new xf86-input for macbooks and other multitouch-touchpads.

apparently, i haven't installed them yet as the ubuntu help page on the mbpro says that there is already a driver. i added the mactel repository for some other dependencies (as the wireless driver), but have not yet installed any touchpad driver manually.

to make sure before proceeding: you are talking about the package bcm5974-dkms, right?

I tried to setup the parameters that you wrote (kosumi68), but nothing changed. Tapping is still bad, I tried to tapping faster/slower harder/softer but nothing happened. 1 click is two tap, and a double click is 4 tap. Right click with tapping is impossible. Thanks

Quote:

---

Originally Posted by joskapista

I tried to setup the parameters that you wrote (kosumi68), but nothing changed. Tapping is still bad, I tried to tapping faster/slower harder/softer but nothing happened. 1 click is two tap, and a double click is 4 tap. Right click with tapping is impossible. Thanks

---

Ok, not what I get here, but hey. :-)

If you feel like debugging, running the command

Code:

---

sudo ./bin/src/test /dev/input/eventX

---

will help. You need to know the event number currently assigned to bcm5974:

Code:

---

grep bcm5974 /proc/bus/input/devices -A4 | grep H:

---

and you need to have your driver set to non-grabbing (recompile, make, make install, log out, log in):

Code:

---

diff --git a/src/mtouch.c b/src/mtouch.c
index 14c0fd9..6341b56 100644
--- a/src/mtouch.c
+++ b/src/mtouch.c
@@ -21,7 +21,7 @@
 
 #include "mtouch.h"
 
-static const int use_grab = 1;
+static const int use_grab = 0;
 
 int configure_mtouch(struct MTouch *mt, int fd)
 {

---

That should make it possible to see exactly how the gestures are interpreted, and why there would be a double event happening.

Quote:

---

Originally Posted by jonas_t

you are talking about the package bcm5974-dkms, right?

---

Yes - the multitouch X driver requires this newer version of the bcm5974 kernel module since the default one shipped with Lucid doesn't contain all the required low-level multitouch event code. So you'll have no luck running multitouch without this.

Quote:

---

Originally Posted by kosumi68

Ok, not what I get here, but hey. :-)

If you feel like debugging, running the command

---

So I did feel like it ;) I'm having exactly the same problem, concerning double-tap to click. When running the test program, a single tap wouldn't output anything (besides motion). only when double tapping i'd get these lines:

Quote:

---

button bit: 0 1
button bit: 0 0
motion: 0 -3
button bit: 0 1
button bit: 0 0
button bit: 0 0
motion: -4 -3
motion: -4 -3
button bit: 0 1
button bit: 0 0
motion: 0 6
motion: 6 0
motion: 5 -3
button bit: 0 1
button bit: 0 0
button bit: 0 0
motion: 0 -3
button bit: 0 1
button bit: 0 0

---

this would have resulted in 5 clicks.
---------

I've been digging through the code, to see if I could help create a patch or something, but I need some info. I'm an embedded systems developer myself, meaning i'm used to low level stuff.. hardware interrupts. I'm not familiar with the whole event and function calls and API from this mouse driver. Is there anywhere I can find some information about this? As in.. what function is called when. And besides that, i'd need to understand how you built your driver. What files contain what functions and what function is doing what.. you lack a certain degree of commentaries haha. Besides that, i'm grateful that you're taking the effort of development!

Quote:

---

Originally Posted by Wbdsgnr

I can find some information about this? As in.. what function is called when. And besides that, i'd need to understand how you built your driver. What files contain what functions and what function is doing what.. you lack a certain degree of commentaries haha. Besides that, i'm grateful that you're taking the effort of development!

---

Look at the function read_input() in driver/multitouch.c. It is the single point of entry of input events, and following the code from there should give you a picture of how it works.

Funny then that ohloh considers the project to be well-commented ;-) Reading the git log might help.

The driver was developed to aid the development of MT drivers and to support new hardware. Any help is appreciated. That would make me grateful, too. :-)

... you are running the current git head, 71168e1fb, right?

EDIT: ok, I realize now that some of the code in the beta1 is old already, and the problem might already be fixed in the upcoming mtdev library. Will return shortly with a bugfix.

Wonderful! If that's the only entry, that would make things easier :) Yeah, no offence meant as for as the comments in the code! I see you did comment on structs and stuff in the includes. And there's enough comments in the functions themselves, but I was pointing to a little function description above the functions. But, now I know the point of entry, the function names should become pretty self-explanatory! :)

---
I now see you did do that for important functions ;) I take back my words; you've commented quite well :P

Quote:

---

Originally Posted by Wbdsgnr

Wonderful! If that's the only entry, that would make things easier :) Yeah, no offence meant as for as the comments in the code! I see you did comment on structs and stuff in the includes. And there's enough comments in the functions themselves, but I was pointing to a little function description above the functions. But, now I know the point of entry, the function names should become pretty self-explanatory! :)

---
I now see you did do that for important functions ;) I take back my words; you've commented quite well :P

---

Well, let's not get carried away. :-P

I pushed a patch to the repo, I am very interested to hear if it makes any difference.

Cheers!

Haha. Hmm.. sorry to inform you, but now no tap will trigger an event. Single, double or triple.. nothing happens!