Fix text truncation in reply quotes and Twitter posts

4 files changed, 35 insertions, 3 deletions

diff --git a/src/main/java/com/juick/service/TwitterService.java b/src/main/java/com/juick/service/TwitterService.java
index d0c105f2..ffcaca13 100644
--- a/src/main/java/com/juick/service/TwitterService.java
+++ b/src/main/java/com/juick/service/TwitterService.java
@@ -27,6 +27,7 @@ import com.github.scribejava.core.model.Verb;
 import com.github.scribejava.core.oauth.OAuth20Service;
 import com.juick.model.Message;
 import com.juick.util.MessageUtils;
+import com.juick.util.formatters.PlainTextFormatter;
 import jakarta.annotation.PostConstruct;
 import org.apache.commons.lang3.StringUtils;
 import org.slf4j.Logger;
@@ -67,7 +68,7 @@ public class TwitterService {
         userService.getTwitterToken(jmsg.getUser().getUid()).ifPresent(t -> {
             String status = MessageUtils.getMessageHashTags(jmsg) + StringUtils.defaultString(jmsg.getText());
             if (status.length() > 253) {
-                status = status.substring(0, 252) + "…";
+                status = PlainTextFormatter.truncateText(status, 252) + "…";
             }
             status += " http://juick.com/" + jmsg.getMid();
             try {
diff --git a/src/main/java/com/juick/util/MessageUtils.java b/src/main/java/com/juick/util/MessageUtils.java
index a4407f17..308d144b 100644
--- a/src/main/java/com/juick/util/MessageUtils.java
+++ b/src/main/java/com/juick/util/MessageUtils.java
@@ -21,6 +21,7 @@ import com.juick.model.Message;
 import com.juick.model.Tag;
 import com.juick.model.User;
 import com.juick.model.Entity;
+import com.juick.util.formatters.PlainTextFormatter;
 import org.apache.commons.lang3.StringUtils;
 import org.springframework.web.util.UriComponentsBuilder;
 
@@ -52,7 +53,7 @@ public class MessageUtils {
 
         if (quote != null) {
             if (quote.length() > 50) {
-                result = ">" + quote.substring(0, 47).replace('\n', ' ') + "...\n";
+                result = ">" + PlainTextFormatter.truncateText(quote, 47).replace('\n', ' ') + "…\n";
             } else if (!quote.isEmpty()) {
                 result = ">" + quote.replace('\n', ' ') + "\n";
             }
diff --git a/src/main/java/com/juick/util/formatters/PlainTextFormatter.java b/src/main/java/com/juick/util/formatters/PlainTextFormatter.java
index 9a885bca..2e7722e7 100644
--- a/src/main/java/com/juick/util/formatters/PlainTextFormatter.java
+++ b/src/main/java/com/juick/util/formatters/PlainTextFormatter.java
@@ -22,6 +22,7 @@ import com.juick.util.MessageUtils;
 import org.apache.commons.lang3.StringUtils;
 import org.ocpsoft.prettytime.PrettyTime;
 
+import java.text.BreakIterator;
 import java.util.Date;
 import java.util.Locale;
 
@@ -62,7 +63,7 @@ public class PlainTextFormatter {
             sb.append(attachmentUrl).append("\n");
         }
         if (txt.length() >= cropLength) {
-            sb.append(StringUtils.substring(txt, 0, cropLength)).append(" [...]");
+            sb.append(PlainTextFormatter.truncateText(txt, cropLength)).append(" [...]");
         } else {
             sb.append(txt);
         }
@@ -97,4 +98,26 @@ public class PlainTextFormatter {
     public static String formatTwitterCard(Message jmsg) {
         return MessageUtils.getMessageHashTags(jmsg) + StringUtils.defaultString(jmsg.getText());
     }
+    /**
+     * Truncate text to the nearest word, up to a maximum length specified.
+     *
+     * @param text
+     * @param maxLength
+     * @return
+     */
+    public static String truncateText(String text, int maxLength) {
+        if(text != null && text.length() > maxLength) {
+            BreakIterator bi = BreakIterator.getWordInstance();
+            bi.setText(text);
+
+            if(bi.isBoundary(maxLength-1)) {
+                return text.substring(0, maxLength-2);
+            } else {
+                int preceding = bi.preceding(maxLength-1);
+                return text.substring(0, preceding-1);
+            }
+        } else {
+            return text;
+        }
+    }
 }
diff --git a/src/test/java/com/juick/server/tests/ServerTests.java b/src/test/java/com/juick/server/tests/ServerTests.java
index a728ac2b..d61725e6 100644
--- a/src/test/java/com/juick/server/tests/ServerTests.java
+++ b/src/test/java/com/juick/server/tests/ServerTests.java
@@ -2693,4 +2693,11 @@ public class ServerTests {
         .with(httpBasic(ugnichName, ugnichPassword)))
         .andExpect(status().isBadRequest());
     }
+    @Test
+    public void textTruncationShouldNotBreakEmojis() {
+        var text = "Так ты же написал, чтоб я сам поправил отступ \uD83D\uDE00\n";
+        var expected = "Так ты же написал, чтоб я сам поправил отступ";
+        var truncated = PlainTextFormatter.truncateText(text, 47);
+        assertThat(truncated, is(expected));
+    }
 }