Re: [AUCTeX-devel] Fix for tex-jp.el

[Uwe Brauer]

>>> "Ikumi" == Ikumi Keita <address@hidden> writes:
Hi Ikumi

(I confess my ignorance and assume Ikumi is your first name and Keita is
your family name, an order which chosen by the european and arabic world. I
hope it is correct. Ikumi is a female name? Is this correct? BTW Uwe is
male. I am saying that since Uwe is not a very common name.)


   >> And you not only want me to compile it but to use and to run it? If so
   >> should I do it for Japanese tex files? Unfortunately I don't speak
   >> Japanese, so maybe you could send me some test files, which I could try
   >> to modify and compile?

   > Ah, no, I'm not going to ask a non Japanese-speaker to do such a
   > complicated task.

   >> So how shall we proceed? First is your patch in the auctex git repo, or
   >> not? Please confirm this.

   > I think I have done it in the former half of this mail.

   >> If it is not, I will checkout the commit you recommended create a branch
   >> and apply your patch?

   > Yes, please.

Ok today and tomorrow I might be very busy.

   >> in the next step I will try to generate an official xemacs pkg using
   >> xemacs 21.4 mule?

   > I think it is not necessary.  Rather than that, could you please
   > clarify the point stated in
   > http://lists.gnu.org/archive/html/auctex-devel/2017-02/msg00015.html
   > ?  That is, does the existence of the functions
   > `get-coding-system-from-locale' and `current-locale' depend on the mule
   > feature?  Is it independent of the xemacs version (21.4, 21.5)? 

No it is not. Both are present in 21.5.X Mule but absent in 21.4 Mule/NoMule



   > In addtition, I would appreciate if you have answers to the following
   > questions and could kindly tell them to me: Does xemacs have a simple
   > and reliable way to know the coding system accomanying the locale?

21.5 might, I am not so sure about 21.4, and I do remember discussions
about this issue on the xemacs beta list some time ago. I will come back
to you soon.


   > If so, how?  I proposed a way of
   >                        ((fboundp 'get-coding-system-from-locale)
   >                         (get-coding-system-from-locale
   >                          (current-locale))))))



   > to do the job for xemacs, but I would like to know if this is
   > appropriate, and to know a better way if exists.

   > Regards,
   > Ikumi Keita

   > P.S. By the way, speaking of the topic that xemacs 21.4 without mule
   > fails to compile the official xemacs pkg, isn't it due to some errors
   > during byte-compiling tex-jp.el?

Yes of course.

   > Then, it might be solved by removing tex-jp.el from the target of
   > byte compilation. That is what AUCTeX-team-Makefile does when the
   > given emacsen lacks mule feature. The file tex-jp.el contains full
   > of Japanese text in its non-comment part encoded in the coding
   > system iso-2022-jp, which will be an obstacle that non-mule emacsen
   > can never overcome because of the nature of that coding system.

Yes you are right, but most of the few Xemacs developers encourage the
use of Mule xemacs anyhow. And I think that is what Norbert Koch uses,
21.4 Mule version. 

But you are right, if both flavors are shipped it should be done as you
say. I have a look into that as soon as possible, but the Xemacs pkg
system has still no fully recovered from its web page breakdown it
suffered some month ago.