Skip site navigation (1)Skip section navigation (2)
Date:      2 Apr 2000 18:33:21 +0200
From:      naddy@mips.rhein-neckar.de (Christian Weisgerber)
To:        freebsd-hackers@freebsd.org
Subject:   Re: Unicode on FreeBSD
Message-ID:  <8c7soh$179g$1@bigeye.rhein-neckar.de>
References:  <20000320194702.11223.qmail@web3101.mail.yahoo.com> <8bitar$2i4f$1@bigeye.rhein-neckar.de> <20000329033908.A14122@happy.checkpoint.com> <20000402051559.A52041@happy.checkpoint.com>

next in thread | previous in thread | raw e-mail | index | archive | help
Anatoly Vorobey <mellon@pobox.com> wrote:

> What do you think?

For what it's worth (me not being a committer and generally not on
the productive side of things), I morally support this idea.

To push certain buttons: what you are suggesting is to bring syscons
up to what the Linux console already supports in this respect.

> - in raster modes (SC_PIXEL_MODE on, etc.) more than 256 characters can
> now be trivially drawn. [...]

I should point out that what you have outlined is very limited
Unicode support. It's great for the primary European application
of Unicode, i.e. having an extended character set that combines a
character repertoire previously not available from 8-bit character
sets, e.g. the ability to write both French and Russian in the same
text. Without double-width and combining characters it won't be
nearly as useful for Asian users, though, so don't expect rampant
enthusiasm from that corner. Whether "full" Unicode support is
desirable in a console driver is another question.

> - the road is wide open for Unicode support in userland, through UTF-8. 

A UTF-8 capable xterm has been capable since, uh, last summer I
think. It's in XFree86 4.0. The road is wide open already. It would
be even more open, if the work done by the Linux people wasn't
consistently GPLed and could be reused <sigh>.

> - The format of screen font files must be changed.

Hardly an issue. You'll have an array of glyphs and an array which
associates a Unicode code point with each glyph. In fact, you just
might want to use the same format the latest Linux console tools
have already pioneered for this purpose.

> - some rendering routines are slowed down due to the fact that simple
> 8-bit array lookups are no longer available for getting characters'
> information. This may be circumvented somewhat by smart searches/hash 
> tables.

Linux uses some kind of hash tables. I don't know the implementation
details, but speedwise the overhead appears to be negligible.

Hint: For anybody interested in unicodification, the linux-utf8
mailing list is a must-read.

I also think the creating of a freebsd-i18n list is long overdue.
I18N issues are largely lost among the traffic on -hackers and
-questions, and it has become something of a specialty area since
most people appear to be served well by the existing non-solutions.

-- 
Christian "naddy" Weisgerber                  naddy@mips.rhein-neckar.de



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?8c7soh$179g$1>