Re: IBM AS/400 Migration

[Antoine Leca]

David Turner wrote:
> 
> >     The big difference is the fact that the AS/400 uses EBCDIC not ASCII. I
> > have to change some of the "Constants"
> > (ie "head") back to ASCII. The compiler makes 'h' EBCDIC (0x88) I need 'h'
> > ASCII (0x68).

Quite common problem when porting to EBCDIC, isn't it?


> >     I will give you back the "changes" that I needed to make, when I
> > complete the task.
> >
> Isn't this problem compiler-specific ?

No, it is platform-specific !

> I mean isn't there a switch to indicate that all characters are encoded
> using ASCII.

Perhaps, that would be helpful.


> I thought that ASCII want part of ANSI C. Is this true ?

You mean, that ISO C would require ASCII? no, that wrong: you can perfectly
have a conforming compiler on a non-ASCII machine. The only requirement
is to have (internally) the same character set as ASCII, including []{}|\^~
($ @ and ` can be dropped); but the values can be different from ASCII.

By the way, there is another thing that is completely overseen in FreeType :
theorically (according to the standard), char can be different from 8 bits,
and there are even some mainframes with C compilers where this is real
(Unisys, Bull). However, this is just for theorical information, since
I do not expect everybody to compile FreeType on such a host ;-).


Antoine