compare regular expression on rAthena & Hercules

The most famous script that uses Regular Expression is GmOcean's Disguise Event
to learn more about Regular Expression, refer to Wikipedia
Visit this site for more in-depth tutorial

This NPC will Copy whatever you type prontera,150,185,5 script dasdasd 1_F_MARIA,{ input .@a$; // if (.@a$ ~= "^(\\w{1,70})$") // unfortunately this line doesn't return true on special characters if (.@a$ ~= "^(.*)$") npctalk $@regexmatch$[1]; end; L_0: npctalk $@p2$; consolemes CONSOLEMES_DEBUG, $@p2$; end; OnInit: defpattern 1, "([^:]+): (.*)$","L_0"; activatepset 1; end; }
It's common practice that you should include ^ at the beginning and $ at the end
without it, players can input whitespace to bypass the check

Example:
(.@a$ ~= "asdf") = it can be "asdf", "aaasdf", "asdfffff", " asdf   " and so on
(.@a$ ~= "^asdf") = it can be "asdf", "asdffffff", "asdfasdf", "asdf   " <-- this just check the begining string
(.@a$ ~= "^asdf$") = it can only be "asdf", no other match

defpattern 1, "([^:]+): asdf","L_0"; <-- it can be "asdf", "asdfffffff", "asdfasdf", "asdf    " and so on
defpattern 1, "([^:]+): asdf$","L_0"; <-- only "asdf" will trigger the label L_0, OR "ASDF", "AsDf" and so on ...

yup, *pcre_match is case-sensitive, but *defpattern isn't, remember to keep this in mind
to make *pcre_match case-insensitive, use (?i) flag
to make *defpattern case-sensitive, use (?-i) flag
prontera,150,185,5 script dasdasd 1_F_MARIA,{ input .@a$; if (.@a$ ~= "^(?i)(asdf)$") // becomes non-case-sensitive npctalk $@regexmatch$[1]; end; L_0: npctalk $@p2$; consolemes CONSOLEMES_DEBUG, $@p2$; end; OnInit: defpattern 1, "([^:]+): (?-i)(asdf)$","L_0"; // only "asdf" will trigger me now activatepset 1; end; }

in practice, this command is most useful to check invalid character input
- only allow 0-9, a-z, A-Z if (.@tmp$ ~! "^[0-9A-Za-z]+$") mes "Invalid Code"; or follow a format, like MySQL datetime format if (.@tmp$ ~= "^(\\d{4,4})-(\\d{2,2})-(\\d{2,2}) (\\d{2,2}):(\\d{2,2}):(\\d{2,2})$")) { mesf "Year: %s", $@regexmatch$[1]; mesf "Month: %s", $@regexmatch$[2]; mesf "Day: %s", $@regexmatch$[3]; mesf "Hour: %s", $@regexmatch$[4]; mesf "Minute: %s", $@regexmatch$[5]; mesf "Second: %s", $@regexmatch$[6]; } else mes "wrong format";

for rAthena counterpart, rAthena has *preg_match instead, which follow PHP style input if (!preg_match("^[0-9A-Za-z]+$", .@tmp$)) mes "Invalid Code";
However unlike Hercules, there are no way to capture the strings within a bracket
rAthena's *preg_match can only returns 0 or 1
correct me if I'm wrong, the source code seems to indicate it can return 0~10 but I couldn't get it display 2 or above
dispbottom ""+ preg_match("asdf", "asdf"); // return 1 dispbottom ""+ preg_match("asdf", "asdfasdf"); // return 1 dispbottom ""+ preg_match("asdf", "asdfasdf", 4); // return 1 dispbottom ""+ preg_match("asdf", "asdfasdf", 5); // return 0


Some older client that diff with multi-language support will return |00 in front of character speech
try check it with *consolemes with the very first script (rAthena is *debugmes)
if that happens, your syntax has to change into
defpattern 1, "([^:]+): (\\|\\d{2})?(.*)$","L_0"; activatepset 1;

External Links:
Disguise Event fix
defpattern regexp
ignore case sensitive in regexp