Author: centic
Date: Tue Mar 24 10:58:16 2020
New Revision: 1875571
URL: http://svn.apache.org/viewvc?rev=1875571&view=rev
Log:
Bug 64244: Take the replacement of RichtText strings into account when
computing length of strings
Modified:
poi/trunk/src/ooxml/java/org/apache/poi/xssf/usermodel/XSSFRichTextString.java
poi/trunk/src/ooxml/testcases/org/apache/poi/xssf/usermodel/TestXSSFRichTextString.java
Modified:
poi/trunk/src/ooxml/java/org/apache/poi/xssf/usermodel/XSSFRichTextString.java
URL:
http://svn.apache.org/viewvc/poi/trunk/src/ooxml/java/org/apache/poi/xssf/usermodel/XSSFRichTextString.java?rev=1875571&r1=1875570&r2=1875571&view=diff
==============================================================================
---
poi/trunk/src/ooxml/java/org/apache/poi/xssf/usermodel/XSSFRichTextString.java
(original)
+++
poi/trunk/src/ooxml/java/org/apache/poi/xssf/usermodel/XSSFRichTextString.java
Tue Mar 24 10:58:16 2020
@@ -201,7 +201,7 @@ public class XSSFRichTextString implemen
CTRElt lt = st.addNewR();
lt.setT(text);
preserveSpaces(lt.xgetT());
-
+
if (font != null) {
CTRPrElt pr = lt.addNewRPr();
setRunAttributes(font.getCTFont(), pr);
@@ -247,7 +247,7 @@ public class XSSFRichTextString implemen
}
/**
- * Does this string have any explicit formatting applied, or is
+ * Does this string have any explicit formatting applied, or is
* it just text in the default style?
*/
public boolean hasFormatting() {
@@ -496,6 +496,32 @@ public class XSSFRichTextString implemen
}
/**
+ * Optimized counting of actual length of a string
+ * considering the replacement of _xHHHH_ that needs
+ * to be applied to rich-text strings.
+ *
+ * @param value The string
+ * @return The length of the string, 0 if the string is null.
+ */
+ static int utfLength(String value) {
+ if(value == null) {
+ return 0;
+ }
+ if (!value.contains("_x")) {
+ return value.length();
+ }
+
+ Matcher matcher = utfPtrn.matcher(value);
+ int count = 0;
+ while (matcher.find()) {
+ count++;
+ }
+
+ // Length of pattern is 7 (_xHHHH_), and we replace it with one
character
+ return value.length() - (count * 6);
+ }
+
+ /**
* For all characters which cannot be represented in XML as defined by the
XML 1.0 specification,
* the characters are escaped using the Unicode numerical character
representation escape character
* format _xHHHH_, where H represents a hexadecimal character in the
character's value.
@@ -512,7 +538,7 @@ public class XSSFRichTextString implemen
if(value == null || !value.contains("_x")) {
return value;
}
-
+
StringBuilder buf = new StringBuilder();
Matcher m = utfPtrn.matcher(value);
int idx = 0;
@@ -528,13 +554,13 @@ public class XSSFRichTextString implemen
idx = m.end();
}
-
- // small optimization: don't go via StringBuilder if not necessary,
- // the encodings are very rare, so we should almost always go via this
shortcut.
+
+ // small optimization: don't go via StringBuilder if not necessary,
+ // the encodings are very rare, so we should almost always go via this
shortcut.
if(idx == 0) {
return value;
}
-
+
buf.append(value.substring(idx));
return buf.toString();
}
@@ -577,7 +603,7 @@ public class XSSFRichTextString implemen
String txt = r.getT();
CTRPrElt fmt = r.getRPr();
- length += txt.length();
+ length += utfLength(txt);
formats.put(length, fmt);
}
return formats;
@@ -605,7 +631,7 @@ public class XSSFRichTextString implemen
}
return stf;
}
-
+
private ThemesTable getThemesTable() {
if(styles == null) return null;
return styles.getTheme();
Modified:
poi/trunk/src/ooxml/testcases/org/apache/poi/xssf/usermodel/TestXSSFRichTextString.java
URL:
http://svn.apache.org/viewvc/poi/trunk/src/ooxml/testcases/org/apache/poi/xssf/usermodel/TestXSSFRichTextString.java?rev=1875571&r1=1875570&r2=1875571&view=diff
==============================================================================
---
poi/trunk/src/ooxml/testcases/org/apache/poi/xssf/usermodel/TestXSSFRichTextString.java
(original)
+++
poi/trunk/src/ooxml/testcases/org/apache/poi/xssf/usermodel/TestXSSFRichTextString.java
Tue Mar 24 10:58:16 2020
@@ -563,4 +563,29 @@ public final class TestXSSFRichTextStrin
assertEquals("Rich Text\r\nTest",
wb.getSheetAt(0).getRow(1).getCell(1).getRichStringCellValue().getString());
wb.close();
}
+
+ @Test
+ public void testUtfDecode_withApplyFont() {
+ XSSFFont font = new XSSFFont();
+ font.setBold(true);
+ font.setFontHeightInPoints((short) 14);
+
+ CTRst st = CTRst.Factory.newInstance();
+ st.setT("abc_x000D_2ef_x000D_");
+ XSSFRichTextString rt = new XSSFRichTextString(st);
+ rt.applyFont(font);
+ assertEquals("abc\r2ef\r", rt.getString());
+ }
+
+ @Test
+ public void testUtfLength() {
+ assertEquals(0, XSSFRichTextString.utfLength(null));
+ assertEquals(0, XSSFRichTextString.utfLength(""));
+
+ assertEquals(3, XSSFRichTextString.utfLength("abc"));
+ assertEquals(3, XSSFRichTextString.utfLength("ab_x0032_"));
+ assertEquals(3, XSSFRichTextString.utfLength("a_x0032__x0032_"));
+ assertEquals(3, XSSFRichTextString.utfLength("_x0032_a_x0032_"));
+ assertEquals(3, XSSFRichTextString.utfLength("_x0032__x0032_a"));
+ }
}
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
