Merge Konflikt behoben bei merge von fix unicode encoding

mediathekview · Nov 23, 2021 · ab544f9 · ab544f9
2 parents cd77047 + 285cd06
commit ab544f9
Showing 1 changed file with 14 additions and 3 deletions.
diff --git a/src/main/java/de/mediathekview/mlib/daten/DatenFilm.java b/src/main/java/de/mediathekview/mlib/daten/DatenFilm.java
@@ -19,6 +19,8 @@
  */
 package de.mediathekview.mlib.daten;
 
+import java.text.Normalizer;
+import java.time.LocalDateTime;
 import java.time.format.DateTimeFormatter;
 
 import org.apache.commons.lang3.time.FastDateFormat;
@@ -132,14 +134,14 @@ public DatenFilm(String ssender, String tthema, String filmWebsite, String ttite
           long dauerSekunden, String description) {
     // da werden die gefundenen Filme beim Absuchen der Senderwebsites erstellt, und nur die!!
     arr[FILM_SENDER] = ssender;
-    arr[FILM_THEMA] = tthema.isEmpty() ? ssender : cleanWhitespaces(tthema.trim());
+    arr[FILM_THEMA] = tthema.isEmpty() ? ssender : normalize(cleanWhitespaces(tthema.trim()));
     setTitle(ttitel.isEmpty() ? tthema : ttitel.trim());
     arr[FILM_URL] = uurl;
     arr[FILM_URL_RTMP] = uurlRtmp;
     arr[FILM_WEBSEITE] = filmWebsite;
     checkDatum(datum, arr[FILM_SENDER] + ' ' + arr[FILM_THEMA] + ' ' + arr[FILM_TITEL]);
     checkZeit(arr[FILM_DATUM], zeit, arr[FILM_SENDER] + ' ' + arr[FILM_THEMA] + ' ' + arr[FILM_TITEL]);
-    arr[FILM_BESCHREIBUNG] = cleanDescription(description);
+    arr[FILM_BESCHREIBUNG] = normalize(cleanDescription(description));
 
     // Filmlänge
     checkFilmDauer(dauerSekunden);
@@ -150,7 +152,7 @@ private void setTitle(String title) {
       title = title.substring(0, title.indexOf(COPYRIGHT_CHAR_HTML));
       title = title.trim();
     }
-    arr[FILM_TITEL] = cleanWhitespaces(title);
+    arr[FILM_TITEL] = normalize(cleanWhitespaces(title));
   }
 
   /**
@@ -162,6 +164,15 @@ public void setFileSize() {
     }
   }
 
+  private static String normalize(String s) {
+    // some websites uses NFD normalization instead of NFC => normalize
+    if (s != null) {
+      return Normalizer.normalize(s, Normalizer.Form.NFC);
+    }
+
+    return null;
+  }
+
   public static String cleanWhitespaces(String text) {
     return text.replaceAll("[\\t\\n\\x0B\\f\\r]", "")
             .replace("\u00a0", " ") // repalce no-break-space with space