Adjust zero-width joiner check to fix IndexError

There was the potential for an IndexError here, with a string that
started with a zero-width joiner and had at least one more character
afterwards.

tildes/tildes/lib/string.py

@@ -189,19 +189,18 @@ def _sanitize_characters(original: str) -> str:
             # newlines, which are replaced with normal spaces
             if char == "\n":
                 final_characters.append(" ")
-            elif char == "\u200D":
-                final_length = len(final_characters)
-                # only check for the ZWJ if it's between two characters
-                if final_length <= index < len(original) - 1:
-                    char_before_category = unicodedata.category(
-                        final_characters[final_length - 1]
-                    )
-                    char_after_category = unicodedata.category(original[index + 1])
-                    # only keep the ZWJ if it's between two symbol characters
-                    if char_before_category.startswith(
-                        "S"
-                    ) and char_after_category.startswith("S"):
-                        final_characters.append("\u200D")
+
+            # Keep zero-width joiner only if it's between two symbol characters, so we
+            # don't break certain emoji variants
+            if char == "\u200D":
+                try:
+                    before_category = unicodedata.category(final_characters[-1])
+                    after_category = unicodedata.category(original[index + 1])
+                except IndexError:
+                    continue
+
+                if before_category.startswith("S") and after_category.startswith("S"):
+                    final_characters.append(char)
         else:
             # any other type of character, just keep it
             final_characters.append(char)