[Draft #2] Samba 3.0 roadmap...idmap storage & central idmap repository

[Steve Langasek]

On Wed, Jul 10, 2002 at 12:25:42AM +0200, Simo Sorce wrote:
> On Tue, 2002-07-09 at 22:59, Jeremy Allison wrote:
> > But Simo, I disagree about the internal rep. I think it
> > needs to be utf8 for Samba internal strings. We already
> > have to deal with mbcs issues - this doesn't make it any
> > worse.

> Have you thought how difficult is to effectively use utf8 strings?
> search/replace/uppercase/lowercase?
> it is very difficult to manipulate correctly utf8 strings without
> introducing errors. I already experimented working with ucs2 null
> terminated strings and it is way more easy and less prone to errors.
> a character is always 2 bytes long and a byte codification doesn't
> change meaning based on which place do it takes inside a string.
> And substituting/manipulating characters in a string do not change the
> string length with ucs2!

> Can you instead tell me what are benefits of using utf8?

Well, for starters, utf8 is forwards-compatible with the full current
Unicode spec, whereas UCS-2 is truncated at 16 bits (hence Apple's use
of UTF-16).  The less work that has to be done to convert Samba from
UCS-2 to UCS-4 on-the-wire when the time comes, the better.

Steve Langasek
postmodern programmer
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://lists.samba.org/archive/samba-technical/attachments/20020713/e3d93bfd/attachment.bin