... er, UNICODE!! Not that there's much of a difference, unicode still destroyed my world.I was up until 5 AM trying to figure out how to enumerate the character sets from a font. It was bloody painful.
For those who are curious, the API function EnumFontFamiliesEx was the key. Inside that bastard's call back function (FONTENUMPROC) there's a structure passed as a parameter: NEWTEXTMETRICEX, and inside that there's another struct (small wonder I can't find anything) called FONTSIGNATURE. In FONTSIGNATURE, there's an array called fsUsb, it's 4 DWORDs. Each of these DWORDs are the seperate 32 bit pieces of a 128 bit value, from this value you can determine which character sets are supported. FUN! The MSDN has a list. Lists are fun.
Anyway, thanks to adding of the unicode crap,I reworked the character selection screen. It works MUCH better and MUCH faster than it did prior.
I also found out something else as I was finishing the unicode implementation today: C# sets the default CharSet on pinvoked functions to CharSet.ANSI by default. I wonder what the reasoning behind that was? Frankly it'd be a fuckload more useful to have it set to CharSet.Auto by default (apparently the default can vary from language to language). I know this to be true because my imported GetCharABCWidths() function seriously messed up on unicode characters above 255. When I changed the DllImport attribute to set the CharSet to CharSet.Auto it worked beautifully.
So, in summary, use BM font, it's much better than anything I can do.
