Thank your for your reply, the infomation is really useful.

On 1/8/07, HuYong <> wrote:
It is really the good news.
1. Consider the situation of users, CJK console often used in some system's modify and configuration, also may used in looking manual and help. So we always keep zhcon's input method simple, easy supplement, rapid respone with low cost cpu and memory.
2. The unicon input method used zhcon before, seems die. Those extern input method always make more problem when compile zhcon project. So if you want feedback new input method as library, we recommand you use open and load library style. This way avoids most compile problems. As the result, we can use same method to load other input method used in zhcon before.
Do you mean to create an instance of InputServer  with  UCIMF?
Or is there any formal input method API I can follow?

btw, UCIMF is implemented as shared library, and IIIMF, OpenVanilla supports are implemented as dynamic loaded modules. I guess, it would not be too difficult to port.

UCIMF take about 1MB memory while excution, due to PCF font storage and modules it loads. It current run well on my i486 machine( that is,...Thinkpad 560X ).
I will try to port first, and shrink it later when go further.

3. I want to use utf8 as native code in zhcon, but this needs much work. This modify can solve some problem in zhcon, like half code of Japanese, and so.
Great :-)

I have built zhcon (svn version), and try to use it to connect IRC and also read some text.
I found that LC_CTYPE is bound to be zh_CN.UTF-8.
Some text can not be correctly displayed, especially with Traditional Chinese text. Due to lack of font? I guess :-o
But it's really exciting to see UTF-8 in zhcon , I would like to help to make it better :-)

4. If we support pcf, zhcon can use wenquanyi as native Chinese font. This will be great, because wqy is GPL. The attach file mybe help you. If render cause problem, this maybe not nessary in console tools like zhcon. As I known, the render is not good solution especially when port zhcon to PDA or embed device.
I try first to use freetype2 as font render backend.

My current trick is to inherit BaseFont class( virtualize GetChar() first...),
and implement it without lots of changes. I am not sure whether this way goes,
so I conserve these detail and keep working.

Something is need to be known.
take "efont-unicode.pcf.gz" as example.
PCF font renders quite good and fast, but only in flat file!
While in compressed " pcf.gz", the font renders quite slow, due to lots of file_gzip_io_xxx.
So "jfbterm" and some other program just unzip it, "freetype" and "libXfont" use buf to save memory, but "freetype" will become very slow, even 1/30 than "TTF".

In common case, unicode pcf font file will take  2.2MB, and 700-800KB in compressed format. It doesn't hurt in comman desktops and laptops. I would suggest to use unicode
font file in uncompress format.

5. At the end, I like to point which is zhcon's main target.
   Full font support, at least CJK, no missing chars
   More and even full display support, can display on different moinitor, driver, and Linux env.
   Simple input method, can input any char at least CJK.
   All in GPL.
We appeciate any work for zhcon, you are welcome.
----- Original Message -----
From: Mat
Sent: Friday, January 05, 2007 1:45 PM
Subject: [Zhcon-devel] IME support and PCF font

Dear zhcon's developers:
I am developing an input method framework which provide IIIMF, OpenVanilla, (even SCIM in the future) support for unicode console.
The project, which called UCIMF, now works with "jfbterm", which is a linux unicode framebuffer console.
Here are project's hosts:
UCIMF's is designed to be a shared library, with a C/C++ API, aims to privode IME independent of console program.
I have successful ported it to fbiterm, which is another unicode framebuffer console, too.

Actually... I use a lot of graphic driver layer of zhcon in UCIMF, and the codes are really excellent,
and zhcon's program architeture teaches me a lot, during whole my work. I often think, to do some feedback to the zhcon project.

It's great to hear that zhcon support UTF-8 now !
I am glad to port my program to support zhcon, and to support TC/SC/KR/JP input methods in near future.

By the way, I current write a article, to discuss why PCF rendering rate is so much slower than TTF.
Here is the article( In Traditonal Chinese ):
So, I read some source code and some spec about PCF recently, and still continue studying PCF font rendering issues.

It's heard in the devel mailing-list that zhcon plans to support PCF font,
I wonder whether I have the chance to help to develope, or even cooperate.

sincerely, Mat.

Take Surveys. Earn Cash. Influence the Future of IT
Join's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash

Zhcon-devel mailing list