Re: unicode (char as abstract data type)

Raul Miller (
Mon, 18 May 1998 13:34:23 -0400

Albert D. Cahalan <> wrote:
> UCS-2 and UCS-4 are acceptable for the kernel interface, internal
> kernel usage, internal libc usage, and the on-disk format of new or
> foreign filesystems. UCS-2 is required by Java, Samba, and Wine.

L"gcc appears to support UCS-4, not UCS-2.\n"

In any event, if libc support can be done right for internal kernel
interfaces which are UCS-2, it can be done right for UTF8.

I see no major performance wins coming out of a switch to UCS-2, and I
see lots of migration nightmares. [Can you imagine redhat distributing
a release of the linux kernel that requires everyone rebuild their hard
disks and to never be able to interoperate with earlier kernel releases?
If not, where do even trivial performance wins come from?]


- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to