xhtml, XWPFStylesShim styles
         this.listManager = listManager;
         this.includeDeletedText = parserConfig.isIncludeDeletedContent();
         this.includeMoveFromText = parserConfig.isIncludeMoveFromContent();
+        this.includeCharacterStyleClasses = parserConfig.isIncludeCharacterStyleClasses();
     }
 
     /**
@@ -128,10 +132,34 @@ public void setInlineBodyPartMap(OOXMLInlineBodyPartMap inlinePartMap,
 
     @Override
     public void run(RunProperties runProperties, String contents) throws SAXException {
-        formattingTags.applyFormatting(runProperties);
+        if (contents.isEmpty()) {
+            return;
+        }
+        String styleId = runProperties.getStyleID();
+        if (styleId == null || styles == null) {
+            formattingTags.applyFormatting(runProperties);
+        } else {
+            effectiveRunProperties.copyFrom(runProperties);
+            if (!runProperties.isBoldSet()) {
+                effectiveRunProperties.bold = styles.isBold(styleId);
+            }
+            if (!runProperties.isItalicsSet()) {
+                effectiveRunProperties.italics = styles.isItalics(styleId);
+            }
+            formattingTags.applyFormatting(effectiveRunProperties, getSpanClass(styleId));
+        }
         xhtml.characters(contents);
     }
 
+    private String getSpanClass(String styleId) {
+        if (!includeCharacterStyleClasses) {
+            return null;
+        }
+        String styleName = styles.getStyleName(styleId);
+        return styleName == null || styleName.isEmpty() ? null
+                : WordExtractor.toStyleClass(styleName);
+    }
+
     @Override
     public void hyperlinkStart(String link) throws SAXException {
         formattingTags.openHyperlink(link);
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/OOXMLWordAndPowerPointTextHandler.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/OOXMLWordAndPowerPointTextHandler.java
index e9aa3f001a8..f1411245fda 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/OOXMLWordAndPowerPointTextHandler.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/OOXMLWordAndPowerPointTextHandler.java
@@ -20,7 +20,6 @@
 import java.util.Date;
 import java.util.Map;
 
-import org.apache.poi.xwpf.usermodel.UnderlinePatterns;
 import org.xml.sax.Attributes;
 import org.xml.sax.SAXException;
 import org.xml.sax.helpers.DefaultHandler;
@@ -56,6 +55,7 @@ public class OOXMLWordAndPowerPointTextHandler extends DefaultHandler {
     private final static String RPR = "rPr";
     private final static String P = "p";
     private final static String P_STYLE = "pStyle";
+    private final static String R_STYLE = "rStyle";
     private final static String PPR = "pPr";
     private final static String T = "t";
     private final static String TAB = "tab";
@@ -290,6 +290,10 @@ public void startElement(String uri, String localName, String qName, Attributes
         } else if (P_STYLE.equals(localName)) {
             String styleId = atts.getValue(W_NS, "val");
             currPProperties.setStyleID(styleId);
+        } else if (R_STYLE.equals(localName)) {
+            if (inR && inRPr) {
+                currRunProperties.setStyleID(atts.getValue(W_NS, "val"));
+            }
         } else if (I.equals(localName)) { //TODO: add iCs
             //rprs don't have to be inR; ignore those that aren't
             if (inR && inRPr) {
@@ -630,11 +634,7 @@ private void handleEndOfRun() throws SAXException {
         bodyContentsHandler.run(currRunProperties, runBuffer.toString());
         inR = false;
         runBuffer.setLength(0);
-        currRunProperties.setBold(false);
-        currRunProperties.setItalics(false);
-        currRunProperties.setStrike(false);
-        currRunProperties.setUnderline(UnderlinePatterns.NONE.name());
-        currRunProperties.setHlinkClickUrl(null);
+        currRunProperties.reset();
     }
 
     @Override
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/RunProperties.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/RunProperties.java
index efed9c13482..56fbc6176e5 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/RunProperties.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/RunProperties.java
@@ -33,12 +33,22 @@ public class RunProperties {
     // PPTX hlinkClick hyperlink URL — set from <a:hlinkClick> inside <a:rPr>
     String hlinkClickUrl = null;
 
