HIGHLIGHT() highlights a hindi word wrong

[githubmanticore]

When I try to highligh Hindi word हिंदी in text english text हिंदी पाठ it leaves one character out of the snippet:

➜  ~ docker stop manticore; docker run --name manticore -p9306:9306 -p9308:9308 --rm -d manticoresearch/manticore:dev && \  
docker exec -it manticore mysql && \  
docker stop manticore  
  
Error response from daemon: No such container: manticore  
983dc20eb06f5812e4cc0f4e91a348fad13fc31636a737748c7dacf643e95de2  
MySQL [(none)]> create table t(f text);  
MySQL [(none)]> insert into t values(0,'english text हिंदी पाठ');  
MySQL [(none)]> select highlight() from t where match('हिंदी');  
 -----------------------------------------------   
| highlight()                                   |  
 -----------------------------------------------   
| english text <b>हिंद</b>ी पाठ                 |  
 -----------------------------------------------   
  
  
MySQL [(none)]> show status like '%version%';  
 --------------- -------------------------------   
| Counter       | Value                         |  
 --------------- -------------------------------   
| version       | 3.5.1 84c8db5e@200817 release |  
| mysql_version | 3.5.1 84c8db5e@200817 release |  
 --------------- -------------------------------   

It also looks strange that the row with the snippet is shorter than expected (3 spaces are missing in the end).

[dmorgoon]

Not sure if it is related or not, but highlighting in general works unexpectedly with some unicode symbols/codepoints

create table test_idx2 (t text);
--- note **é** bellow
insert into test_idx2 values ('includé');
select highlight() as h from test_idx2 where match('include');
+------------------+
| h                |
+------------------+
| <b>include</b>́  |
+------------------+