Description of problem: If I click into vertical scrollbar of Firefox, the document may scroll varying number of pages. Version-Release number of selected component (if applicable): kernel-2.6.31-0.35.rc1.git5.fc12.x86_64 xorg-x11-server-Xorg-1.6.99-8.20090618.fc12.x86_64 xorg-x11-drv-ati-6.12.2-19.fc12.x86_64 xorg-x11-drv-synaptics-1.1.99-1.20090622.fc12.x86_64 firefox-3.5-0.21.beta4.fc12.x86_64 How reproducible: It only reporduces on one computer. Reproduction is reliable on that computer. But the problem does not occure on every click. Steps to Reproduce: 1. Open a long document in Firefox 2. Click into scrollbar below scroller Actual results: Excessive scroll Expected results: Scroll by 1 page Additional info: I only put this before the feet of synaptics tentatively, to document circumstances and software set. This may be a bug in Firefox, or a bug in X server. When I had a bug like that in synaptics before, other applications were affected, but so far it appears that no other application is (tried vim-X11, Sylpheed, evince). Also, the last time it was a logic issue in synaptics that made it forget the button status. So, the scrollbar scrolled like crazy until I moved finger or produced other event. This time it's different: the uncommanded scroll stops by itself after less than a second.
I forgot to mention that this started after the upgrade yesterday. Unfortunately, there was a big gap in upgrade sequence on this box, the previous synaptics driver was xorg-x11-drv-synaptics-1.1.0-7.fc11.x86_64, installed on May 26. Also, I tried to run synclient -m 1, but it asks for SHMconfig and I forgot how to set it.
I'd need some more details please: - how do you click? tap? physical buttons? other buttons? - does it happen with other methods of clicking (tapping, physical buttons, etc.) - can it be a timeout issue? holding the button for a tad longer causes more scrolling, so maybe you're hitting just that magical threshold (that's a timeout in firefox afaict) - does it happen with a standard usb mouse? > Also, I tried to run synclient -m 1, but it asks for SHMconfig and > I forgot how to set it. https://fedoraproject.org/wiki/Input_device_configuration, the key is input.x11_driver.SHMConfig. to narrow down whether it is in fact synaptics, you could also move /usr/share/hal/fdi/policies/20thirdparty/10-synaptics.fdi out of the way, restart HAL + X and then try again. this should load the evdev driver for the touchpad - which will feel a bit different but should still work (no tapping though). does the problem still happen?
The problem is gone after an update. kernel-2.6.31-0.42.rc2.fc12.x86_64 <----------------- changed xorg-x11-server-Xorg-1.6.99-8.20090618.fc12.x86_64 xorg-x11-drv-ati-6.12.2-19.fc12.x86_64 xorg-x11-drv-synaptics-1.1.99-1.20090622.fc12.x86_64 firefox-3.5-1.fc12.x86_64 <----------------- changed I suspect it was Firefox, or Firefox triggering something in GTK2.
Oh joy, it's still here. Apparently needs a background CPU load to reproduce reliably. Reopening... I'll address questions from comment #2 in a moment.
I'm switching the component to firefox, because: - this is reproducible with a USB touchpad - Firefox is the only application that fails It may be something in GTK2, and Firefox only triggering it, but I honestly have no idea. I need an expert to look at it. I create a CPU load by building kernels, this helps reproduce it easily. This very bug in Bugzilla serves as a test webpage.
The problem is gone with update to firefox-3.5.2-2.fc11.x86_64. However, comments by Christopher in bug 515772 suggest that it was built with old toolchain, presumably in old buildroot. So, we may almost be certain that the problem is not in Firefox as such, but in some library it uses on F12.
Something new every release. The problem is back but at least now that Firefox gets a stuck and/or lost event and starts scrolling unstoppably, it will stop (eventually) if I click like mad into the uppper region of the scrollbar. Before, it was necessary to wait until it scrolls to the bottom stop. firefox-3.5.2-3.fc12.x86_64 gtk2-2.17.6-4.fc12.x86_64 glibc-2.10.90-12.x86_64
Looks like it fixed itself now. firefox-3.5.3-1.fc12.x86_64 gtk2-2.17.11-2.fc12.x86_64 glibc-2.10.90-21.x86_64