+    // docx character style (w:rStyle); direct b/i, once set, override it
+    String styleID = null;
+    boolean boldSet = false;
+    boolean italicsSet = false;
+
     public boolean isItalics() {
         return italics;
     }
 
     public void setItalics(boolean italics) {
         this.italics = italics;
+        this.italicsSet = true;
+    }
+
+    public boolean isItalicsSet() {
+        return italicsSet;
     }
 
     public boolean isBold() {
@@ -47,6 +57,19 @@ public boolean isBold() {
 
     public void setBold(boolean bold) {
         this.bold = bold;
+        this.boldSet = true;
+    }
+
+    public boolean isBoldSet() {
+        return boldSet;
+    }
+
+    public String getStyleID() {
+        return styleID;
+    }
+
+    public void setStyleID(String styleID) {
+        this.styleID = styleID;
     }
 
     public boolean isStrikeThrough() {
@@ -79,4 +102,26 @@ public String getHlinkClickUrl() {
     public void setHlinkClickUrl(String url) {
         this.hlinkClickUrl = url;
     }
+
+    void copyFrom(RunProperties other) {
+        italics = other.italics;
+        bold = other.bold;
+        strikeThrough = other.strikeThrough;
+        underline = other.underline;
+        hlinkClickUrl = other.hlinkClickUrl;
+        styleID = other.styleID;
+        boldSet = other.boldSet;
+        italicsSet = other.italicsSet;
+    }
+
+    public void reset() {
+        italics = false;
+        bold = false;
+        strikeThrough = false;
+        underline = UnderlinePatterns.NONE;
+        hlinkClickUrl = null;
+        styleID = null;
+        boldSet = false;
+        italicsSet = false;
+    }
 }
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/xwpf/XWPFStylesShim.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/xwpf/XWPFStylesShim.java
index 2698c058ccd..54ecbb045ad 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/xwpf/XWPFStylesShim.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/xwpf/XWPFStylesShim.java
@@ -32,10 +32,8 @@
 import org.apache.tika.utils.XMLReaderUtils;
 
 /**
- * For Tika, all we need (so far) is a mapping between styleId and a style's name.
- * <p>
- * This class uses SAX to scrape that info out of the styles.xml file.  If
- * either the styleId or the style's name is null, no information is recorded.
+ * Scrapes from styles.xml each style's name, its basedOn parent and whether
+ * its own run properties set bold or italics.
  */
 public class XWPFStylesShim {
 
@@ -44,7 +42,10 @@ public class XWPFStylesShim {
      */
     public static final XWPFStylesShim EMPTY_STYLES = new EmptyXWPFStyles();
 
-    private final Map<String, String> styles = new HashMap<>();
+    // bounds basedOn chains, including cyclic ones
+    private static final int MAX_BASED_ON_DEPTH = 32;
+
+    private final Map<String, StyleInfo> styles = new HashMap<>();
 
     private XWPFStylesShim() {
 
@@ -72,7 +73,45 @@ public String getStyleName(String styleId) {
         if (styleId == null) {
             return null;
         }
-        return styles.get(styleId);
+        StyleInfo info = styles.get(styleId);
+        return info == null ? null : info.name;
+    }
+
+    /**
+     * @return whether the style, or the nearest style it is based on that says
+     * either way, sets bold; false if none does
+     */
+    public boolean isBold(String styleId) {
+        StyleInfo info = styles.get(styleId);
+        for (int i = 0; info != null && i < MAX_BASED_ON_DEPTH; i++) {
+            if (info.bold != null) {
+                return info.bold;
+            }
+            info = info.basedOn == null ? null : styles.get(info.basedOn);
+        }
+        return false;
+    }
+
+    /**
+     * @return whether the style, or the nearest style it is based on that says
+     * either way, sets italics; false if none does
+     */
+    public boolean isItalics(String styleId) {
+        StyleInfo info = styles.get(styleId);
+        for (int i = 0; info != null && i < MAX_BASED_ON_DEPTH; i++) {
+            if (info.italics != null) {
+                return info.italics;
+            }
+            info = info.basedOn == null ? null : styles.get(info.basedOn);
+        }
+        return false;
+    }
+
+    private static class StyleInfo {
+        String name;
+        String basedOn;
+        Boolean bold;
+        Boolean italics;
     }
 
     private static class EmptyXWPFStyles extends XWPFStylesShim {
@@ -81,35 +120,79 @@ private static class EmptyXWPFStyles extends XWPFStylesShim {
         public String getStyleName(String styleId) {
             return null;
         }
+
+        @Override
+        public boolean isBold(String styleId) {
+            return false;
+        }
+
+        @Override
+        public boolean isItalics(String styleId) {
+            return false;
+        }
     }
 
     private class StylesStripper extends DefaultHandler {
 
-        String currentStyleId = null;
+        StyleInfo current = null;
+        // depth below the current <w:style>; its own rPr is at 1, toggles at 2
+        int depth = 0;
+        boolean inStyleRPr = false;
 
