grep.7aa698d3.diagnosis.txt

If grep conducts a case-insensitive search (-i) on an input that contains
multibyte characters and the locale is UTF8, then grep prints a match of
incorrect length. When conducting the case-insensitive search, EXECUTE_FCT
first computes a lower-case of the input (search.c:388).  The length of the
match is computed for the match in the lower-case input (search.c:555).
However, the lower-case of a multibyte character can take 1 byte less. So, the
length of the normal-case and lower-case input differ. The computed value of
match_size could be half the expected value (grep.c:1081-1085). Hence, the
match in the normal-case input is printed with incorrect length (grep.c:1091).

Example of Correct Fix: 
Add a mapping between normal-case and lower-case string to compute the length
of the match in the normal-case string from the length of the match in the
lower-case string. 

Examples of Incorrect Fixes: 
1) Do not lower the case (Regression because a case-insensitive search is
case-sensitive). 

2) If matched string contains a multibyte char, double the match_size
(Incomplete Fix because it works only of all are multibyte characters). 

3) Print complete line if there is a match (Regression because only match
should be returned).