[Fix] Fix normalization of non-alphabet based languages

src/libstat/backends/redis_backend.c

@@ -527,7 +527,8 @@ rspamd_redis_tokens_to_query (struct rspamd_task *task,
 								"HSET %b_tokens %b %b",
 								prefix, (size_t) prefix_len,
 								n0, (size_t) l0,
-								tok->t1->stemmed.begin, tok->t1->stemmed.len);
+								tok->t1->stemmed.begin,
+								tok->t1->stemmed.len);
 					}
 				}
 				else {

src/libstat/tokenizers/tokenizers.c

@@ -627,14 +627,10 @@ rspamd_uchars_to_ucs32 (const UChar *src, gsize srclen,
 			}
 #endif
 
-			if (cat == U_UPPERCASE_LETTER ||
-					cat == U_LOWERCASE_LETTER ||
-					cat == U_DECIMAL_DIGIT_NUMBER ||
+			if ((cat >= U_UPPERCASE_LETTER && cat <= U_OTHER_NUMBER) ||
 					cat == U_CONNECTOR_PUNCTUATION ||
 					cat == U_MATH_SYMBOL ||
-					cat == U_CURRENCY_SYMBOL ||
-					cat == U_INITIAL_PUNCTUATION ||
-					cat == U_FINAL_PUNCTUATION) {
+					cat == U_CURRENCY_SYMBOL) {
 				*d++ = u_tolower (t);
 			}
 		}