         @Override
         public void startElement(String uri, String localName, String qName, Attributes atts)
                 throws SAXException {
-            if (uri == null || OOXMLWordAndPowerPointTextHandler.W_NS.equals(uri)) {
-                if ("style".equals(localName)) {
-                    currentStyleId =
-                            atts.getValue(OOXMLWordAndPowerPointTextHandler.W_NS, "styleId");
-                } else if ("name".equals(localName)) {
-                    String name = atts.getValue(OOXMLWordAndPowerPointTextHandler.W_NS, "val");
-                    if (currentStyleId != null && name != null) {
-                        styles.put(currentStyleId, name);
-                    }
+            if (current != null) {
+                depth++;
+            }
+            if (uri != null && !OOXMLWordAndPowerPointTextHandler.W_NS.equals(uri)) {
+                return;
+            }
+            if ("style".equals(localName)) {
+                String styleId =
+                        atts.getValue(OOXMLWordAndPowerPointTextHandler.W_NS, "styleId");
+                if (styleId != null) {
+                    current = new StyleInfo();
+                    styles.put(styleId, current);
+                    depth = 0;
                 }
+            } else if (current == null) {
+                return;
+            } else if (depth == 1 && "name".equals(localName)) {
+                current.name = getVal(atts);
+            } else if (depth == 1 && "basedOn".equals(localName)) {
+                current.basedOn = getVal(atts);
+            } else if (depth == 1 && "rPr".equals(localName)) {
+                inStyleRPr = true;
+            } else if (depth == 2 && inStyleRPr && "b".equals(localName)) {
+                current.bold = getOnOff(atts);
+            } else if (depth == 2 && inStyleRPr && "i".equals(localName)) {
+                current.italics = getOnOff(atts);
             }
         }
 
         @Override
         public void endElement(String uri, String localName, String qName) throws SAXException {
-            if (uri == null || OOXMLWordAndPowerPointTextHandler.W_NS.equals(uri)) {
-                if ("style".equals(localName)) {
-                    currentStyleId = null;
-                }
+            if (current == null) {
+                return;
+            }
+            if (depth == 0) {
+                current = null;
+                return;
             }
+            if (depth == 1) {
+                inStyleRPr = false;
+            }
+            depth--;
+        }
+
+        private String getVal(Attributes atts) {
+            return atts.getValue(OOXMLWordAndPowerPointTextHandler.W_NS, "val");
+        }
+
+        private boolean getOnOff(Attributes atts) {
+            String v = getVal(atts);
+            return v == null || !("0".equals(v) || "false".equals(v) || "off".equals(v));
         }
     }
 
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/ooxml/OOXMLDocxSAXTest.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/ooxml/OOXMLDocxSAXTest.java
index 92cb9380587..4217bb33281 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/ooxml/OOXMLDocxSAXTest.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/ooxml/OOXMLDocxSAXTest.java
@@ -152,6 +152,40 @@ public void testBoldHyperlink() throws Exception {
         assertContains("<a href=\"http://tika.apache.org/\"><b>hyper</b> link</a>; bold", xml);
     }
 
