UTF-8 (multibyte chars) and the game

[Nightborn]

In several locations, you might have problems when a user writes over the hard character limit (i.e. 50 in the title) with an multibyte character.

In this case, the char will "break in two" with only the first part saved... an invalid char.

As I just had with a clan motd, this can mean you can display it with the wrong char at the end (an ? normally as the browser can't interpret  that) but an textarea box i.e. will be empty... because the htmlentities function will return nothing there...

Solution: remove the last character and it will work again.

I don't know if I do a complete core solution to work with those char problems... it's hard to grab as html elements only take "char" as counter, and char=1 for them, they don't calculate *how much space* I need in the database.

However, we could simply scale the DB fields larger, to fetch and save such overheads. In the clans,however a substr() is done which breaks it an mb_substr would suffice.

[Wasili]

I overload, in my version of 0.9.7, all string functions with a mb-equivalent. And also, of course, I converted all files and the database to Unicode. It works perfectly and I never hat any problem with broken names.
I don’t need HTMlEntities anymoire, only HTMLSpecialchars which would replace <, > and ".
The only limitation is that you must edit the php.ini or overwrite the settings in a .htaccess-file.

Code:

php_value mbstring.func_overload 7 

php_value mbstring.internal_encoding UTF-8

[Nightborn]

I suspect the broken stuff comes from the database format and field length there.

Lotgd does not set any charset there, just uses the default one. Hence the errors I presume.

it has not really something to do with the php mb_ functions.

Also, I have currently in the titlechanger hardcoded a strlen (no mb_) in it to make at least a proper insert as I *don't know* what kind of charset to expect in the db.

I am aware this is not 100% properly solved