[jtablet-dev] When will the alpha alert dialog go away?

[Garret Cole]

Out of curiosity, how close are we to turning alpha into a beta (gmail style..)? Curious when to expect that dialog that pops up when you intialize the tablet property to go away from the build.

--
You received this message because you are subscribed to the Google Groups "jtablet-dev" group.
To post to this group, send email to jtabl...@googlegroups.com.
To unsubscribe from this group, send email to jtablet-dev...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/jtablet-dev?hl=en.

[Marcello Bastéa-Forte]

Well, we've only had one alpha so far. :-) I haven't had a chance to
put together a second alpha. At the very least we'll have an alpha
with the linux version and windows 64-bit versions before any betas or
release candidates.

I would like to do another pass over the API and remove things that
aren't critical or are confusing, since this is the last real chance
to remove stuff.

How has the alpha been working for everyone?

Marcello

p.s. I'm glad the popup is doing its job of reminding people this
isn't a final release. :-)

[Thor Harald Johansen]

For me, seemingly well. I have a little thing I'm wondering about
though... Not sure if this is a bug specific to the Bamboo Fun or
something with JTablet... But window switching seems to momentarily kill
pressure sensitivity. Full pressure is measured for the first stroke.
Then normal functionality resumes for the 2nd stroke. Any idea of what
could be causing this?

Thor

[Marcello Bastéa-Forte]

Does this happen on the demo app? Can you paste the TabletEvents that
occur around the time of interest? (I'm assuming you're on Windows.)

Marcello

[Garret Cole]

Marcello, this is the bug I was emailing you about previously. The bug also exists in jtablet1 and inside the demo app. You reproduce it by focusing the application that is using the jtablet appi with your tablet, and the tablet events that are reported are all non-pressure reported. I believe the device ID is UNKNOWN at that time.

Garret

[Thor Harald Johansen]

Yeah. I forgot to mention, it exists in JTablet1 as well. Doesn't seem
to noticably affect other WINTAB applications, or Linux for that matter,
as far as I can tell. I have a vague recollection of openCanvas being
afflicted with the same problem, though. I'll do some more testing and
get back to you with results.

Thor

> <riftg...@gmail.com <mailto:riftg...@gmail.com>>

> >> wrote:
> >>>
> >>> Out of curiosity, how close are we to turning alpha into a beta
> (gmail
> >>> style..)? Curious when to expect that dialog that pops up when you
> >>> intialize
> >>> the tablet property to go away from the build.
> >>>
> >>> --
> >>> You received this message because you are subscribed to the
> Google Groups
> >>> "jtablet-dev" group.
> >>> To post to this group, send email to

> jtabl...@googlegroups.com <mailto:jtabl...@googlegroups.com>.

> >>> To unsubscribe from this group, send email to
> >>> jtablet-dev...@googlegroups.com

> <mailto:jtablet-dev%2Bunsu...@googlegroups.com>.

> >>> For more options, visit this group at
> >>> http://groups.google.com/group/jtablet-dev?hl=en.
> >>>
> >>
> >
> > --
> > You received this message because you are subscribed to the
> Google Groups
> > "jtablet-dev" group.
> > To post to this group, send email to jtabl...@googlegroups.com

> <mailto:jtabl...@googlegroups.com>.

> > To unsubscribe from this group, send email to
> > jtablet-dev...@googlegroups.com

> <mailto:jtablet-dev%2Bunsu...@googlegroups.com>.

> > For more options, visit this group at
> > http://groups.google.com/group/jtablet-dev?hl=en.
> >
> >
>
> --
> You received this message because you are subscribed to the Google
> Groups "jtablet-dev" group.
> To post to this group, send email to jtabl...@googlegroups.com

> <mailto:jtabl...@googlegroups.com>.

> To unsubscribe from this group, send email to
> jtablet-dev...@googlegroups.com

> <mailto:jtablet-dev%2Bunsu...@googlegroups.com>.

[Thor Harald Johansen]

Okay.

Some observations: With JTablet1 the error manifests as MINIMUM pressure
on the FIRST stroke (pen down, motion, pen up). I had installed, but
forgotten to remove JTablet1 earlier, to test if both versions work with
Sketcher2. I removed every trace of JTablet1 just now, and installed
JTablet2 again. The same problem exists, but now manifests as MAXIMUM
pressure on the first stroke.

Here is the output from the demo applet, demonstrating what happens as
the pen is lifted from the tablet for the first time after the
application window has been refocused.

cello.jtablet.event.TabletEvent[DRAGGED,when=1272478568590,device=SystemDevice[UNKNOWN],x=366.0,y=88.0,modifiers=Button1,pressure=1.0]
on DemoSurface@10519800
cello.jtablet.event.TabletEvent[DRAGGED,when=1272478568606,device=SystemDevice[UNKNOWN],x=370.0,y=86.0,modifiers=Button1,pressure=1.0]
on DemoSurface@10519800
cello.jtablet.event.TabletEvent[DRAGGED,when=1272478568606,device=SystemDevice[UNKNOWN],x=374.0,y=85.0,modifiers=Button1,pressure=1.0]
on DemoSurface@10519800
cello.jtablet.event.TabletEvent[RELEASED,when=1272478568622,device=SystemDevice[UNKNOWN],x=374.0,y=85.0]
on DemoSurface@10519800
cello.jtablet.event.TabletEvent[MOVED,when=1272478568622,device=SystemDevice[UNKNOWN],x=380.0,y=84.0]
on DemoSurface@10519800
cello.jtablet.event.TabletEvent[ENTERED,when=1272478568637,device=WinTabDevice[STYLUS-4022-1],x=380.0,y=84.0]
on DemoSurface@10519800
cello.jtablet.event.TabletEvent[MOVED,when=1272478568653,device=WinTabDevice[STYLUS-4022-1],x=383.0,y=85.0,sidePressure=NaN,rotation=NaN]
on DemoSurface@10519800
cello.jtablet.event.TabletEvent[LEVEL_CHANGED,when=1272478568653,device=WinTabDevice[STYLUS-4022-1],x=383.0,y=85.0,sidePressure=NaN,rotation=NaN]
on DemoSurface@10519800
cello.jtablet.event.TabletEvent[MOVED,when=1272478568668,device=WinTabDevice[STYLUS-4022-1],x=386.0,y=88.0,sidePressure=NaN,rotation=NaN]
on DemoSurface@10519800

