Version: (using KDE 4.3.2) Installed from: openSUSE RPMs The web shortcuts in konqueror don't provide a possibliity anymore to use percent encoding for the search time. However, for various sites it is necessary to use percent encoding - these sites will fail with UTF8 encoding or other encodings. Example: I'm using a web shortcut to access the "Duden", the most common german ortografic dictionary. This is the search term: http://www.duden.de/suche/index.php?suchwort=\{@}&suchbereich=mixed Result when searching for "Täter": http://www.duden.de/suche/index.php?suchwort=Täter&suchbereich=mixed Expected result: http://www.duden.de/suche/index.php?suchwort=T%E4ter&suchbereich=mixed
Try to change charset to ISO 8859-1. I don't know german but seems to work. You misunderstand how this system works because you think that is a percent encoding and really is an encoding conversion problem. If you view text source result of that site you can noted that <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1" /> so the site is using iso-8859-1 charset. Then, if you send it a text in UTF-8 charset the site can't handle it, in fact this was the https://bugs.kde.org/show_bug.cgi?id=182934 bug. When you create a search engine you must select in charset the server's charset because without this information, or with a wrong one, konqueror can't convert text from your local charset to the server's one.
Thanks, it works with 8859-1. (I was searing for the percent encoding because the URL, that is generated by Duden's site itself when entering something in the search field, has percent encoding.) However, percent encoding still has a mayor advantage: When you use another browser as konqueror as standard browser, then this browser will not use 8859-1 (or whatever other encoding used in the shortcut configuration), but it's standard encoding (probably UTF8). For this reason, when I use KRunner to use a this shortcut, firefox is called and uses "Täter" with UTF8. When firefox would get the URL in percent encoded form, than this would be no problem.
Don't forget to mark this as WONTFIX and explain that is not a bug. Percent encoding is "more human reading" but, if encoding conversion is broken, don't solve anything.
Reporting Comment #2 as Bug 212736 and closing this one as wontfix.