Talk:OS2 API:Unicode

Anyone know of good documentation for the Unicode API?

What I'd like to know is if what I'm referencing is really using c parameters/returns versus the OS/2 declared ones like ULONG, etc. For example, UniStrToUcs returns an int according to the docs I'm reading. Ought this be APIRET? Prokushev, I think this API will be problematic with keeping language specific features out. How shall I handle void**? - Daniel

I think we should just describe such types like we does for, for example, ULONG. Like integer is signed 32-bit integer in range -xxxx..+xxxx.

void** is pointer to pointer on undefined type. For example, in Pascal we have ^Pointer for this construction.

So, actually, no many problems here, I think.

---Prokushev

OK, who's correct? I've found UniCode.INF where fields are defined SHORT versus c's short (ya, I know, they are the same value-wise) in the html file supplied with eComStation 1.2MR's Toolkit. If the INF is correct my question above is moot on the integer declarations. -- Daniel

Well, according to usual OS/2 practic all compiler depended things shadowed by defines (SHORT instead of short). It seems later libraries (with now common OS/2 knowlidges) broke this rule. If we want to be still OS/2 aware them my vote for SHORT and such macros usage.

---Prokushev