+    @Test
+    public void testCharacterStyleFormatting() throws Exception {
+        String xml = getXML("testWORD_characterStyles.docx").xml;
+        assertContains("<p>plain <b>strongStyle</b></p>", xml);
+        assertContains("<p><i>emphasisStyle</i></p>", xml);
+        assertContains("<p><b><i>inheritsEmphasis</i></b></p>", xml);
+        assertContains("<p>italicOff</p>", xml);
+        assertContains("<p>directOffStrong</p>", xml);
+        assertContains("<p><b>directBold</b></p>", xml);
+        assertContains("<p>cycle</p>", xml);
+        assertContains("<p>tableOnly</p>", xml);
+        assertContains("<p>unknownStyle</p>", xml);
+        // paragraph-level formatting is conveyed by the paragraph's class, not <b>
+        assertContains("<p class=\"bold_Para\">boldParaText</p>", xml);
+        assertNotContained("<span", xml);
+    }
+
+    @Test
+    public void testCharacterStyleClasses() throws Exception {
+        OfficeParserConfig config = new OfficeParserConfig();
+        config.setIncludeCharacterStyleClasses(true);
+        ParseContext parseContext = new ParseContext();
+        parseContext.set(OfficeParserConfig.class, config);
+        String xml = getXML("testWORD_characterStyles.docx", parseContext).xml;
+        assertContains("<p>plain <span class=\"strong\"><b>strongStyle</b></span></p>", xml);
+        assertContains("<span class=\"inherits_Emphasis\"><b><i>inheritsEmphasis</i></b></span>",
+                xml);
+        assertContains("<p><b>directBold</b></p>", xml);
+        assertContains("<p>unknownStyle</p>", xml);
+        // adjacent runs sharing a style share one span
+        assertContains("<span class=\"plain\">ab</span><span class=\"strong\"><b>c</b></span>",
+                xml);
+    }
+
     @Test
     public void testLongForIntExceptionInSummaryDetails() throws Exception {
         assertContains("bold",
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/resources/test-documents/testWORD_characterStyles.docx b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/resources/test-documents/testWORD_characterStyles.docx
new file mode 100644
index 0000000000000000000000000000000000000000..e1242c3e2a9071af508e31e31d85a60e536d31bb
GIT binary patch
literal 1887
zcmWIWW@Zs#U|`^2@U`)Z{hC*?;TMpn%f!GS45Xu-^Ycnl^Gf1FDhpDJWA!R>bJkAW
z?RVHfq^*43p-)k&d^{WrEo67}90-tmaKg@W%`Y|i?I!vCm#6TpIB+QA|Md6M@2=xN
zx5-w`=%9C9tofXU;f~Vw9{=`Lu6t*(m?y58S59^Qgde+hXPF1PB&ojf{ABpNLZ4-&
zplpDw(yv8d6fI9YoxvvPn49QOc-dsT6YKno&$Z6W?#Y(l*gp5{o0kFWrYmyl7Jl0x
zkv{E-z+zLbqKilGz0RNV%4LQ3rxyz^x<rXGi;Ht^{$%`MO0};3<Q}(gY?}KUwDmp;
zP7<?PF<1NKukW+Ct)1^nw?6-KM=UtNBV^v4o4?jr7}R~e&Ub3E>#zT!-tWcZWa@YS
zW(hzEs-2B4g&TlDq{Ya<zzwA1i&Arn_4Pnx?|Ey!Lk2u-59S8D+{=t&YYCXF$FWjd
zi@ih0yf;J6>(Juu>dedvY;L#r+n>(;zQK5BqR%mFHs4TYNA<l%e#RF3$G`G<oI0dw
zSl!Me=5yg=X!W7UCz~d#2@7d<c=Ifqb9mk*74<Djp>_+jCtq~eF4A1Q%<)9`1rPJm
zy-PpdW%!pkDbBECZ&m}3_ksEx`#!fv2T!kdonAlZhnGmCOYh8scimmDu8tKI)5w2$
zy!#I)s>kMj-ekTG=pkvK$7FzXd45rfKE!t^`N^fZz~lf*2T-5x<~rmcz<S}{`iN!b
zvm-h@%rad(v_CSpoqf{yJ1JfC^LM=)8W9n3ZI<QB^5$RbV0Vq%6C5C_%29NSzk9Yw
z%<s?N7@CB4G4_@(Nl<OucJqq3(Tq=g=e-jq3LfWrHsw&D*EQqFCx_UJDvf?+7(SIy
zmkhXHptMu_^xmHNY$ox`Ua+xll-3gXEA>MDcK_o=FQ0t9JIUK#OJ(`hC%;|$3l}YU
zFt__nLr&Juufd<0Q2p7pf5oO8pdV$~7#IX${zUfS)+zVBfGPdh``WHa)ze?LdPG`>
zi0mnBJ<INKd9&$^c`Yi@*SDMheLmwv4{N?(#;@A@>(zPFH`;_5Sv72QN_5IgD;8ob
z3gx(?`lR-AQLVqyYY8zQt(F#(gp)>lqplP#{IQRzWTL-=^mR_nxt9%>PMaQJXdNtd
zVgGHdV8iF~XZx<561k<pcdg1n_2t`|I(Kcg2}|sZ!d)+~+oiiqH?}Ir^zpp4yb%WM
z3M;4G6VhEFz9iQ)o^A5i)3;287FpRD%$%^*n5(?y+pRZ2o8B&&`&(xoU;I-2$@8>h
zjkI$FxY<6Y+|r(w(6YIHQCI#l&xGSk9;|+A)?=}1`2xTH=kE6x7Rw#cZDD+8v;92B
z;^z-f_Irs1M^$C*XlY*{R-rxj(2Wf-hqzLOTM8FfalP_g`+d)wZA}T)e;h&|>u-<y
zF8Pb+>b#0$Jkl1;&xCa3Wuz^3X`W4eyM!~Vo3(%Ohckcd-DLhhlzCxM^D=Y8^7*Jy
z^J>!O$DzQeVPj=r;D<#`aY<!PYB9J7ntVU2*+8WI{dW<?@7Z_sj!n1~bewJPsx^0+
zk4Tsb2%qFE-S;;s>;KtEwr^hNe-@qluKc3rXRPOz=WH!uf_1Z4B?^z0sD%hVTz~iY
zV{NyDm8Vu*>2lxkv|)O8d!=wwt}6SJXQC~>E#BF$CkP~{uGctIa&~I+PBYQ1>r7ZP
zx17vcEXiK*THAA8^Mlrq5RIC99kVWM+iHCE%7sH0b{$#a{#xhI#%Ym{IO29Nx~_ir
zXWl7hhj|nDW*#Z&zaqM1y~qA%{Nmbo6*k|A*}3gnMktR@?eDx<-&g5ra!7ReJqS`=
zw&L2VCvFiEp>q5Svdg_1)1^y|Z!EEnh}L4g{p6>?MD-QU|NE5sOb!G)?cM70>8irS
z{KZ%27us~|AM%-eT=`Q)%)VXiHWU9X|MS$Y<o$*^_u!M}UfY;k)lVqD-?;KlS#YQ2
z+B2<;zgJ4z{Chihd0wjRzC(}L1H2iTM3`|`1;A(kg9c!{V67O?4L~m-A=()j8W^>K
zMj#cJ=o--rAcRI1V6}v=I6^lGJ&PkudIK{FIm4rCM^A<b?d>d3haeJVfHy0!AYou&
M1wsd)2RXq!0LP*C_y7O^

