I have used "Default" encoding many years, and never had problems with encoding detection. Since a recent commit, Konqueror fails to detect the encoding of the page, and I have to manually force Unicode detection. To reproduce, visit https://bugs.kde.org/show_bug.cgi?id=260855 and watch the accented characters in the bug title.
This is probably fixed (or worked around) by r1208522.
*** Bug 261735 has been marked as a duplicate of this bug. ***
r1208522 is completely irrelevant for KHTML.... And if you're not using KHTML you're reporting bugs in the wrong component.
I don't really know the commit number that fixed it, but it works now. If you find the commit, feel free to add it to this bug.
The bug is back.
Can confirm that, actually happens on this very big tracker.
Git commit 13770201a98d95e5d807185c80cd0bb4f31b3aa1 by Maks Orlovich. Committed on 06/03/2011 at 16:02. Pushed by orlovich into branch 'master'. Fix regression in charset parsing. Implicit LWS is permitted between tokens in key/value pairs. BUG: 260885 M +2 -0 kioslave/http/http.cpp http://commits.kde.org/kdelibs/13770201a98d95e5d807185c80cd0bb4f31b3aa1
Git commit 0b952035a61707fb870ea7239160a57fdeac631d by Maks Orlovich. Committed on 06/03/2011 at 16:02. Pushed by orlovich into branch '4.6'. Fix regression in charset parsing. Implicit LWS is permitted between tokens in key/value pairs. BUG: 260885 M +2 -0 kioslave/http/http.cpp http://commits.kde.org/kdelibs/0b952035a61707fb870ea7239160a57fdeac631d