Possible one more i++ in cmsys.strip_nonebig5？ about pttbbs

[IepIweidieng]

According to common/sys/big5.c:

• Unicode code points without the corresponding Big5-UAO code point are always converted to 0xfffd (see u2b_table).
• Big5-UAO code points 0xfeff-0xffff are left unchanged when converted into Unicode (see b2u_table).

Consider the case for strip_nonebig5: str is "\xff\xfdQ" and maxlen > 3.

When i is 0, both else if(str[i]==255) { and if(251<=str[i+1] && str[i+1]<=254) { hold, and the first i++ following them skips the byte \xff.

if(i+1<maxlen && str[i+1]) also holds, and the second i++ skips the byte \xfd.

However, the for loop also performs i++ after the continue and makes the byte Q ignored.

Therefore, the second i++ is redundant and erroneous. It makes the byte following "\xff\xfd" ignored.

Note that the UTF-8 encoding of U+FFFD is ef bf bd (binary: 11101111 10111111 10111101), but UTF-8 is only used for handling input and output for UTF-8 clients and is converted to Unicode code points internally, so it should be irrelevant to this issue.

Function usages of strip_nonebig5:

• mbbsd/mbbsd.c: setup_utmp for user nickname display
• mbbsd/bbs.c: readdoent for article title display

Only the displayed strings are processed through strip_nonebig5, while the strings which are stored to the disk are not affected.