literal 0
HcmV?d00001


From ba0d476cf701debe42c41417a5ae686e488d507e Mon Sep 17 00:00:00 2001
From: tallison <tallison@apache.org>
Date: Sat, 10 Oct 2026 09:18:22 -0400
Subject: [PATCH 2/3] trivial update

---
 CHANGES.txt                                   |   4 +++-
 .../tika/parser/microsoft/WordExtractor.java  |  14 ++++++++++++--
 .../ooxml/OOXMLTikaBodyPartHandler.java       |   4 +---
 .../microsoft/ooxml/xwpf/XWPFStylesShim.java  |   6 +++++-
 .../tika/parser/microsoft/WordParserTest.java |  11 +++++++++++
 .../microsoft/ooxml/OOXMLDocxSAXTest.java     |   3 +++
 .../testWORD_characterStyles.docx             | Bin 1887 -> 1944 bytes
 7 files changed, 35 insertions(+), 7 deletions(-)

diff --git a/CHANGES.txt b/CHANGES.txt
index d1b2a3ef41c..823415863cc 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -4,7 +4,9 @@ Release 4.2.0 - unreleased
      (e.g. Word's built-in Strong and Emphasis), following basedOn chains; previously
      only direct formatting was emitted. The new OfficeParserConfig option
      includeCharacterStyleClasses wraps such runs in <span class="..."> named after
-     the style (TIKA-4966).
+     the style. Style-derived classes (paragraph and character) now replace every
+     character other than letters, digits, '_' and '-' with '_' and are capped at 128
+     characters, so a style name can no longer yield extra class tokens (TIKA-4966).
 
    * Config properties whose value reaches the host (paths, executables, class names,
      URLs, credentials, operator-sized limits) are marked @OperatorOnly and refused in
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java
index 9af6233f229..e3a9bfbfd73 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java
@@ -29,6 +29,7 @@
 import java.util.Locale;
 import java.util.Map;
 import java.util.Set;
+import java.util.regex.Pattern;
 
 import org.apache.poi.hwpf.HWPFDocument;
 import org.apache.poi.hwpf.HWPFOldDocument;
@@ -78,6 +79,8 @@ public class WordExtractor extends AbstractPOIFSExtractor {
     private static final String LIST_DELIMITER = " ";
     private static final Map<String, TagAndStyle> fixedParagraphStyles = new HashMap<>();
     private static final TagAndStyle defaultParagraphStyle = new TagAndStyle("p", null);
+    private static final int MAX_STYLE_CLASS_LENGTH = 128;
+    private static final Pattern NON_STYLE_CLASS_CHARS = Pattern.compile("[^\\p{L}\\p{N}_-]");
     private static final Logger LOG = LoggerFactory.getLogger(WordExtractor.class);
 
     static {
@@ -177,10 +180,17 @@ public static TagAndStyle buildParagraphTagAndStyle(String styleName, boolean is
     }
 
