[19:37] <cr3> has anyone used a wacom bamboo tablet as a multi touch device? is there anything special I need to do?
[19:44] <cr3> running mtdev-test /dev/input/event12 returns nothing after supported mt events :(
[19:45] <cr3> however, the hardware support page seems to indicate that wacom usb has been supported since 11.04: https://wiki.ubuntu.com/Multitouch/HardwareSupport
[20:04] <dandrader> not me
[20:06] <cnd> cr3: the bamboo touch should work out of the box as a touchpad
[20:06] <cnd> which means that X synaptics will grab the dev event node
[20:07] <cnd> so mtdev and any similar programs like evemu and evtest will not work unless you stop X or vt switch away
[20:13] <cr3> cnd: it should work as a touchscreen rather than a touchpad, right? ie, I shouldn't expect to do two finger scrolling, not that I know how to get it working with my finger anyways
[20:13] <cnd> cr3: you can make it behave like a touchscreen if you want
[20:13] <cnd> but by default it behaves like a touchpad
[20:13] <cnd> because it isn't a touchscreen
[20:14] <cnd> and the bamboo touch line is usually a bit too small to actually behave like a touchscreen
[20:15] <cr3> cnd: this is a bamboo tablet (mte-450) though, it behaved like a touchscreen when I plugged it in: http://www.tablet4u.co.uk/product/en/bamboo-tablet.html
[20:15] <cnd> cr3: when you say it behaved like a touchscreen, what exactly do you mean?
[20:15] <cnd> maybe I am misunderstanding
[20:18] <cr3> cnd: heh, and maybe I'm not expressing myself well because this is all new to me :) first, in system settings, it appears as a wacom graphics tablet. second, the tablet only responds to the stylus so I can't imagine how it could be used as a touchpad
[20:19] <cnd> cr3: in X terms, we use "direct" and "indirect" input devices
[20:19] <cnd> a direct input device, like a touchscreen, makes the cursor move to wherever you touch on the device
[20:19] <cnd> if you touch the top left corner, the cursor will move to the top left corner of the screen, no matter where it was previously
[20:20] <cnd> an indirect device will behave like a touchpad, where the cursor movement is controlled by relative motion of the input device
[20:20] <cnd> an indirect device is also a relative device, like a mouse
[20:20] <cnd> in fact, indirect is very similar to "relative", and direct is very similar to "absolute" positioning
[20:21] <cnd> I'm not exactly sure how your bamboo tablet would behave, I was thinking about my bamboo touch, which only has touch input (no pen)
[20:21] <cnd> I actually do expect your tablet to behave like a touchscreen, but we'd have to check to be sure
[20:22] <cnd> if it's behaving like a touchscreen, then it should work with mtdev
[20:22] <cnd> can you run xinput --list --all and pastebin the results?
[20:22] <cr3> cnd: I could try mtdev-test in a vt, one sec
[20:22] <cnd> ok
[20:23] <cr3> cnd: ok, stuff is happening when I run mtdev-test, but I don't see the usual ABS_MT_* at the top of running mtdev-test
[20:24] <cr3> cnd: as for running xinput --list --all, that returns: unable to find device --all
[20:24] <cnd> then my guess would be that your tablet isn't being recognized as a multitouch device
[20:24] <cnd> if it only has pen input, no touch input, then it's just single touch input anyway, right?
[20:25] <cr3> cnd: it could be single touch input, I was hoping it might be more to touch real multitouch devices
[20:25] <cnd> xinput --list --long is what we need
[20:27] <cr3> cnd: http://pastebin.ubuntu.com/1187794/
[20:28] <cnd> cr3: yeah, no multitouch :(
[20:28] <cnd> it's just a singletouch pen tablet
[20:28] <cr3> cnd: do you think that the bamboo tablet is essentially like a touchscreen? in other words, are touchscreens multitouch?
