[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#215647: [patch] xterm 4.3.0-0pre1v3 i18n


From: Branden Robinson <branden@debian.org>
Subject: Bug#215647: [patch] xterm 4.3.0-0pre1v3 i18n
Date: Sat, 18 Oct 2003 16:11:24 -0500

> There are *also* discussion and communication problems, in that I feel
> you were not sharing vital information with me, namely that you had
> already submitted your patch upstream, and it had been rejected.

As I said, the patch is different.

> That doesn't mean I will *automatically* reject your patch, but it means
> that I need to understand upstream's objections, and that you may need
> to make a stronger case for why upstream's objections are not relevant
> to Debian, are outweighed by other considerations unique to our OS.

I remember that Dickey objected my patch because introduction of
new resources will slow down the invocation procedure of xterm.
I felt that the reason is very strange, because my additional
resources are not too fat - "ufont", "ufont1", "ufont2", ... which
are Unicode fonts which are used instead of "font", "font1", "font2",
and so on, while usage of non-iso10646-1 fonts in UTF-8 mode is
apparently wrong though xterm automatically use UTF-8 mode in
some condition.  If iso10646-1 font is not used automatically,
xterm's feature to automatically use UTF-8 mode is meaningless,
I think.

He might say some additional reasons which I forget.

Anyway, I will have to find discussion from my mail archive or
discuss with Mr. Dickey again....

> He didn't say that, but he may have read the patch hastily and assumed
> it was.

Well, then did he change his idea on the current patch?

Anyway, I think different patch should be appropriate between XFree86
and Debian.  In XFree86 (upstream) level, addition of resources is
more appropriate than my current patch.  It is because xterm in
upstream level may be run on various platforms with completely
different set of fonts.  Thus, the default font must be a font
named "fixed" (which is always available) in the upstream level.
Addition of resources will enable automatic usage of Unicode fonts
and the default of "fixed" at the same time.  On the other hand,
addition of resources in Debian level is, as you might think, not
very good because it is too much modified from the upstream.
Also, since fonts in my patch are available in xfonts-base package
in Debian, these font names are less problematic than upstream level.

So, please think your own reason to reject/accept/modify my patch,
apart from the different patch for the upstream and Mr. Dickey's

BTW, unlike Mr. Martin Quinson may think, my patch is useful not
only for CJK people but also for all non-ISO-8859-1 people.  Since
there are less ISO-8859-1 countries because of introduction of
Euro, my patch benefits many people (and no harm for other people
in US, UK, and so on).

> For years now I've wanted to arrange things such that the very generic
> X font aliases "fixed", "proportional", "5x7", and so forth would go
> through another layer of indirection; one that would allow people to
> choose a codeset for the font that would make sense for their locale.  I
> had this idea way back when xfonts-cyrillic was called xfonts-cyr.

I think this way may be useful for some people, if the aliases are
set automatically according to the current LC_CTYPE locale.
(There are already a standardized way (LC_CTYPE) to "choose a codeset"
and we should not introduce different configuration ways for users
to achieve the same thing.)  However, since this is not useful for
multibyte people nor UTF-8, we will need additional works anyway.
Especially, UTF-8 will be more and more important in future.
Thus, this idea does not reduce the amount of needed work.

Tomohiro KUBOTA <kubota@debian.org>

Reply to: