Author: msahyoun
Date: Thu Feb 19 20:31:03 2026
New Revision: 1931940
Log:
PDFBOX-5049: optimize recursion to break a word into box width
Modified:
pdfbox/branches/2.0/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/interactive/form/PlainText.java
Modified:
pdfbox/branches/2.0/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/interactive/form/PlainText.java
==============================================================================
---
pdfbox/branches/2.0/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/interactive/form/PlainText.java
Thu Feb 19 20:14:14 2026 (r1931939)
+++
pdfbox/branches/2.0/pdfbox/src/main/java/org/apache/pdfbox/pdmodel/interactive/form/PlainText.java
Thu Feb 19 20:31:03 2026 (r1931940)
@@ -25,6 +25,9 @@ import java.util.Collections;
import java.util.List;
import org.apache.pdfbox.pdmodel.font.PDFont;
+import org.apache.pdfbox.pdmodel.interactive.form.PlainText.Paragraph;
+import org.apache.pdfbox.pdmodel.interactive.form.PlainText.TextAttribute;
+import org.apache.pdfbox.pdmodel.interactive.form.PlainText.Word;
/**
* A block of text.
@@ -210,19 +213,26 @@ class PlainText
// single word does not fit into width
// PDFBOX-6082: at least 1 character
wordNeedsSplit = true;
- while (true)
- {
- splitOffset--;
- String substring = word.substring(0, splitOffset);
- float substringWidth = font.getStringWidth(substring)
* scale;
- if (substringWidth < width)
- {
- word = substring;
- wordWidth = font.getStringWidth(word) * scale;
- lineWidth = wordWidth;
- break;
- }
- }
+
+ // PDFBOX-5049: The original approach was to decrement
splitOffset
+ // until the substring fits, but this can be very
expensive for long words and
+ // narrow widths (e.g. a long URL in a narrow column).
+ //
+ // Optimization: instead of decrementing splitOffset one
step at a time and
+ // calling getStringWidth on progressively shorter
substrings:
+ // - compute the scaled width of every individual
character once
+ // - build a prefix-sum array
+ // - binary-search for the largest prefix that fits
+ //
+ // TODO: The special case in PDFBOX-5049 should be handled
by not generating an appearance
+ // stream at all as the the height of the text box is only
1pt and the text is not visible.
+
+ float[] prefixWidth = buildPrefixWidths(word, font, scale);
+ splitOffset = findMaxFittingChars(prefixWidth, width);
+
+ word = word.substring(0, splitOffset);
+ wordWidth = prefixWidth[splitOffset];
+ lineWidth = wordWidth;
}
AttributedString as = new AttributedString(word);
@@ -245,6 +255,72 @@ class PlainText
textLines.add(textLine);
return textLines;
}
+
+ /**
+ * Build the prefix-sum array of scaled character widths for the given
word.
+ *
+ * @param word the word to measure.
+ * @param font the font used to obtain glyph advance widths.
+ * @param scale {@code fontSize / FONTSCALE}, pre-computed by the
caller.
+ * @return the {@code float[word.length() + 1]} prefix-sum array.
+ * @throws IOException if the font cannot provide a glyph width.
+ */
+ private static float[] buildPrefixWidths(String word, PDFont font,
float scale)
+ throws IOException
+ {
+ int wordLen = word.length();
+ float[] prefixWidth = new float[wordLen + 1];
+ int i = 0;
+ while (i < wordLen)
+ {
+ int codePoint = word.codePointAt(i);
+ int charCount = Character.charCount(codePoint);
+ // Measure this code point as a single string (handles
surrogate pairs).
+ float cpWidth = font.getStringWidth(word.substring(i, i +
charCount)) * scale;
+ // Propagate the cumulative width across all Java chars of
this code point.
+ for (int j = 0; j < charCount; j++)
+ {
+ prefixWidth[i + j + 1] = prefixWidth[i + j] + (j == 0 ?
cpWidth : 0f);
+ }
+ i += charCount;
+ }
+ return prefixWidth;
+ }
+
+ /**
+ * Find the maximum number of Java chars from a prefix-width array
that fit
+ * within the given available width.
+ * <p>
+ * Binary search over the pre-computed {@code prefixWidth} array.
+ * The result is always at least {@code 1} so that the caller is
guaranteed to
+ * make forward progress (PDFBOX-6082).
+ * </p>
+ *
+ * @param prefixWidth array as returned by {@link #buildPrefixWidths};
length is
+ * {@code wordLength + 1}.
+ * @param width the available line width in the same unit as the
widths stored
+ * in {@code prefixWidth}.
+ * @return the largest index {@code k >= 1} such that {@code
prefixWidth[k] < width},
+ * or {@code 1} if even a single character exceeds the
available width.
+ */
+ private static int findMaxFittingChars(float[] prefixWidth, float
width)
+ {
+ int lo = 1;
+ int hi = prefixWidth.length - 1;
+ while (lo < hi)
+ {
+ int mid = (lo + hi + 1) >>> 1; // upper-mid to avoid infinite
loop
+ if (prefixWidth[mid] < width)
+ {
+ lo = mid;
+ }
+ else
+ {
+ hi = mid - 1;
+ }
+ }
+ return lo;
+ }
}
/**
