[facebook] decode surrogate pairs in metadata values (#6599)

gallery_dl/extractor/facebook.py

@@ -40,7 +40,8 @@ class FacebookExtractor(Extractor):
     @staticmethod
     def decode_all(txt):
         return text.unescape(
-            txt.encode("utf-8").decode("unicode_escape")
+            txt.encode().decode("unicode_escape")
+            .encode("utf_16", "surrogatepass").decode("utf_16")
         ).replace("\\/", "/")
 
     @staticmethod

test/results/facebook.py

@@ -113,6 +113,15 @@ __tests__ = (
     "username" : "Facebook",
 },
 
+{
+    "#url"     : "https://www.facebook.com/photo.php?fbid=1156625586261770",
+    "#comment" : "surrogate pair in 'caption' data (#6599)",
+    "#category": ("", "facebook", "photo"),
+    "#class"   : facebook.FacebookPhotoExtractor,
+
+    "caption"  : "A century of innovation parked side by side.\n\n📸: Vocabutesla via X",
+},
+
 {
     "#url"     : "https://www.facebook.com/watch/?v=1165557851291824",
     "#category": ("", "facebook", "video"),