Re: unicode (char as abstract data type)

Alex Belits (abelits@phobos.illtel.denver.co.us)
Sat, 18 Apr 1998 23:33:18 -0700 (PDT)


On Sun, 19 Apr 1998 linker@nightshade.ml.org wrote:

> Perhaps there are more Linux boxen out then the all commercial unixen
> combined..

Well, all *BSD that also have more installations than all commercial
unices combined, use UFS.

> And can't *bsd r/w Ext2?

I never checked, even though I use FreeBSD, too, and FreeBSD can
partially emulate Linux, so it makes sense to use ext2.

The point is, the need for compatibility at the filesystem level with
something that one doesn't need enough to sacrifice the consistency of
text/filename representation. Now I can read() or even mmap() file, take
arbitrary byte range from it, and pass it as an argument to stat()
or open(). And the same string of bytes I will get from readdir(), from
any file, socket, pipe, or device when it mentions that file. In whatever
local encoding that exists in the world, and I don't have any need to
know, what encoding is actually used, because if application needs it, it
will know how to display or input it.

And sure as hell even if I never again in my life will be able to
see the content of a directory, mounted from Linux server by Windows
client, in correct glyphs without a convertor, I still don't want to lose
the consistency of characters representation on the OS.

--
Alex

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu