[Basilisk]

Hi,

I'm working on a new implementation of trackpad scrolling for PowerBooks (I'm the guy who wrote trackpad scrolling in uControl). I have a few questions for potential users that will help me polish the project:

1. Is your PowerBook multiuser (i.e. regularly has more than one user)?

2. If the answer to question 1 is yes, how important is it that different users have independent trackpad settings? Bear in mind that Panther's fast user switching may change your usage.

3. How important is it that you be able to disable the replacement trackpad driver from a pref pane without a reboot? i.e. If you are a uControl user do you ever turn off uControl manually? If a reboot with Shift key down (Safe Boot) was required to disable the replacement driver would this be acceptable?

4. If you use uControl's trackpad scrolling implementation what scrolling scale setting do you use? The default or have you changed it?

5. Do you actively use the "Ignore USB mouse" setting of the mouse preference panel?

Please note that these questions are not hypotheticals, I have a working driver right now. I'm not, however, ready for beta testers so please don't ask (I'll announce when I'm ready for testing).

Thanks,

Bas

http://www.ragingmenace.com

[Basilisk]

No takers? No one uses uControl's trackpad scrolling?

[ngrundy]

I do, and i do use the 'ignore trackpad while typing' option.

I keep uControl loaded all the time

[Musti]

3.

No I don't turn it off. And yes, if you want to make it a part of the system, go ahead.

4.

For some reason the trackpad scrolling speed does not stick if I change it. So I use the default.

5.

Sometimes.



Thanks

No, thank YOU for that scrolling thingy. That is the among the first five things I install right after a fresh OS X installation, even though I mainly use a mouse.

[geekwagon]

1. No
3. Not very. Nice to have, but it wouldn't keep me from using it. I never turn uControl off manually (never needed to)
4. Default
5. I don't use the ignore trackpad while USB mouse is attached. I DO use ignore trackpad while typing.

I'm assuming this is an implementation that uses the right of the pad rather than the fn key, similar to the Synaptics drivers for Windows? That would be awesome.

[Basilisk]

I'm assuming this is an implementation that uses the right of the pad rather than the fn key, similar to the Synaptics drivers for Windows? That would be awesome.

Correct. Its not feature-equivalent with the Synaptics drivers becuase they include some features (like auto centering the pointer on the default dialog button) that can't be done at the driver level in MacOS X.

That said the driver does feature:
- Optional vertical scrolling at left or right edge of pad.
- Optional horizontal scrolling at top or bottom edge.
- Map hardware button to left or right click.
- Map trackpad taps to no action, left click, left click drag (with or without drag lock), or right click.
- Complete uControl compatibility (you can still use uControl's scrolling to turn the whole pad into a scroll surface).

For the techincally curious the reasons these questions are relevant (particularly number 3 and 5) are that the driver is a replacement for Apple's driver, not an inline hack like uControl. This should make it more stable when Apple revises the interface (it uses only documented APIs), but makes it much more difficult to reliably deactivate/activate it. Throw in Panther's fast user switching and it becomes even harder to offer a driver toggle option.


Thanks for all the responses, and keep em coming,

Bas

[geekwagon]

Wont a replacement for the Apple driver conflict with other companies mouse drivers like Logitech's? It seems like they also replace Apple's drivers, because the "ignore trackpad on typing" option doesn't work when they are installed. Or is the trackpad a seperate driver than the USB mouse? I would hate to lose the ability to map my other buttons but I guess I could always use USB Overdrive for that. Also, please mimic the acceleration/ballistics settings that Apple is using in Panther rather than what they use in Jag.. Much improvement there, IMHO.

Sounds like you are implementing all the features of the Synaptics drivers that I would want. I hope you also leave off some features that I experience with them on my work provided Dell laptop:
- neither mouse nor trackpad work after waking from sleep (have to dig into the control panels and hit the "reset driver" button..)
- slower than hell scroll wheel on external mouse with no option to speed it up

[Paul Guyot]

1. This is *my* powerbook.

3. You're sure you cannot do it otherwise? I can try to help if you want.

4. The setting actually doesn't work with the current version. I had speed set to the maximum with earlier versions.

5. Only ignore trackpad while typing, although I wish it worked better.

Paul

[Basilisk]

Wont a replacement for the Apple driver conflict with other companies mouse drivers like Logitech's?

