http://bugzilla.suse.com/show_bug.cgi?id=1162654 http://bugzilla.suse.com/show_bug.cgi?id=1162654#c1 --- Comment #1 from Martin Doucha <martin.doucha@suse.com> --- Created attachment 829097 --> http://bugzilla.suse.com/attachment.cgi?id=829097&action=edit Kiten search results 2) When I shorten the search string to only the first character (辞), Kiten finds some results but lots of Japanese characters are rendered wrong. (No comparison needed this time). Note that the wrong rendering appears only in the result list. When I run Kiten with LC_ALL=ja_JP.utf8 locale, Japanese characters in the UI will be rendered correctly (except for the search result list which will show the same garbage). -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.suse.com/show_bug.cgi?id=1162654 http://bugzilla.suse.com/show_bug.cgi?id=1162654#c3 Christophe Giboudeaux <christophe@krop.fr> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |CONFIRMED --- Comment #3 from Christophe Giboudeaux <christophe@krop.fr> --- 15:06:46 - kiten(5235) - unknown: Loading edict from "/usr/share/kiten/edict" 15:06:46 - kiten(5235) - unknown: Dictionary Loaded : "edict" 15:06:46 - kiten(5235) - unknown: Loading kanjidic from: "/usr/share/kiten/kanjidic" 15:06:46 - kiten(5235) - unknown: Dictionary load FAILED: "" 15:06:46 - kiten(5235) - unknown: Dictionaries loaded! # ls -l /usr/share/kiten total 36 lrwxrwxrwx 1 root root 14 janv. 29 17:48 edict -> ../edict/edict lrwxrwxrwx 1 root root 17 févr. 4 15:09 kanjidic -> ../edict/kanjidic lrwxrwxrwx 1 root root 17 janv. 29 17:48 radkfile -> ../edict/radkfile -rw-r--r-- 1 root root 4015 janv. 8 07:25 romkana.cnv -rw-r--r-- 1 root root 18576 janv. 8 07:25 vconj # ls -l /usr/share/edict total 52920 -rw-r--r-- 1 root root 19994047 janv. 3 17:43 edict -rw-r--r-- 1 root root 32077830 janv. 3 17:43 enamdict -rw-r--r-- 1 root root 542923 janv. 3 17:43 kanjd212 -rw-r--r-- 1 root root 56469 janv. 3 17:43 kanjd213u -rw-r--r-- 1 root root 1274391 janv. 3 17:43 kanjidic -rw-r--r-- 1 root root 145206 janv. 3 17:43 kradfile -rw-r--r-- 1 root root 82111 janv. 3 17:43 radkfile The files are there, no idea yet why kiten doesn't load them -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.suse.com/show_bug.cgi?id=1162654 http://bugzilla.suse.com/show_bug.cgi?id=1162654#c5 --- Comment #5 from Martin Doucha <martin.doucha@suse.com> --- (In reply to Christophe Giboudeaux from comment #3)

The files are there, no idea yet why kiten doesn't load them

Yes, the files are there. Kiten even logs into console that it's loading them on startup. My first hunch is that Kiten is trying to load them in some other encoding than UTF-8. Don't ask me why. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.suse.com/show_bug.cgi?id=1162654 http://bugzilla.suse.com/show_bug.cgi?id=1162654#c10 --- Comment #10 from Martin Doucha <martin.doucha@suse.com> --- (In reply to Wolfgang Bauer from comment #6)

Maybe the kanjidic file included in the edict package is broken?

Yes, I was about to say that. Checking for dictionary updates in the menu will fix search and kanji browser but radical selector will remain broken. $ file .local/share/kiten/kanjidic .local/share/kiten/kanjidic: ISO-8859 text, with very long lines $ file /usr/share/edict/kanjidic /usr/share/edict/kanjidic: UTF-8 Unicode text, with very long lines I guess that's what's going on. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.suse.com/show_bug.cgi?id=1162654 http://bugzilla.suse.com/show_bug.cgi?id=1162654#c12 --- Comment #12 from Wolfgang Bauer <wbauer@tmo.at> --- Downloading radkfile from kiten's git repository and copying it to ~/.local/share/kiten/ does seem to fix the radical selector. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.suse.com/show_bug.cgi?id=1162654 http://bugzilla.suse.com/show_bug.cgi?id=1162654#c14 --- Comment #14 from Wolfgang Bauer <wbauer@tmo.at> --- (In reply to Wolfgang Bauer from comment #11)

It could be that all the problems come from the fact that we replace the datafiles included in kiten (edict kanjidic radkfile) with symlinks to the ones from edict. Btw, this was introduced a bit over a year ago: https://build.opensuse.org/request/show/651896

I'm currently trying to build a package that doesn't do that... Available here now if you want to try it out: https://download.opensuse.org/repositories/home:/wolfi323:/branches:/KDE:/Ap...

It indeed fixes all the mentioned problems here AFAICS. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.suse.com/show_bug.cgi?id=1162654 http://bugzilla.suse.com/show_bug.cgi?id=1162654#c16 --- Comment #16 from Wolfgang Bauer <wbauer@tmo.at> --- (In reply to Christophe Giboudeaux from comment #15)

(In reply to Wolfgang Bauer from comment #12)

...

Downloading radkfile from kiten's git repository and copying it to ~/.local/share/kiten/ does seem to fix the radical selector.

Interesting, the file from the edict package is UTF-8 and the file in the kiten repository is ISO-8859

Yes, converting edict's files to EUC-JP encoding seems to fix things as well. But it still looks different than in the Gentoo screenshots (might not be a problem though, I suppose the files are a newer version). The reason why it cannot read the UTF-8 encoded files, is probably this code in lib/DictEdict/dictfileedict.cpp and DictKanjidic/dictfilekanjidic.cpp: QTextStream fileStream( &file ); fileStream.setCodec( QTextCodec::codecForName( "eucJP" ) ); But if we patch this, we'd break the update functionality in kiten I suppose. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.suse.com/show_bug.cgi?id=1162654 http://bugzilla.suse.com/show_bug.cgi?id=1162654#c18 --- Comment #18 from Martin Doucha <martin.doucha@suse.com> --- (In reply to Wolfgang Bauer from comment #16)

Yes, converting edict's files to EUC-JP encoding seems to fix things as well. But it still looks different than in the Gentoo screenshots (might not be a problem though, I suppose the files are a newer version).

The buttons in the radical selector appear in random order in each column (even on Gentoo) so what you see won't match the screenshot exactly. But if you can find all symbols from the screenshot somewhere in the correct column of your radical selector, it works as intended. -- You are receiving this mail because: You are on the CC list for the bug.

http://bugzilla.suse.com/show_bug.cgi?id=1162654 http://bugzilla.suse.com/show_bug.cgi?id=1162654#c22 --- Comment #22 from Swamp Workflow Management <swamp@suse.de> --- openSUSE-RU-2020:0193-1: An update that has one recommended fix can now be installed. Category: recommended (moderate) Bug References: 1162654 CVE References: Sources used: openSUSE Leap 15.1 (src): kiten-18.12.3-lp151.2.4.1 openSUSE Backports SLE-15-SP1 (src): kiten-18.12.3-bp151.3.3.1 -- You are receiving this mail because: You are on the CC list for the bug.