Re: [Aiptektablet-devel] Current CVS under 2.6.6 driver. AiptekReadEvents problems?

[Bryan W. H. <bwh...@ea...>]

Darryl Luff wrote:

> Hi all,
>
> I've spent the morning trying to get the aiptek driver working with 
> the 2.6.6 kernel and XFree86.
>
> - I got the current CVS drivers (5th June) from sourceforge. To get 
> the driver to work under the 2.6.6 kernel I had to copy in an 
> implementation of msleep from another driver. I think the global 
> msleep only appears in 2.6.7?? Anyway, it then worked. evtest shows 
> Pressure events that seem to be normal.
>
> - In then looked at the XFree86 driver. It compiled OK, but in the 
> Gimp, the devices show up but don't seem to do anything. If I 
> understand it correctly, changing to the
> erasor tool in Gimp should make the Erasor active in the Device Status 
> dialog? In mine, the "Core Pointer" is always active, and the three 
> aiptek devices seem
> insensitive and never change. I can draw in the gimp but it's not 
> pressure sensitive.
>
It probably "should", yes, but doesn't. I was talking with one of the 
Wacom driver guys, and he was thinking about sending patches to Gtk/Gimp 
to fix it's behavior with tablets.

> I'm assuming that this is because the device never changes, but I dont 
> really know anything about this.
>
> - In AiptekReadEvents, the eventType field is always zero at the end 
> of the checks that use AiptekIsEventValid(common, i) to check for an 
> available driver. This is because AiptekIsEventValid returns 
> "Success", which has the value 0 on my system, so the check:
>
> if (AiptekIsEventValid(common, i))
> {
>
> always fails. I changed it to check for a "Success" return value.
> if (AiptekIsEventValid(common, i) == Success)
> {
>
> also changed the other calls to this function to be the same.


Good catch!

>
> - Now, eventType seems to be set properly. But I dont understand where 
> the distinction between erasor, stylus and cursor comes from? Do I 
> have to somehow tell the tablet to switch modes?


Yes. Write to the tool_mode file. Or use Gaiptek.

>
> - I know the loop around the AiptekIsEventValid call is trying to make 
> sure there is a valid driver, but I'm not really following the method. 
> If eventType is originally 2 (CURSOR_ID), then 'start' is set to 2. It 
> is then used to initialise 'i' to the value of deviceIdArray[start], 
> which is deviceIdArray[2], which is ERASER_ID (4). So the call to 
> AiptekIsEventValid is AiptekIsEventValid(common, ERASOR_ID). This 
> returns Success, so the eventType is set to ERASOR_ID?
>
> - Shouldnt the line "if (deviceIdArray[start] != deviceIdArray[i])" 
> that triggers the re-routing warning, be "if (deviceIdArray[start] != 
> i)", as 'i' was already set using "i = deviceIdArray[j]".


This is why I get headaches...

common->currentValues.eventType = 1 << i;

>
> Thanks for any pointers!

Two quality patches; what pointers do you think you need?