Japanese incorrect/archaic Kanji

[NeverSage]

Binary package hint: ibus-anthy

IBus 1.2.0.20091215

Anthy 9100h-0ubuntu2

Some Japanese characters (Kanji) are missing or replaced with archaic forms of the Kanji. For example when attempting to type 隠れん坊 (かくれんぼう), I first type the Hiragana, then when changing to Kanji 隠 (かく) is not available. Instead the archaic form 隱 is. 隠 is only available when typing other readings such as 隠 (いん).

It appears some Kanji are not properly connected to their readings.

ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: ibus-anthy 1.2.0.20100115-1
ProcVersionSignature: Ubuntu 2.6.32-24.39-generic 2.6.32.15+drm33.5
Uname: Linux 2.6.32-24-generic i686
Architecture: i386
Date: Wed Aug 11 10:54:40 2010
InstallationMedia: Ubuntu-Netbook 10.04 "Lucid Lynx" - Release i386 (20100429.4)
ProcEnviron:
 LANGUAGE=en_US:en
 LANG=ja_JP.utf8
 SHELL=/bin/bash
SourcePackage: ibus-anthy

[Hideki Yamane (henrich)]

I cannot reproduce it (with Debian unstable).

> For example when attempting to type 隠れん坊 (かくれんぼう), I first type the Hiragana,
> then when changing to Kanji 隠 (かく) is not available. Instead the archaic form 隱 is.
> 隠 is only available when typing other readings such as 隠 (いん).

I just type it as "kakurenbou", then anthy properly converted to 隠れん坊 as we expect.
Could you try it with new user?

[NeverSage (sages451)]

I've released my issue was not with ibus-anthy. By installing a new font the problem was fixed.

[Dave M G (martin-autotelic)]

I would like to reopen this bug. I am not sure about the process for doing so, so forgive me if I'm not doing this right.

I am using Ubuntu 11.04, ibus/anthy. However, I also experienced this problem in Ubuntu 10.04, Ubuntu 10.10, and also in UIM/Anthy.

It has come and gone, sometimes by replacing fonts, but it comes back, and I hope something can be done to create a permanent solution.

When viewing or entering Japanese, I sometimes see archaic or incorrect kanji.

the clearest example is 社. You may see the correct character there, but viewed on my system, the radical to the left is "shimesu" (示), not "nehen" (礻) as it should be.

I have attached a screenshot. The larger kanji (which is an image) is correct, the one in the small blue box, which should be the same, is the old/archaic/incorrect form.

This has been reported before, as above, and in the following places:
http://ubuntuforums.org/showthread.php?t=1487447
http://lists.tlug.jp/ML/1012/msg00044.html

However, none of the suggested fixes have rectified the problem.

I have attempted to reinstall fonts and remove any Chinese fonts as suggested.

However, the problem returns. It exists in at least FireFox and LibreOffice.

I have some known working fonts, so if, for example, I am working in LibreOffice and switch to a font known to work, then I can ensure I see the correct kanji.

However, if I go to web sites or open documents that are created by other people, then I can not as easily ensure control over fonts and be confident I am getting the right kanji.

In any case, this shouldn't be something I should have to manually monitor. Having installed Japanese support on my system, the default behaviour for all relevant character handling should be to show me the correct characters.

I hope this can be

[era (era)]

The status Incomplete basically means there is information missing. In this case I would actually say that it is correct, as (at least) one crucial detail is missing, namely the names of the fonts. You allege that some fonts are broken -- which ones? And others are correct -- which fonts would that be?

With this information, we could attempt to reproduce the problem, although it is possible that there are still some factors which need to be discovered before it can be reliably diagnosed.

[Dave M G (martin-autotelic)]

Era,

Thank you for responding. however, how do I determine which font is broken?

For example, in Firefox, I have my default font set to "Meiryo UI", which is a Microsoft font I know works. However, web pages set their own fonts, and so sometimes when it uses my default things display correctly, but other times, when they choose their own fonts, things go awry.

From what I can tell, when a web site or or program uses the Ubuntu default "Sans Serif", or any font that is only English characters, the system selects some default that is broken. But I don't know what that font is.

For my systen font, I use the "Ubuntu" font, and there doesn't seem to be an option to select a default font for Japanese in particular.

In LibreOffice, when I specifically select any known working font, then things are fine. However, if the applied font is not a Japanese font, then Japanese characters default to the broken font.

In all these cases, I can't tell where this broken font is being selected from.

Please let me know how I can reliably determine where the default broken font is coming from, and I will provide the names.

[Dave M G (martin-autotelic)]

I've solved this problem for me.

Following information here:
http://ubuntuforums.org/showthread.php?t=1487447
Basically I removed all Japanese fonts listed as installed in Synaptic, and then installed only these:
ttf-kochi-mincho-naga10
ttf-kochi-gothic-naga10
ttf-sazanami-gothic
ttf-sazanami-mincho

Now I am getting the correct kanji consistently.

Unfortunately, I do not know which of the fonts I uninstalled was corrupted, as I do not have the time to test and experiment. I will report if I work it out, though.

[Launchpad Janitor (janitor)]

[Expired for anthy (Ubuntu) because there has been no activity for 60 days.]