Nope, USB mice and trackpad are totally separate drivers. No conflict.

It seems like they also replace Apple's drivers, because the "ignore trackpad on typing" option doesn't work when they are installed.

All this means is that the Apple trackpad driver doesn't recognize the (separate) Logitech driver as a USB mouse. This might be a bug in Apple's driver, or just something unexpected in Logitech's driver.

I don't implement "ignore USB mouse" in my first version for precisely this reason. Its hard to catch all the cases unless you have every conceivable mouse to test with.

Also, please mimic the acceleration/ballistics settings that Apple is using in Panther rather than what they use in Jag

I'm using a curve based on the Jaguar curve. I don't detect any differences in Panther, but I'll look into it.

3. You're sure you cannot do it otherwise? I can try to help if you want.

I have a workaround I use during testing, its just very hacky. Toggling the driver involves a second kernel extension to cleanly unload the original Apple driver. This requires the use of IOKit internal structures which may change, and thus makes the whole thing more fragile in respect to OS versions.

The reason I asked the question is to get a sense of what folks really prefer. If I don't give the option to toggle directly then some flexibility is lost, but unlike uControl, it is reasonable to expect the driver will work on any version of 10.2 or 10.3 without modification. That seems like a big win to me.

Bas

[geekwagon]

Originally posted by Basilisk:
All this means is that the Apple trackpad driver doesn't recognize the (separate) Logitech driver as a USB mouse. This might be a bug in Apple's driver, or just something unexpected in Logitech's driver.

I hadn't thought of it that direction. I've never done kernel-module level programming so I have never looked into how this stuff works.

I'm using a curve based on the Jaguar curve. I don't detect any differences in Panther, but I'll look into it.

It could be that the trackpad driver is unchanged between the two. Now that I think about it, I don't think I have ever used the trackpad while booted into Panther (because it is installed on an external disk I never boot it unless I am at my desk.) External mice are definitely different though, and the breakpoint between the ultra-slow and ultra-fast speeds feels to be at a better spot. That point is the main problem I have in Jaguar and previous.

[Musti]

I am looking forward to what you're working on right now...The ability to scroll without the Fn key and the other stuff...using the right/left edge of trackpad...

[Chuckit]

Originally posted by Basilisk:
Toggling the driver involves a second kernel extension to cleanly unload the original Apple driver. This requires the use of IOKit internal structures which may change, and thus makes the whole thing more fragile in respect to OS versions.

Why not just have kextload/kextunload do it for you? I imagine it does it the same way you're doing it, but if Apple changes their internal mojo, they'll have to update those tools as well.

[Basilisk]

Why not just have kextload/kextunload do it for you? I imagine it does it the same way you're doing it, but if Apple changes their internal mojo, they'll have to update those tools as well.

Becuase the problem isn't kextload/kextunload. The problem is that the Apple driver has a bug that causes it not to unregister from its provider nub. This gets cleaned up when the replacement driver loads, but during the transition (which is not instantaneous, kextload takes a second or so) ANY contact with the pad will cause a kernel panic.

This made for lots of fun when I was debugging at first. The hack I wrote watches for the Apple driver to unload and properly forces the provider nub to unregister. This makes my life easier when debugging, but as I said, its a hack, and unless the ability to dynamically unload is a make-or-break deal for most folks, I'd rather not ship it.

Bas

[jasong]

First off, I love uControl, and I LOVE the idea of what you are trying to do.

1. No

2.

3. I have never manually disabled uControl before. It would be great to keep the ability to change the driver on the fly, but I don't think I would ever use it.

4. Default

5. I do not use the ignore trackpad when mouse is present feature, but I do ignore the trackpad while typing.

Hope to see this soon.

-- Jason

[JasonQG]

Oh, man. I can't wait for this update. I hope the wait won't be too long.

1. no
2. N/A
3. If it works properly, I won't ever want to disable it.
4. Default
5. I never use a USB mouse