Version: (using KDE KDE 3.5.5) Installed from: SuSE RPMs The Unicode table for Arabic characters has a list of combining characters, like ً == fathatan ٌ == dammatan ٍ == kasratan َ == fatha ُ == damma ِ == kasra ّ == shadda ْ == sukun Konqueror displays these combining characters perfectly for every character that is not the last character in the string. Here's an example: <html> <body> <p>حُبٌّ</p> </body> </html> This is a word pronounced like 'Hubboon' and means 'love'. The first character is the 'hah' (ح). It's that hook at the right. Remember we write Arabic from right to left. On the 'hah' (pronounced as a hard 'H') there's a damma (ُ), a combining character. It looks a little bit like a '9'. In English you pronounce it as a short 'oo'. Then there's the 'beh' (ب), pronounced as a 'B' with two combining characters, (1) the shadda (looks like a 'w'; you must pronounce the character twice 'bb') and (2) the dammatan (looks like '2e' or as '99'; must pronounce it like 'oon'). As you see if you save the html example as a file and open the document in the Konqueror, the damma is drawn above the 'hah' perfectly, but the shadda and the dammatan appear before (right of) the 'hah' where they should appear above the 'beh'. For every combining character for the last character of a string, the combining characters appear before the string. I see the same behaviour with Konqueror under Debian.
Last week I saw one of my HTML pages printed from a Safari browser under Mac OS/X. Safari has this problem too, so it's a bug in khtml. If a longer sentence wraps to the next line, and both lines end with a combining character, both lines start with the combining character for the last character. I'll send a bug report to Apple too, with an URL to this page.
I can confirm this bug. It also appears in Hebrew, where it is easier to follow because the script is not cursive. Add any of the diacritics, e.g., segol, to any of the characters, e.g., tsadi. צֶ and you see that the segol (three dots below) appear right of the tsadi instead of directly below. Now add any letter, e.g. another tsadi (צֶצ), and you see that (from right to left) a tsadi with a correctly placed segol followed by an undecorated tsadi. The second rendering is correct, the first should look like the right half of the first. An example plus screenshot is here: http://bthec11.uni-graz.at/~katzer/hebrew_diacritic_konqueror_bug.html http://bthec11.uni-graz.at/~katzer/hebrew_diacritic_konqueror_bug.png
i also can reproduce this bug using '3.5.7 (KDE 3.5.7 "release 72.6" , openSUSE 10.3)'. Additional info: It doesn't happen with the form tag 'input', but inside of normal divs. In case you need a test page, i offer you http://www.mrt-gutesundneues.de/script/arabconn.php, where i do my first arabic steps ;-)
Created attachment 24452 [details] wrong one wrong one with konqueror kde 4.00.69 (KDE 4.00.69 >= 20080410)
Created attachment 24453 [details] right one
I can confirm this but with konqueror kde 4.00.69 (KDE 4.00.69 >= 20080410)
Still present in konqueror svn trunk 813735
As a workaround for this bug you can add dir=rtl e.g. <html> <body> <p dir=rtl>حُبٌّ</p> </body> </html>
Message from the Bugsquad and Konqueror teams: This bug is closed as outdated, as we do not have the manpower to maintain the KDE3 version anymore. If you still can reproduce this issue with Konqueror 4.8.4 or later, please open a new report. Thank you for your understanding.