I've noticed the same thing, primarily in my most-used application: FBReader. It will count a single down-tap on the D-pad as several, or as mentioned, scroll continuously until you press another key. It happens less for the center button so I've mapped that to the "next page" action. Very annoying, and all-new for OS2008.

Larry

It's been fixed in the latest FBreader version. Just select the new option for scrolling on key release instead of key press. :-)

bluehash, of course, I've forgotten. In apps installer "down" sticks very often.
Karel Jansens, screen-protector is not the reason. I have no it, but the problem exists.

Yep, this is brand new for me in OS2008. For me, it happens in MicroB all the time. Also, I can't hold scroll-up in RSS Reader AT ALL, but I can tap tab-up no problem
whatsoever.

Don't know about the App Manager, the targets are sooo small there I'm always using the stylus and the GUI to navigate.

You can vote for #2723 if you want.
https://bugs.maemo.org/show_bug.cgi?id=2723

I had opened up bug 2751 on this issue a few days ago (https://bugs.maemo.org/show_bug.cgi?id=2751). From doing a bit of research and testing, it appears to be load related -- if a key press results in a spike in cpu utilization, the key release code doesn't go through so the app thinks that the key is still pressed down. So I previously didn't notice this until I configured microb to do a page down with the d-pad.

I did a bit of additional research, and I've seen this same issue reported in various Linux distro forums (google keywords: "linux key repeats stuck"), and a common theme seems to relate to cpu load, and power management. Apparently on some portable devices the power management circuit is shared with the keyboard interrupt, so back-to-back key interrupts get lost (one interrupt for a key press, a second one for the key release). I don't know if that relates to the same probelem that we see on the Nokia devices, but it sounds plausible. Here's a thread talking about this on Ubuntu: https://bugs.launchpad.net/ubuntu/+s....15/+bug/39315

Same problem here (no screen protector).

I've been reading long PDFs in Evince, this has happened daily.

It is separate from the CPU lock I get every once in a while.
In fact, it scrolls pretty energetically, I'd say the CPU was in good form.

Screen protector? Might have been a good idea, bit late now!

I thought of remapping the key function to see if it's the switch or the driver, but Evince doesn't let me.

Sounds like you've got that covered, though, with FBReader activating on keyup. Notice anything different when you do that?

Yeah, happens with my n800 too. Guess its not an isolated machine/incident

Just wanted to chip in to say I've noticed this happen an awful lot on mine too.
Doesn't happen if I use the screen soft-scroll; Just only if I use the D-pad.

Heavily doubt it's screen-protector related 'tho (Coming from 1src.com, I've read enough screen-protector problem posts to know this doesn't fit the profile )

I suspect the OS is missing the key-up event when you release the pad, because it seems to happen most often under high-load situations, and also if you hold it down for a few seconds so that lots of screen updates are happening due to the scrolling.

MicroB is a bloody terrible app - The single biggest hog on the N800, so if my theory is correct it's no surprise that that is where the problem shows up most often.


The soft-keyboard problem mentioned in the bug report is unrelated - I suspect it's partly due to something like the coders not putting in a 'bounce' timer for the digitizer (We learned this in 1st year mechatronics; Even with an industrial microswitch, you need to put a small delay in after each 'click' before reading the input again, otherwise you get multiple clicks from spring bounce. Calibrating the bounce can be a bit trial and error 'tho.)
The fact that the digitizer is very hard, unlike PPC and PalmOS digitizers (Where the plastic 'flex' of the digitizer tends to absorb the bounce) exacerbates the problem.

The iPhopod digitizers are hard too, but don't suffer from this problem because they are capacitive (Like a laptop trackpad) and not pressure sensitive (But that means you can't use a plastic stylus or fingernail with them...)


EDIT: Well well... I've just duplicated one of the other scrolling weirdnesses.
In the File manager, I was just scrolling through a large list of MP3s. During the scroll, the cpu hit 100% (At least according to the LoadApplet meter), and then it just continued scrolling!
If I tapped the up arrow, it'd reverse direction and 'auto-scroll' upward. Click the down arrow, and it 'auto-scrolls' down again. Tapping on the scroll-bar stops it.
Significantly, it doesn't do it in 'list view' mode, only in 'thumbnail' mode (Which, annoyingly, is the only way to see the full name of a lot of the files here...)
Is this actually programmed behaviour or a bug??