[sugj-tech:2027] Samba, SWAT and clients I18N/L10N frame work [Was: SWAT interna...]

Kenichi Okuyama okuyama at trl.ibm.co.jp
Fri Oct 27 01:44:47 GMT 2000

Dear All,

>>>>> "Miura" == Hiroshi MIURA <miura at samba.gr.jp> writes:
Miura> Why does anyone unlike i18n?  :-)
Miura> Is none of the members interesting in it?

I don't really know why. At least if any members are interested in
using UTF-8, then i18n code should be of interest for him/her.

Using Unicode does not escape you from using Multi-byte code.  UTF-8
is Multibyte itself, and if you think about performnace, converting
UTF-8 <-> UCS-* is nonsense. you can easilly treat UTF-8 directly
with good performance ( at least in case like Samba ). i18n code is
full of this kind of codes you(members)'ll want to see.

So, i18n-ed code holds good performance not only against Far-East
countries, but to anybody who uses UTF-8.

Also, Japanese Samba WorkGroup seems to have many patches for
bug-fix, performance gain aginst "ALL" languages ( or should I say
language independent part ), as well, coming up.

# You can find that out if you download Samba with their patches
# places on 
# ftp://ftp.samba.gr.jp/pub/samba-jp/samba-2.0.7-ja/samba-2.0.7-ja.2.0.tar.gz
# There's fix against bitmaps, string trimming, etc.

But if there's no reply, they will have confusing time thinking

" To which version should they attach the patch, exactly???"

for their patch are already growing into quite a large size.

Thinking about new functionality is OK. But before that, we should
re-brush up the current code, so that we can TRUST on codes that we
will rely on at adding functionality.

And I'd like to suggest start doing that from adding i18n patches.

Kenichi Okuyama at Tokyo Research Lab. IBM-Japan. Co. Japan.

More information about the samba-technical mailing list