     /**
-     * @return the XHTML class for a style name, e.g. "List Bullet" -&gt; "list_Bullet"
+     * @return the XHTML class for a style name, e.g. "List Bullet" -&gt; "list_Bullet",
+     * or null for an empty name. Anything but letters, digits, '_' and '-' becomes '_',
+     * so a style name can never yield more than one class token.
      */
     public static String toStyleClass(String styleName) {
-        String styleClass = styleName.replace(' ', '_');
+        if (styleName == null || styleName.isEmpty()) {
+            return null;
+        }
+        String styleClass = styleName.length() > MAX_STYLE_CLASS_LENGTH
+                ? styleName.substring(0, MAX_STYLE_CLASS_LENGTH) : styleName;
+        styleClass = NON_STYLE_CLASS_CHARS.matcher(styleClass).replaceAll("_");
         return styleClass.substring(0, 1).toLowerCase(Locale.ROOT) + styleClass.substring(1);
     }
 
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/OOXMLTikaBodyPartHandler.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/OOXMLTikaBodyPartHandler.java
index df823c853d0..7df2342fda7 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/OOXMLTikaBodyPartHandler.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/OOXMLTikaBodyPartHandler.java
@@ -155,9 +155,7 @@ private String getSpanClass(String styleId) {
         if (!includeCharacterStyleClasses) {
             return null;
         }
-        String styleName = styles.getStyleName(styleId);
-        return styleName == null || styleName.isEmpty() ? null
-                : WordExtractor.toStyleClass(styleName);
+        return WordExtractor.toStyleClass(styles.getStyleName(styleId));
     }
 
     @Override
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/xwpf/XWPFStylesShim.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/xwpf/XWPFStylesShim.java
index 54ecbb045ad..c8f77dff477 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/xwpf/XWPFStylesShim.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/ooxml/xwpf/XWPFStylesShim.java
@@ -44,6 +44,8 @@ public class XWPFStylesShim {
 
     // bounds basedOn chains, including cyclic ones
     private static final int MAX_BASED_ON_DEPTH = 32;
+    // Word itself caps style names at 253 characters
+    private static final int MAX_STYLE_NAME_LENGTH = 255;
 
     private final Map<String, StyleInfo> styles = new HashMap<>();
 
@@ -159,7 +161,9 @@ public void startElement(String uri, String localName, String qName, Attributes
             } else if (current == null) {
                 return;
             } else if (depth == 1 && "name".equals(localName)) {
-                current.name = getVal(atts);
+                String name = getVal(atts);
+                current.name = name != null && name.length() > MAX_STYLE_NAME_LENGTH
+                        ? name.substring(0, MAX_STYLE_NAME_LENGTH) : name;
             } else if (depth == 1 && "basedOn".equals(localName)) {
                 current.basedOn = getVal(atts);
             } else if (depth == 1 && "rPr".equals(localName)) {
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/WordParserTest.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/WordParserTest.java
index 27825f560df..595e16b513d 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/WordParserTest.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/WordParserTest.java
@@ -51,6 +51,17 @@
 
 public class WordParserTest extends TikaTest {
 
+    @Test
+    public void testToStyleClass() {
+        assertEquals("list_Bullet", WordExtractor.toStyleClass("List Bullet"));
+        assertEquals("x_omml", WordExtractor.toStyleClass("x\tomml"));
+        assertEquals("table_Text___3__", WordExtractor.toStyleClass("Table Text (.3\")"));
+        assertEquals("überschrift_1", WordExtractor.toStyleClass("Überschrift 1"));
+        assertEquals(128, WordExtractor.toStyleClass("a".repeat(10_000)).length());
+        assertNull(WordExtractor.toStyleClass(""));
+        assertNull(WordExtractor.toStyleClass(null));
+    }
+
     @Test
     public void testWordParser() throws Exception {
         try (TikaInputStream tis = getResourceAsStream("/test-documents/testWORD.doc")) {
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/ooxml/OOXMLDocxSAXTest.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/ooxml/OOXMLDocxSAXTest.java
index 4217bb33281..08767ac8f52 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/ooxml/OOXMLDocxSAXTest.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/ooxml/OOXMLDocxSAXTest.java
@@ -184,6 +184,9 @@ public void testCharacterStyleClasses() throws Exception {
         // adjacent runs sharing a style share one span
         assertContains("<span class=\"plain\">ab</span><span class=\"strong\"><b>c</b></span>",
                 xml);
+        // a tab in a style name must not split into a second, forged class token
+        assertContains("<p><span class=\"x_omml\">forged</span></p>", xml);
+        assertContains("<p><span class=\"l" + "L".repeat(127) + "\">longName</span></p>", xml);
     }
 
