Version: 3.4.0 (using KDE 3.4.0, compiled sources) Compiler: gcc version 3.4.3-20050110 (Gentoo Linux 3.4.3.20050110, ssp-3.4.3.20050110-0, pie-8.7.7) OS: Linux (i686) release 2.6.11-gentoo-r3 Steps to recreate: 1. Browse to a webpage that overflows past the bottom of the screen (i.e., requires scrolling). 2. Initiate the "auto-scroll" mode by clicking the center mouse button (Note: this bug does not exist with the shift+arrow mode). 3. Move the mouse cursor below the initiation point to start scrolling down. 4. Move the mousewheel either up or down. Results: The page will start scrolling up instead of down. The page will start scrolling down again as soon as the mouse is moved (as long as the cursor is under the initiation point).
Some additional findings: When scrolling UP (in auto-scroll), moving the mousewheel won't reverse the scroll direction, however it will increase the scroll speed. The scroll speed will be properly set as soon as the mouse is moved again.
Where did you set mousewheel to autoscroll? It does "paste selection" or "back in history" here.
This is a pretty old report. :) Under Konqueror's settings -> Web Behavior, uncheck the "Middle click opens URL in selection" option, and then the middle click initiates autoscroll. (using Konqueror 3.5.2) On 1 Sep 2006 18:09:41 -0000, Philip Rodrigues <phil@kde.org> wrote: [bugs.kde.org quoted mail]
Ah, neat. I can't reproduce the behaviour you describe on KDE 3.5 though - in autoscroll mode, the mouse wheel just "jumps" the page up or down as appropriate, and then the page carries on scrolling as it was before. Have you tested with 3.5?
I experience the same behaviour in KDE 3.5.4. Note that to reproduce this, you cannot move the mouse cursor after moving the mouse wheel.
Now I *can* reproduce it. Odd.
For me, mouse wheel in autoscroll mode jumps to the top of the page, pretty much. No matter where the wheel is turned.
This bug had disapeared for me in KDE 3.5.7, and it reapeared in KDE 3.5.8. Very annoying.
Confirmed on 3.5.9. It works right on 4.1 (trunk r802881). Should we set this bug as "LATER" ?
i agree with finex