[20:29] <cnd> a touchscreen can be single-touch
[20:29] <cnd> so your tablet is like a singletouch touchscreen
[20:31] <cr3> cnd: when a touchpad supports two finger scrolling, is that reall multi touch?
[20:31] <cr3> s/reall/really/
[20:31] <cnd> it depends
[20:32] <cnd> it might be multi-finger singletouch
[20:32] <cnd> where it reports only one location, but guesses the number of touches on the trackpad
[20:32] <cnd> or it may be full multitouch
[20:41] <cr3> cnd: so, on the CheckingMTDevice wiki page, it says to use mtdev-test to test for multi touch. however, I ran it on a few laptops, some of which didn't even support ABS_MT_*, and it always outputs stuff when moving the touchpad mouse
[20:42] <cr3> cnd: I'm quite sure these laptops are not all multi touch though, so perhaps I'm not looking for the right thing in the output
[20:42] <cnd> hmmm… that page is out of date...
[20:42] <cr3> some laptops are from 2006, so not recent
[20:42] <cnd> xinput --list is the best way
[20:42] <cnd> you're right that it will always output stuff
[20:43] <cnd> you have to check that mtdev outputs ABS_MT_*
[20:43] <cnd> but that's easier said than done
[20:48] <cr3> cnd: so, if I understand correctly, when mtdev outputs ABS_MT_* that means that the touchpad can report two or more absolute coordinates in order to be considered multitouch, right?
[20:48] <cnd> correct
[20:49] <cr3> cnd: is there a possibility that mtdev outputs ABS_MT_* but physically pressing two fingers on the touchpad doesn't work? in other words, do you think that physical validation of multitouch support will uncover potential problems?
[20:49] <cnd> not likely
[20:49] <cnd> devices need linux drivers to have multitouch suppor
[20:50] <cnd> it doesn't magically occur :)
[20:50] <cnd> so anything that does say it is multitouch almost assuredly is
[20:51] <cr3> cnd: that's great because the test can be automated! however, do you know of a way to script such a test considering that ctrl-alt-f1 to vt in order to run mtdev-test is going to be really difficult to script :)
[20:51] <cnd> well, mtdev-test isn't the best way to go at this point
[20:51] <cr3> cnd: I mean, is there a way to determine whether a touchpad outputs ABS_MT_* other than running mtdev-test in a vt :)
[20:51] <cnd> it would be better to use the output of xinput --list --long
[20:52] <cnd> and look for XITouchClass
[20:52] <cr3> cnd: ok, let me compare a couple systems, one moment..
[20:52] <cnd> k
[20:57] <cr3> cnd: seems to be quite apparent: Class originated from: 0. Type: XITouchClass
[20:59] <cr3> cnd: could multitouch screens be tested the same way? ie, will they also appear with XITouchClass for a separate input device in the output of xinput --list --long?
[21:02] <cnd> yes, they will have XItouchClass too
[21:05] <cr3> cnd: interesting, a colleague just sent me the output of their xinput --list --long from a laptop that apparently has a multitouch touchscreen: http://pastebin.ubuntu.com/1187841/
[21:05] <cr3> cnd: from what I can see, it doesn't look like the touchscreen supports multitouch though, right?
[21:06] <cnd> so serial wacom tablets are a bit of a sore point
[21:06] <cnd> they can be multitouch
[21:06] <cnd> but you have to force them into a specific mode using a user-space tool
[21:06] <cnd> I don't really know how to do it myself
[21:06] <cnd> but I know bdmurray has been able to get it to work
[21:07] <cnd> but in its default startup state it behaves like a single touch touchscreen
[21:07] <cr3> cnd: hm, I remember seeing something about wacom serial on the HardwareSupport page under the MultiTouch wiki pages: " Need to use setserial to configure. "
[21:07] <cnd> yeah
[21:07] <cr3> cnd: ok, for enablement purposes, I still want to fail the device because user experience is bad and I want the vender to know about it :)
[21:08] <cnd> I would say so
[21:09] <cnd> if it doesn't boot up into multitouch mode, then it's a poor experience
[21:09] <cr3> cnd: apparently, multitouch worked out of the box in quantal. if you'd like, I could ask him to join this channel if you might have a few questions
[21:09] <cnd> I don't have any questions?
