Hello Steve,
Thanks for your comments.
SS> XIM is working very well with the CSI (code-set independent)
SS> version of xterm provided by Li18nux.org (patches from IBM).
Just to clarify matters: as far as I am aware, XIM works fine in
XFree86 XTerm 158 on a properly configured system, thanks to
Tomohiro's work.
The only issue is that our XTerm currently crashes instead of
gracefully aborting if the system's configuration is broken; we are
committed to fixing this very minor bug.
SS> It works equally well in our UTF-8 locales and our non-UTF-8
SS> locales.
It goes without saying that I gratefully acknowledge Sun's pioneering
work in the area of international support on X11 systems. While some
of the engineering choices that Sun's engineers have made are diffe-
rent from the ones we did, all that matters in the end is proper in-
ternational support at the user level.
I am extremely curious to know how you intend to implement combining
characters in your framework.
SS> Because of this (CSI), Sun will be adopting this version
SS> of xterm, rather than the utf-8 hardwired one, for a future
SS> release of Solaris.
As with most engineering decisions, there can be multiple choices that
are correct. Historically, Sun have used the CSI approach, and I have
no doubt that your engineers are well versed in its arcana.
XFree86 is committed to a model in which applications use Unicode
internally and convert at the application's boundary. Sun would
appear to be committed to a different model. Assuming that you intend
to keep your code under a sufficiently liberal license for us to use,
this can only make me glad, as crossbreeding between different
implementations can only be a good thing.
SS> I also encourage the Xfree86 community to strongly consider the
SS> value of a code-set independent version of xterm, since there are
SS> many Linux platforms with non-utf8 international locales which
SS> will benefit from an implementation which works cleanly with both
SS> the encoding and the input methods of those locales.
As far as I know, there is no user-visible difference between the two
models. User-level code set independence is provided in XFree86 XTerm
by Tomohiro's iconv patches or, alternatively, by the ``luit'' layer.
If you disagree with that assessment, please do discuss any short-
comings of our approach that you may have noticed -- as I mentioned
above, your technical expertise is highly valued.
While I have your ear: the version of dtterm that Sun distributes
since Solaris 2.6 does selection conversion correctly. Last time I
checked, this was not the case of the IBM patches. I would like to
encourage you to check interoperability of your version of XTerm with
dtterm as well as our version of XTerm in wide character mode; this is
especially important with Europe switching to the Euro by the end of
the year.
Support for the UTF8_STRING mechanism would be a welcome addition, now
that Ienup Sung has blessed the protocol.
SS> P.S. I agree with Owen's sentiments about XIM, and Sun strongly
SS> encourages the use of IIIM as an alternative.
Thanks to both of you for the information. I will try to understand
IIIM and get back to you.
See you, and take care in those difficult times,
Juliusz
P.S. As usual, I am speaking for myself and not for XFree86.
_______________________________________________
I18n mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/i18n
