Since Mandriva 2009.1 https://qa.mandriva.com/show_bug.cgi?id=48314 the titles of each screen in the installer are in uppercase letters. You can see some screen-shots here: http://docteam.mageia.nl/el/installer/content/images/ . In Greek language the capital letters are used without accents. It gives an ugly aspect which gives a first negative impression. I'd like to find a solution to have the words in capital without accents I think that it could be easy if i translate these phrases without accents in lowercase so when they will be converted they will not have accents. But i dont know if those sentences are used only in installer , or they are used in other places too which will give spelling errors. If this solution is the better one, there is a "tag" or something that i can distinct the sentences in titles in the po files ? Another example, someone to correct this in websites, did this in css : http://leonart.wordpress.com/2010/10/30/κεÏαλαίοι-ελληνικοί-ÏαÏακÏήÏεÏ-ÏÏÏί/ http://babelfish.yahoo.com/translate_url?lp=el_en&trurl=http%3A%2F%2Fleonart.wordpress.com%2F2010%2F10%2F30%2F%25CE%25BA%25CE%25B5%25CF%2586%25CE%25B1%25CE%25BB%25CE%25B1%25CE%25AF%25CE%25BF%25CE%25B9-%25CE%25B5%25CE%25BB%25CE%25BB%25CE%25B7%25CE%25BD%25CE%25B9%25CE%25BA%25CE%25BF%25CE%25AF-%25CF%2587%25CE%25B1%25CF%2581%25CE%25B1%25CE%25BA%25CF%2584%25CE%25AE%25CF%2581%25CE%25B5%25CF%2582-%25CF%2587%25CF%2589%25CF%2581%25CE%25AF%2F Another solution could be , while converting to uppercase, to use a function as above to convert the Greek letters to letters without accents.
Assignee: bugsquad => thierry.vignaudSource RPM: (none) => drakx-installer-stage2
Those CSS "fixes" really are ugly and in no way are a real fix
Priority: Normal => LowSeverity: normal => enhancement
This is not an installer bug. Both C/C++[0][1] (using glib) and perl[2] do the same thing and keep accents when using the el locale. eg: $ perl -e 'use locale; use utf8; print uc("ÎκκίνηÏη ΡÏθμιÏη ÏÎ¿Ï Î§ // ΡÏθμιÏη\n")' Wide character in print at -e line 1. ÎÎÎÎÎÎΣΠΡÎÎÎÎΣΠΤÎÎ¥ Χ // ΡÎÎÎÎΣΠWhich is the right thing according to http://en.wikipedia.org/wiki/Greek_diacritics#Upper_case [0] using wchar.h & towupper() [1] http://kakollu.blogspot.fr/2010/11/how-to-convert-utf-8-to-upper-case-in-c.html [2] perl -e 'use locale; use utf8;warn uc("ÎκκίνηÏη ΡÏθμιÏη ÏÎ¿Ï Î§ // ΡÏθμιÏη")'
Source RPM: drakx-installer-stage2 => glibcResolution: (none) => INVALIDAssignee: thierry.vignaud => bugsquadStatus: NEW => RESOLVEDComponent: Installer => RPM Packages
I think that the treatment of perl and C takes in charge character by character. While the capital letter takes an accent when it is in the first position of a lowercase word, when all letters are in capital it looses the punctuation. http://en.wikipedia.org/wiki/Greek_diacritics#Position_in_letters http://en.wikipedia.org/wiki/Capitalization#Accents
It is described well in this article: http://multilingualtypesetting.co.uk/blog/greek-typesetting-tips/
A similar case in mozilla bugzilla: https://bugzilla.mozilla.org/show_bug.cgi?id=307039
CC: (none) => almaza24map