On Fri, Mar 05, 2010 at 01:02:18AM -0800, Bryan Hundven wrote:
> >> When I draw fast with the pen the circle is more of a polygon, but
> >> when I draw the circle slowly it looks fine. Some kind of tracking
> >> issue?
> >
> > hmm. good question. Please run "xinput --test-xi2" and then move in a
> > circle. The raw events should look something like this:
> > EVENT type 17
> >    device: 2
> >    detail: 0
> >    valuators:
> >        0: 0.00 (-1.00)
> >        1: -4.00 (-9.00)
> 
> I attached 4 files. The failures are with 0.10.4 04 Mar 2010 21:49:00.
> The two issues are with the ubuntu/debian (same exact src) package
> 1:0.10.3+20100109-1ubuntu1.
> 
> When I just touch the screen or put the pen in proximity of the
> screen, input fails with:
> X Error of failed request:  BadAtom (invalid Atom parameter)
>   Major opcode of failed request:  17 (X_GetAtomName)
>   Atom id in failed request:  0xbfa17ebc                    << this is
> 0xb7740000 on touch
>   Serial number of failed request:  30                        << this
> and the next are 31 on touch
>   Current serial number in output stream:  30             <<
>                 Button labels: None None None None None None None None
> None None None None None None None None None None None None None None
> None None None None None None None None None None
> 
> The touch issue is the xi2 output while I was reproduced the touch
> issue I mentioned previously.

oh, whoops. that looks like either the driver doesn't initialize some data
or there's a bug in the xinput tool. didn't expect that to happen :)

> The pen issue is the xi2 output of the issue in bug #2952501.
> 
> This is ubuntu 10.04... Its always possible that I am behind xorg git
> master enough to get some bugs that are fixed upstream.

I don't think so, I keep Fedora relatively close to upstream and we see a
bunch of issues mainly with ISDV4 and ISDV4-like tablets. so chances are
that the bug isn't fixed yet though I'd appreciate it if you could poke the
ubuntu maintainers to update to at least 0.10.4.
 
> > where the two numbers are unaccelerated and accelerated values. Are they
> > always the same? if not, then acceleration kicks in and that's causing the
> > problem. If they are the same, it's the driver's fault.
> >
> >> I also noticed that sometimes when I flipped to the eraser that it was
> >> still drawing with the stylus. Only after using the stylus did the
> >> eraser go back to being the eraser. This might also be an issue with
> >> xournal.
> >>
> >> And after doing this test, touch stopped working. Touch returned after
> >> restarting the xserver.
> >>
> >> I checked the Xorg.0.log.old after restarting and still see:
> >>
> >> (WW) wcmSerialValidate: bad magic at 3 v=a4 l=9
> >
> > out of interest - are all these bugs introduced by this latest patch set?
> > what was the last known working version? If you have one, can you git bisect
> > down to the patch that breaks it (or introduces one of these issues)?
> 
> I only started using using ubuntu lucid in January. The version
> probably was and still is 1:0.10.3+20100109-1ubuntu1.

hmm, ok. my immediate worry is introducing new bugs. we currently have a
bunch of issues that are known or semi-known and I'm going through the code
with some general cleanup and refacturing work. I expect a fair number of
bugs getting fixed in the course of this work. Especially the ISDV4 code I
found to be quite tricky to read, so I'm trying to streamline that part a
bit.
so the bad news for you is that this bug may be present for a bit until it's
sweeped up. If you're willing to help, please let me know and I'll show you
how to get started building from git and hunting bugs down.

> EVENT type 17
>     device: 2
>     detail: 0
>     valuators:
>           0: 3078.00 (3078.00)
>           1: 1548.00 (1548.00)
>           2: 0.00 (0.00)
>           3: 0.00 (0.00)
>           4: 0.00 (0.00)
>           5: 0.00 (0.00)

these suggest that there's no accel applied to the tablet, so it's somewhere
in the driver code. not sure what yet, I need to dig into this. I thought
just purging the in-driver accel code was enough but maybe that's not the
case. thanks for providing all the info.

Cheers,
  Peter

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Linuxwacom-devel mailing list
Linuxwacom-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linuxwacom-devel

Reply via email to