[21:09] <cnd> I personally don't care about wacom
[21:09] <cnd> well, old wacom serial I mean
[21:10] <cnd> they are only two touch
[21:10] <cnd> which doesn't get you much in ubuntu right now
[21:16] <cr3> cnd: if a toushcreen is not reporting XITouchClass, as in the case of wacom serial, but it does actually support two finger touch, how would you suggest I test for that?
[21:16] <cnd> well, that's a practically impossible task
[21:16] <cnd> you can't do anything if the kernel says it's not supported
[21:18] <cr3> cnd: if the kernel says it's not support, how can it work?
[21:18] <cr3> supported even?
[21:20] <cnd> it can't?
[21:20] <cnd> I don't understand what you're asking
[21:24] <cr3> cnd: in this output of xinput, the touchscreen apparently doesn't support multitouch (no XITouchClass) but my colleague says it supports multitouch: http://pastebin.ubuntu.com/1187841/
[21:24] <cnd> yes
[21:24] <cnd> there's no way for a script to know that
[21:27] <cr3> cnd: to recap, a touchpad or a touchscreen that reports XITouchClass in the output of xinput will necessarily work, but the opposite is not necessarily true where these devices don't report XITouchClass then they might work, right?
[21:27] <cnd> you could say that, but most likely devices that don't have a touch class just aren't multitouch devices
[21:27] <cnd> like your tablet
[21:27] <cnd> it just isn't a multitouch device
[21:28] <cnd> most multitouch devices these days work out of the box
[21:30] <cr3> cnd: if it's not a multitouch device but it behaves well enough according to criteria I'm not in control of, I might want to test physically
[21:31] <cr3> cnd: could you recommend a way for me to detect something like: when user presses two fingers on the touchscreen, the test passes?
[21:31] <cnd> I don't understand
[21:31] <cr3> cnd: for example, would it be possible to listen on the output of xinput test-xi2 for specific events?
[21:31] <cnd> if xinput says it's not multitouch, then it's not multitouch
[21:31] <cnd> it won't work
[21:32] <cnd> you can listen to the output of events from xinput test-xi2
[21:32] <cnd> it will tell you when there are touch events
[21:32] <cnd> but you won't get any touch events from non-multitouch devices
[21:34] <cr3> cnd: so how does the system simulate support for two finger touch for a wacom serial touchscreen?
[21:34] <cnd> it doesn't
[21:34] <cnd> there's no "simulation" of multitouch
[21:34] <cnd> if it's not multitouch, then it's not multitouch :)
[21:35] <cnd> or the driver doesn't support multitouch, which is basically the same
[21:35] <cnd> as far as any script can tell
[21:37] <cr3> so, it doesn't seem to be possible to determine whether a device that is physically a multitouch actually works as a multitouch device:
[21:37] <cr3> 1. if it's detected as a multitouch device, it'll just work
[21:37] <cr3> 2. if it's not detected as a multitouch device, then who's to know it's physically a multitouch device
[21:38] <cr3> cnd: does that sound about right?
[21:38] <cnd> exactly :)
[21:51] <cr3> cnd: if I wanted to make double sure that multitouch was working, like wearing a belt and suspenders, I would first check for XITouchClass but I would then want to verify physically...
[21:51] <cr3> cnd: if I listen to the output of events from xinput test-xi2, is there anything I should be paying attention to for multitouch events?
[21:56] <cnd> cr3: yes, look for XITouch{Begin,Update,End} events
[21:57] <cr3> cnd: sweet, I think that will make everyone happy :)
[21:57] <cnd> good :)