string.len with cyrillic in Gideros

unlying

string.len("база") return 8 in gideros. And any other cyrillic combination will return doubled number. With latin works fine. I thought that may be it is ok in Lua, but in Quick works fine...
So it is gideros bug.

Unknown

its not a bug. unless it is bug in standard lua interpreter (library liblua).
in short
> You might want to know how many Unicode characters are in a string. Depending on the encoding used, a single Unicode character may occupy up to four bytes.

for more information see http://lua-users.org/wiki/LuaUnicode

unlying

May be it is really Lua bug. But as i said before Quick returns right numbers.

ar2rsawseen

By default in Lua string.len in this case should return how many bytes string occupies, to get size for buffer, etc. And it's doing just that, so maybe it's a bug in Quick

I know I'm not being helpful, just the fact

http://stackoverflow.com/questions/10097941/print-number-of-characters-in-utf-8-string

So basically you'd need a specific library to deal with unicode strings, I just don't know any of them, maybe someone else will.

unlying

I used this lib in my Rebus game. Looks like it fit to this problem too. If anybody will need it - file attached to message.
May be Quick has some sort of autodetect... curiously...

atilim

Gideros always saves the files in UTF-8 format. For Quick, I think there are two possibilities:
1. You've used a encoding like http://en.wikipedia.org/wiki/ISO/IEC_8859-5 other than UTF-8.
2. Quick has modified string.len (and other string functions) to work with UTF-8 seamlessly.

atilim

And I think the most convenient way to work with different encodings is to use http://www.gnu.org/software/libiconv/ library. And it seems there is a Lua bindings too: http://ittner.github.com/lua-iconv/

But it's a big library and its binaries takes about ~2MB.

MauMau

I'm not sure if my issue fits into the same category - I found that string.upper and string.lower do not work as expected.

According to lua.org, the following should work:

string.upper("ação") -- returns: AÇÃO

With Gideros, however, it returns: AçãO

Is Gideros perhaps using an outdated library?

uzubari

Hello
Here is a solution which works for me for getting correct unicode string length

-- Return the character count in a unicode string word function wordLength( word ) local wordlength = 0 for c in string.gmatch(word, ".") do if( string.byte(c) < 128 or string.byte(c) > 191) then --print("c "..string.byte(c)) wordlength = wordlength + 1 end end return wordlength end