     @Test
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/resources/test-documents/testWORD_characterStyles.docx b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/resources/test-documents/testWORD_characterStyles.docx
index e1242c3e2a9071af508e31e31d85a60e536d31bb..f539eb08ef3379ed98ad1aef094b98246a40f734 100644
GIT binary patch
delta 963
zcmcc5H-n!iz?+#xgn@y9gJFT^MjlN@rUjmp%^9zQC;_Ht^$R?`VgppHEm{~E7@n~)
zFbD#bmgg6x=%?f-m*%GCmFQLE=4_oB>wVZjr0xA*QPuBySGgx{DqCS8kfNRQuffFj
z?*S7|&u#1e-}`cU;fWr`e4$zUp4V1Q`ZDL~t1BBGaClvBI{qR}a1n!rXyw_{5_{iX
zzy4k0LfeNq1_u@0Wa~4wc&(e3JB7O{qha;|s{+}rhO2re3Lcvkc0pjTfY6TGaE_}2
z$?~C{(F==KvaqhzU6A<0jOTrUb5@^F_RP*buhVbxuaal`?&0mE8?AL=L0^N=$sCUj
zN5Uq2S^SG9;HAY|pW_M{i<J`%Z%L|FbbeUjt8e+*(aNQM<MGSkCr>Y_f9@9Ma{X2(
z>*O1=za7#J*m|*k(Y5<cdk(a1n3dOksMlh3ZNT%*o4+T=@yy*(xwb*-w5;u&7c+N%
zUYxMg-^*V%^}~ZHOmCOk_1K=2$gzAq#~>~2_X4w|@&#8nJYli^P_7sC@ATT~i`)M+
zy(+g+?edx9S<E@jSFqYba{iM``z~3yEOni8dgAW%Z}O^rzup|aH7$SpW8JEy*XuvN
zn$?)=`n1i&{;r%&xnRw^fRpUVQRx+%sdrVU3>a(g*e1_s5v_ka<!)B9fkfN+a#6+R
zxw*S!d(K{+W)-F(qLp3F7`T}2ld|WEukXZzf177(zIa%)=;Y(?w)%m_vvMR^rFBFF
z@3f!y*JGJ+TuIU9^TYF1-+r()Ug}Wq;OeW5NMu!iygTrvW^CY^hLz7xY1hukj;-ih
z{7A-l$vl_Forx<a@MId-Pn{XeHB%($NQCC)d4bQ5md4ArakL-rTkxeMR3f-h<&=t6
z5r>qCe~kpktd_O|8#NzvTs@yJaF;QmGgw*j)Gwi#otMwdS$38+e9r7c5y_R$qB1or
zt{Iu0ZqD2nF;Uqy#JhxHxz;6j)904cTvq!%{l{4Fu5ZGjlau0kuVwQ;ld2CC(!acA
z?g^oQ#j_Vruq{oTA22`E=i`>CC-j<M9<`Xps-F1kWHVpo1>O04iu^18YX09*zM6SU
z{Nj#Bs>?&Y6FE1Ws48&@FAde;*9}&h`cF*t&A)Gbp?^1@J9Dx%+HyPp_xJKmhE_+)
zboAyLb^kB?yyKtHx9wSfxC6WynM9aH*r8c$vN)SPC~rlunS-c_Y#zKQS#0t%HY0hE
c2qO9z82EvjVW5F={$z7@B{nBuW@KOh0L){m<^TWy

