[javalia]

i made this function because GetClientName doesnt cuts string in utf8 safe way

new const UTF8MULTIBYTECHAR = (1 << 7);

//this function assumes string is a correctly encoded utf8 string that is cutted in not utf8 safe way.
stock terminateUTF8String(String:buffer[], const maxlength = -1){
	
	if(maxlength > 0){
	
		buffer[maxlength - 1] = '\0';
		
	}
	
	new length = strlen(buffer);
	new bytescounted = 0;
	
	if(length <= 0){
	
		return 0;
	
	}
	
	for(new i = length - 1; i >= 0; i--){
		
		if(UTF8MULTIBYTECHAR & buffer[i] == '\0'){

			return 0;//its a single byte character, we have nothing to do.

		}else{
		
			//j is not a good idea...
			for(new j = 1; j <= 7; j++){
			
				if((UTF8MULTIBYTECHAR >> j) & buffer[i] == '\0'){
				
					if(j == 1){
					
						//its part of multi byte character
						bytescounted++;
						break;
					
					}else{
					
						//its starting byte of multi byte character, so lets see if we readed enough amount of utf8 strings before and cut it if its not.
						if(bytescounted != (j - 1)){
						
							buffer[i] = '\0';
							
						}
						
						return 0;
					
					}
				
				}
			
			}
		
		}

	}
	
	return 0;

}

changed code a bit and tested and working.

[Powerlord]

I'm not sure exactly what all this does, but...

if(UTF8MULTIBYTECHAR & buffer[i] == '\0'){

Why not just use IsCharMB here?

Having said that, if you're dealing with GetClientName, the String you put it in should be MAX_NAME_LENGTH in size.

[javalia]

ah, i forgot about that function, however, because of other bit operations, i still need that const.
so i will just leave that.


and, even u give string buffer with enough size that corresponds MAX_NAME_LENGTH, because game`s client name string itself is cutted in non utf8 safe way, it can return strings that r having a sliced multibyte character on end of string.

[javalia]

eh...and that IsCharMB looks like not so good to handle bytes in String[].

p.s.
or is it ok? well, i have no idea. but i will leave that code because its not bad.

p.s.
well, the function seems like ok.

[Root_]

When I am trying to rename player with utf8 character(s), it prints corrupted result (symbols). Can your snippet fix that?

[RedSword]

Originally Posted by Root_ When I am trying to rename player with utf8 character(s), it prints corrupted result (symbols). Can your snippet fix that?

Could you give an example on how to reproduce the problem ? Could that be a file encoding problem on your side ?

[Root_]

I used default playercommands plugin. Ie sm_rename #userid|name, which is containing UTF8 symbols (see russian alphabed)
It showed ********* instead of correct symbol

[GoD-Tony]

Originally Posted by javalia i made this function because GetClientName doesnt cuts string in utf8 safe way

I ran into this same problem today, thanks. Here is a modified version I use which calls IsCharMB:

PHP Code:

stock TerminateNameUTF8(String:name[])
{
    new len = strlen(name);
    
    for (new i = 0; i < len; i++)
    {
        new bytes = IsCharMB(name[i]);
        
        if (bytes > 1)
        {
            if (len - i < bytes)
            {
                name[i] = '\0';
                return;
            }
            
            i += bytes - 1;
        }
    }
} 


Apparently Steam allows more bytes in a name than the game does, so you can end up with a half-UTF8 character at the end of a GetClientName call. MySQL errors out when attempting to Insert these strings. Very useful fix in this case.