A piece of information that may or may not be significant: With the
JTablet2 Demo Applet open in one tab of Firefox, and an arbitrary
website open in another tag, the same problem can be reproduced, simply
by switching tabs back and forth and attempting to draw.

For consistency, I opened Adobe Photoshop CS3 and performed the same
window switching test, and experienced no issues with pen pressure.

Thor

[Marcello Bastéa-Forte]

Ok, so the problem is it's not detecting the tablet right away, and
you're seeing the default mouse listener behavior in effect. I hope
this isn't a limitation of wintab... I have a couple questions:

Does it happen when running as an application (vs applet)?

What happens at the beginning of that first UNKNOWN device stroke?
What happens when you press down again after the events you pasted?

Are you focusing the app first? Or is the first stroke drawing onto
the app that is failing? Does alt tabbing first not cause the problem?
Or clicking on the titlebar?

Last of all, if you just hover on the app, do you get any events, or
are they all UNKNOWN device?

Marcello

[Thor Harald Johansen]

Marcello Bastéa-Forte wrote:
> Ok, so the problem is it's not detecting the tablet right away, and
> you're seeing the default mouse listener behavior in effect. I hope
> this isn't a limitation of wintab... I have a couple questions:
>
> Does it happen when running as an application (vs applet)?

There is a difference, as explained below.

> What happens at the beginning of that first UNKNOWN device stroke?
> What happens when you press down again after the events you pasted?

Upon first entry of cursor, device is UNKNOWN (this is for APPLETS):

cello.jtablet.event.TabletEvent[ENTERED,when=1272480221646,device=SystemDevice[UNKNOWN],x=223.0,y=106.0]
on DemoSurface@24562387
cello.jtablet.event.TabletEvent[MOVED,when=1272480221662,device=SystemDevice[UNKNOWN],x=220.0,y=98.0]
on DemoSurface@24562387
cello.jtablet.event.TabletEvent[MOVED,when=1272480221677,device=SystemDevice[UNKNOWN],x=219.0,y=91.0]
on DemoSurface@24562387

Upon drawing, device remains UNKNOWN until button is released, as shown
before. Then the ENTRY of the tablet device is registered, and normal
functioning, as one'd expect, begins.

> Are you focusing the app first? Or is the first stroke drawing onto
> the app that is failing? Does alt tabbing first not cause the problem?
> Or clicking on the titlebar?

Alt+Tab or any kind of task switching method does not improve it for
applets.

> Last of all, if you just hover on the app, do you get any events, or
> are they all UNKNOWN device?

Hovering on the APP provides tablet events. Hovering on the APPLET
provides UNKNOWN events.

The only way I am able to provoke this to happen to the test APP is by
making the first click on the window surface, the focusing click, be the
same click that starts the stroke. A small glitch like that is probably
within the acceptable, though.

This contrasts to applets, where, even after various forms of focusing
the window has been tried, the problem is consistantly reproducable.

Thor

[Marcello Bastéa-Forte]

Ok, I did a little digging. It doesn't seem to be anything in
JTablet's code... I'm simply not getting any packets from WinTab at
all in this situation. The JPen's demo app also has a similar problem.

My test case is:
1. run the demo app as an application
2. alt-tab to another window
3. then click and drag on the demo app window
The problem: No wintab packets are generated until I let go of that
initial drag.

Some variant of this is likely happening in the applet, but more
obvious since all applets require a click to gain focus.

I'm not sure how to fix this problem. Open to ideas! :-)

Marcello

[Garret Cole]

Can you briefly explain what wintab is? Is it the native driver that you are hooking with?

Also, for the time being, is it possible that you can identify when this is happening? If its possible that we could receive some indication or flag of this happening when the event is generated, I think we could handle it at the application layer.

Garret

[Marcello Bastéa-Forte]

On Wed, Apr 28, 2010 at 11:25 PM, Garret Cole <riftg...@gmail.com> wrote:
> Can you briefly explain what wintab is? Is it the native driver that you are
> hooking with?

Read up on it here: http://www.wacomeng.com/windows/

> Also, for the time being, is it possible that you can identify when this is
> happening? If its possible that we could receive some indication or flag of
> this happening when the event is generated, I think we could handle it at
> the application layer.

Unfortunately there's no good way to do this. It's impossible to
distinguish from regular mouse events (which is why this is
happening). If there were a good way to detect it, I'd prefer to solve
it at the JTablet level rather than require every single developer to
deal with it independently.

Marcello

[Marcello Bastéa-Forte]

Thanks to Nicolas, I found a workaround for this bug! 

You can go download the latest intermediate build (v1.2.4) here that should fix the problem:

http://groups.google.com/group/jtablet-dev/browse_thread/thread/89ba16884cc2b995

Please go try it out and report back your findings.

Marcello

[Garret Cole]

revisiting this: when will the dialog for alpha go away? I have a few developers who are feeling uncomfortable putting it into their program.