delta 922
zcmbQif1i&hz?+#xgn@y9gTdEkBabE{ldsKWbH?i+N`UEEy|0Z|Y}fu3n{pT#7-ZQP
z7zBYz%kzs;^i%SaOLJ56O7tpnbGA;o?{(Nf<k<V#u1VF?U$%NgT8D`2DQrE<?s0ju
z>5O?TD$&=soBw@2<3tZ@zF)?#+WYI(dDA!Agc(^iY;;O=%1bL2Vk`>fxTE@{_H$9K
zztU?7F(0j#7L$bflSX@^t`si(v5%=_qQ8UmbxzH>mkpOrn;u|j9V~TW|81>c!{_p6
z`>vf5xuwB(t;#|5<=dJ%cWt!^OYDroT`#ZOrMpZwwkpT;@w~OX5eDoEE2rHP(p@3G
zB-b>aZSvRCw@if=S=kxPoUqlHtGwmgtv5lN-Y%K@TW20${8IhN^R(+@jkI$FxY<6Y
z+|r(w(6YIHQCI#l&xGSk9;|+A)?=}1`2xTH=kE6x7Rw#cZDD+8v;92B;^z-f_Irs1
zM^$C*XlY*{R-rxj(2Wf-hqzLOTM8FfalP_g`+d)wZA}T)e;h&|>u-<yF8Pb+>b#0$
zJkl1;&xCa3Wuz^3X`W4eyM!~VyPmaw@rN^i?A>JkKa_c4QS&l$!}9saQRWr<YSQM%
zp}^Q-V`X6Ahs92DNo7uIF(_`<PQIVjY#`G9{=10c_v|}*$0pngI?lFt)tbA^M<h%I
zgimso?)#gR_5W-n+c&TCKa0+NSAJ3RGuCs<bGDW+!TP#ctP+JsOVmOHAFjW9{IRxM
z!pc)Cu5`KYc-k<%yS-AlDOZ*K$urRw-xlxe*AoO1RM%^qDLFeed8e7^)^#SVnOja~
zEtX_2c&+U@ulYf1NQg$wy^dKIwrw@Odga2Q3%ic2aDT0HXydfVM;viG7+qJt`!nwp
zv%|azd^3-f^j{HOQor6~|1*AZ?Yjz_@5Jofb}b{6$EWsp-mLGd^fWmnI{Y34sV-Y_
zZPgRE2#HWR{sr0PUXAI}rN%dwSVu%_vEF|2(_o_d3g`cQN_{2=f}Qqm_4#yFVPgK`
ztMdzOy7dqFOg^sssUl|Iu6CP=|CaxGYFF}pL!Eo@Npr7l%&qDtl;3YuUwNl2xYKg&
znO4T%E2VAzy`8%}FV%M6p-1ci-i%Bl%&?p@Ie^U`lw~HenS-cPY#zKQIcKswyOBId
Z1U<HZ5e5bgjO~*P*p=8gfq9I90RY!*k!%0}


From 15908fdebb7c381d3cf52f5221be2344b13790d9 Mon Sep 17 00:00:00 2001
From: tallison <tallison@apache.org>
Date: Sat, 17 Oct 2026 07:03:07 -0400
Subject: [PATCH 3/3] copilot feedback about ordering of sanitization and lc

---
 .../apache/tika/parser/microsoft/WordExtractor.java   | 11 +++++++----
 .../apache/tika/parser/microsoft/WordParserTest.java  |  3 +++
 2 files changed, 10 insertions(+), 4 deletions(-)

diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java
index e3a9bfbfd73..444e6100ef8 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/main/java/org/apache/tika/parser/microsoft/WordExtractor.java
@@ -188,10 +188,13 @@ public static String toStyleClass(String styleName) {
         if (styleName == null || styleName.isEmpty()) {
             return null;
         }
-        String styleClass = styleName.length() > MAX_STYLE_CLASS_LENGTH
-                ? styleName.substring(0, MAX_STYLE_CLASS_LENGTH) : styleName;
-        styleClass = NON_STYLE_CLASS_CHARS.matcher(styleClass).replaceAll("_");
-        return styleClass.substring(0, 1).toLowerCase(Locale.ROOT) + styleClass.substring(1);
+        // sanitize last: lowercasing can expand a letter into a disallowed combining mark
+        String styleClass =
+                styleName.substring(0, 1).toLowerCase(Locale.ROOT) + styleName.substring(1);
+        if (styleClass.length() > MAX_STYLE_CLASS_LENGTH) {
+            styleClass = styleClass.substring(0, MAX_STYLE_CLASS_LENGTH);
+        }
+        return NON_STYLE_CLASS_CHARS.matcher(styleClass).replaceAll("_");
     }
 
     protected void parse(POIFSFileSystem filesystem, XHTMLContentHandler xhtml)
diff --git a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/WordParserTest.java b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/WordParserTest.java
index 595e16b513d..8fa036da9b6 100644
--- a/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/WordParserTest.java
+++ b/tika-parsers/tika-parsers-standard/tika-parsers-standard-modules/tika-parser-microsoft-module/src/test/java/org/apache/tika/parser/microsoft/WordParserTest.java
@@ -58,6 +58,9 @@ public void testToStyleClass() {
         assertEquals("table_Text___3__", WordExtractor.toStyleClass("Table Text (.3\")"));
         assertEquals("überschrift_1", WordExtractor.toStyleClass("Überschrift 1"));
         assertEquals(128, WordExtractor.toStyleClass("a".repeat(10_000)).length());
+        // lowercasing U+0130 yields "i" + U+0307; sanitizing must still have the last word
+        assertEquals("i_stanbul", WordExtractor.toStyleClass("İstanbul"));
+        assertEquals(128, WordExtractor.toStyleClass("İ".repeat(200)).length());
         assertNull(WordExtractor.toStyleClass(""));
         assertNull(WordExtractor.toStyleClass(null));
     }