Bugzilla – Bug 128686
[susehelp] incorrect chars in czech links
Last modified: 2008-12-11 11:41:56 UTC
In czech environment (locale cs_CZ.UTF-8) in SuSE Help center, when search results are displayed, there are correct czech characters (diacritics) in "Pre-made" parts of the result page, but in >>links<< to the pages, there are czech characters replaced by (probably) unicode escape sequences. Seems like 2 different encodings are used in one document or that help browser doesn't interpret the esc.sequences correctly Steps to reproduce (probably requires czech environment AND CZECH SUSE DOCUMENTATION "Start-Up (cs)"): 1. Open SuSE help center 2. into search enter: "mp3" 3. do search wrong results are: Výsledky hledání 'mp3': SUSE Linux Documentation (cs) Start-Up (cs) 6.3. Důležité nástroje a postupy (09/07/05) 7.3. Důležité nástroje (09/07/05) Slovník pojmů (09/07/05) SUSE Linux Documentation (en) Start-Up (en) 7.2. Konqueror as a File Manager (09/12/05) 8.3. Important Utilities (09/12/05) 7.3. CDs: Playback and Ripping (09/12/05) Chapter 14. Using Beagle (09/12/05) Glossary (09/12/05) Reference (en) 7.2. Konqueror as a File Manager (09/12/05) 8.3. Important Utilities (09/12/05) 7.3. CDs: Playback and Ripping (09/12/05) Chapter 14. Using Beagle (09/12/05) Glossary (09/12/05) Application Manuals KDE Application Manuals k3b (07/26/05) amarok (09/04/05) kaffeine (09/04/05) kmid (05/23/05) krusader (04/09/05) Kořen dokumentace Poznámky k vydání SUSE Linux Documentation (cs) 6.3. Důležité nástroje a postupy (09/07/05) 7.3. Důležité nástroje (09/07/05) Slovník pojmů (09/07/05) SUSE Linux Documentation (en) 7.2. Konqueror as a File Manager (09/12/05) 8.3. Important Utilities (09/12/05) 7.3. CDs: Playback and Ripping (09/12/05) Chapter 14. Using Beagle (09/12/05) Glossary (09/12/05)
Cornelius, there seems to be a problem with the cs environment.
ping
Reassign to new susehelp maintainer.
*** Bug 210895 has been marked as a duplicate of this bug. ***
Lukas, is this still a problem in 10.2 or SLE10?
Both 10.2 and SLED contains no Czech suse documentation (well, it has translated categories, etc, but all help itself is in English) However on kde SLED there is a problem in exactly same situation: in kde susehelp with czech locale, one of the caption is displayed incorrectly (white square instead of accented char): There is: Ko[[this represents white square]]en There should be: Kořen
Please, provide a screenshot. In general it looks fine to me (SLED10 SP1, taken one of the late RC cantidates). See attachment.
Created attachment 143247 [details] khelpcenter (cs)
Do I have to install a czech manual to reproduce the problem?
I had similar problem under _Hungarian_ locale, which has been reported as Bug #210895 for openSUSE 10.2 (linked above). On the screenshot I posted "Fejlesztés" means in fact "Development" and only some of the special characters were effected. Therefore likely not the Czech locale itself is defected (or not only that one)... Maybe this info helps you to trace back the problem.
I guess either the font lacks the wanted character (but isn't it considered to try other fonts as fallback?) or one of the involved .desktop or .directory files isn't properly UTF-8 encoded. We probably have to check case by case. On my 10.2 system, /usr/share/susehelp/meta/Development/.directory containing Comment[hu]=Dokumentáció fejlesztőknek looks fine, so I'd like to blame the font... I'm inclined to assume it is the same with Czech. See attachment.
Created attachment 143442 [details] Hungarian variant Green square with "u with oblique double quote" is fine; the black circle with a placeholder is not what we would like to see.
Created attachment 143444 [details] now the one with markers...
You do not need to install the Czech guides. The problem is possible to reproduce with the Czech Release Notes. Install Czech Release Notes and use keyword "Instalace", see screenshot.
Created attachment 143488 [details] Czech screenshot
Yes, I was able to reproduce it. But the string in the update-desktop-files.cs.po is fine: /usr/share/susehelp/meta/release-notes.desktop #: /usr/share/susehelp/meta/release-notes.desktop msgctxt "Comment(release-notes.desktop)" msgid "" "A short description of the most important changes for this SUSE Linux release" msgstr "Krátký popis nejdůležitějších změn v této verzi SUSE Linuxu" Probably a font issue. I'm running khelpcenter on top of GNOME, does it also happen in KDE?
Yes, the same happens in KDE as well! e.g. my previously mentioned similar issue on Hungarian locale was screenshot-ed while working under KDE.
Created attachment 143607 [details] czech-screen-shot-detail.png This part from Klara's screen shot looks like an encoding problem, *not* a font problem.
The files RELEASE-NOTES.cs.{html,rtf} in openSUSE 10.2 seem to be OK though: mike@nozomi:~$ ll /usr/share/doc/release-notes/openSUSE/RELEASE-NOTES.cs* -rw-r--r-- 1 root root 14989 2006-12-14 03:16 /usr/share/doc/release-notes/openSUSE/RELEASE-NOTES.cs.html -rw-r--r-- 1 root root 12391 2006-12-14 03:16 /usr/share/doc/release-notes/openSUSE/RELEASE-NOTES.cs.rtf mike@nozomi:~$ grep Pozn /usr/share/doc/release-notes/openSUSE/RELEASE-NOTES.cs* /usr/share/doc/release-notes/openSUSE/RELEASE-NOTES.cs.html:<title>openSUSE 10.2 Poznámky k vydání</title> /usr/share/doc/release-notes/openSUSE/RELEASE-NOTES.cs.html: <h1>openSUSE 10.2 Poznámky k vydání</h1> /usr/share/doc/release-notes/openSUSE/RELEASE-NOTES.cs.html: <p>Poznámky k vydání jsou neustále doplňovány. Nejnovější verzi najdete na <a href="http://www.suse.com/relnotes/i386/openSUSE/10.2/RELEASE-NOTES.en.html">http://www.suse.com/relnotes/i386/openSUSE/10.2/RELEASE-NOTES.en.html</a>. Vzhledem k častým aktualizacím anglické verze, ze které je překládána česká verze, může dojít k tomu, že překlad může být po krátkou dobu nekompletní.</p> /usr/share/doc/release-notes/openSUSE/RELEASE-NOTES.cs.rtf: <h1>openSUSE 10.2 Poznámky k vydání</h1> /usr/share/doc/release-notes/openSUSE/RELEASE-NOTES.cs.rtf: <p>Poznámky k vydání jsou neustále doplňovány. Nejnovější verzi najdete na <tt>http://www.suse.com/relnotes/i386/openSUSE/10.2/RELEASE-NOTES.en.html</tt>. Vzhledem k častým aktualizacím anglické verze, ze které je překládána česká verze, může dojít k tomu, že překlad může být po krátkou dobu nekompletní.</p> mike@nozomi:~$
Agreed. This (Klara's example) looks as if the indexer got something wrong. I checked .desktop files only and those are properly encoded, AFAICS.
As the files seem to be OK, there seems to be a problem with the generation of the results page.
I don't see this happening anymore, do you Klara?
still problems with encoding in 10.3 - attaching screenshot. Steps to reproduce same as in Comment #14
Created attachment 175957 [details] czech 10.3 screenshot
will handle this in the jan susehelp sweep
please review if this still an issue and close if fixed. moving to 10.3
Unable to reproduce it anymore in os10.3