Ok, I ran this code snippet to see what a charmap should look like on iPhone OS :
char ch = ' '; // first char
char *str = malloc(17*sizeof(char));
str[16] = '';
for (int l = 0; l < 15; l++) { // lines
for (int c = 0; c < 16; c++) { // columns
str[c] = ch++;
}
NSLog(@"%@", [NSString stringWithCString:str length:16]);
}
free(str);
This is the output (it's a screenshot from the console cause I'm not sure this forum can display all these special chars) :

So, theoretically, with a charmap like this one, you should be able to use special chars from the extended ASCII character set and display them with a LabelAtlas. However, there's an issue with character encoding. These special chars are encoded in UTF8 and when the updateAtlasValues function of the LabelAtlas class retrieves the C characters from the string (with const char *s = [string_ UTF8String];), the special chars are now the size of 2 chars each. For instance, the letter "é" is retrieved as "é". But these two characters "é" are displayed correctly by the LabelAtlas, and they're beyond the 128th char in the charmap.
I'd say a workaround is needed to fix the character encoding issue. I tried something like that in updateAtlasValues :
//const char *s = [string_ UTF8String];
char *s = malloc((n+1)*sizeof(char));
[string_ getCString:s maxLength:n+1 encoding:NSEncodingXXX];
and played with both the encoding parameter AND the encoding of the .m file, but never got it to work correctly.