Tap is sometimes not registered on touchscreens

SRU: patch for maverick attached

False negatives are felt directly and affect the overall utouch experience. The fix is very simple, therefore I am nominating this for maverick inclusion.

The patch has been run locally for ten days, tested on both touchpads and touchscreens of varying frame rate. A package is available in the utouch unstable ppa (the version name there is 1.0.15+repack2-0ubuntu2, for historical reasons. It is, however, exactly the same package).

The patch affects the logic of single-finger pointer emulation, which is the subject of the test case. The regression potential is thus relatively low.

Doesn't quite work for me. Still need to drag a tiny amount instead of simply tapping.

How did you test this? Using the ppa:utouch-team/unstable ppa? Are you using a touchscreen? Which kind? Did you reboot?

Yes, added ppa:utouch-team/unstable ppa, apt-get update, apt-get upgrade & reboot

Using the WeTab touchscreen that you included in egalax-hid this morning

It's not bad if you know you have to drag the cursor a little bit, but for a user not aware of this it could be frustrating

I am asking because I have tested this very explicitly on a joojoo, with a similar DWAV device. Since the support for your device is brand new, it would be good to go through some other tests in that other bug first, before attributing your problem to this particular bug.

Point well taken. After all, we have only a couple of hours experience. If you need something tested pls. let me know.

If you are on IRC, you find us in #ubuntu-touch

debb1046, just to make sure - we have a second patch judged too complex for the SRU. If you can apply and test it, does it make any different for you?

debb1046,

If you are up for it, this issue seems to be a perfect example to start our user test database.

1. Compile evemu (bzr branch lp:utouch-evemu)

2. Find the event node of your device (grep -i galax /proc/bus/input/devices -A4 | grep H:)

3. Extract the device definition (sudo ./tools/evemu-describe /dev/input/eventX > wetab.prop)

4. Start a recording (sudo ./tools/evemu-record /dev/input/eventX > wetab.event)

5. Perform the test that fails - the recording will stop after ten seconds of inactivity.

6. Attach the wetab.prop plus wetab.event files here with a description of the failure.

With this data, it is possible to reproduce the problem exactly, and the test can be used during development.

Thanks!

I applied patch from #9 to utouch-grail_1.0.15-0ubuntu1. The problem persists and there is an additional problem where it thinks I have lifted my finger when dragging (I get interrupted lines in tuxpaint)

Attaching wetab.prop

Here is wetab.event for utouch-grail_1.0.15-0ubuntu1 with patch 0002. I am hitting keys on onboard.

Went back to the version 1.0.15+repack2-0ubuntu2~utouch1 from unstable ppa and recorded same activity

I also upgraded to the repack2 version.
I found no improvements in the pointer behaviour after a reboot.
I still have to move my finger sometimes.

I have the GeneralTouch Win7 TwoFinger device.

Is some form of calibration required?

Thank you very much for the evemu data. With it, I was able to reproduce the problem, and it seems the package in the unstable ppa is not applying the patch properly. Without the patch, replaying the sequence would only move the pointer a couple of times. With the patch applied, the pointer would move between 11 distinct positions. Those positions can also be seen using mtview.

Please find attached two debian packages. The first one, test1, is the package as it should have been in the unstable ppa, with the patch in question properly applied. The second package is with the additional patch which is applied for natty only (so far). Please test them - if everything is correct, you should find that both work.

Thanks!

Hi Henrik,

Thank you very much for your quick response.
I am running the 32bit version, is there a package for that too?
Then I can test this too.

The attached package is built yet another way, using the packaging.quilt branch, proposed to be merged into this SRU. The irony is that the quilt patch was included in the first version of the package, and later removed as someone assured it would work. It probably does in some settings, but apparently not in mine. :-)

So: expected result with this i386 package (test3) is the same as with test1.

Hi Henrik,

Whatever works, and this one definitely does :).
Thank you very much!
If you need any additional logs ore other info please let me know.

I could also install only the -test3, having 32 Bit version installed. I confirm that -test3 works very well. Hooray!

I am glad to hear that. :-)

So, to recap, the fix does solve the problem, but there was a problem with the package itself. The fix for it has been merged, and a new version has been uploaded to ppa:utouch-team/unstable. The uploaded package should be identical to test3 above, but please test. With confirmation, we can hopefully get the fix into -proposed this week.

Thanks!

I removed the manual install in Synaptic and let it do it's thing with the unstable ppa active.
I now have this version: 1.0.15+repack2-0ubuntu2~utouch2 of libutouch-grail1.

My pointer still behaves like it should.
Is that the test you were looking for?
Thank you, also on behalf of MacRules :)

All is well on WeTab, too.

So that's a wrap - time for a sponsored upload. :-)

Thanks for testing!

----- Original message -----
> So that's a wrap - time for a sponsored upload. :-)

Nicely done!

I haven't noticed any regressions with my configurations using the new package.
Tested with ntrig and magic trackpad.

Hey Henrik, just to confirm from lp:~utouch-team/utouch-grail/packaging branch, I get this:

dpkg-source: info: construction de utouch-grail à partir de ./utouch-grail_1.0.16.orig.tar.gz
patching file src/grail-api.c
Reversed (or previously applied) patch detected! Skipping patch.
1 out of 1 hunk ignored -- saving rejects to file src/grail-api.c.rej
dpkg-source: erreur: LC_ALL=C patch -t -F 0 -N -p1 -u -V never -g0 -E -b -B .pc/0001-Report-touch-down-on-one-finger-entry.patch/ < utouch-grail-1.0.16.orig.shaNdj/debian/patches/0001-Report-touch-down-on-one-finger-entry.patch a produit une erreur de sortie de type 1
dpkg-buildpackage: erreur: dpkg-source -i -I -b utouch-grail-1.0.16 a produit une erreur de sortie de type 2
debuild: fatal error at line 1337:
dpkg-buildpackage -rfakeroot -D -us -uc -i -I failed
bzr: ERROR: The build failed.

-> So debian/patches/0001-Report-touch-down-on-one-finger-entry.patch doesn't apply, I think you included it to 1.0.16 from the source. Once removed, it builds fine on both natty and maverick. Just a confirmation is needed before I sponsor to -proposed.

Hi Didier,

> Hey Henrik, just to confirm from lp:~utouch-team/utouch-grail/packaging

> branch, I get this:

This is the packaging branch we use for natty.

> dpkg-source: info: construction de utouch-grail à partir de ./utouch-grail_1.0.16.orig.tar.gz
> patching file src/grail-api.c
> Reversed (or previously applied) patch detected! Skipping patch.
> 1 out of 1 hunk ignored -- saving rejects to file src/grail-api.c.rej
> dpkg-source: erreur: LC_ALL=C patch -t -F 0 -N -p1 -u -V never -g0 -E -b -B .pc/0001-Report-touch-down-on-one-finger-entry.patch/ < utouch-grail-1.0.16.orig.shaNdj/debian/patches/0001-Report-touch-down-on-one-finger-entry.patch a produit une erreur de sortie de type 1
> dpkg-buildpackage: erreur: dpkg-source -i -I -b utouch-grail-1.0.16 a produit une erreur de sortie de type 2
> debuild: fatal error at line 1337:
> dpkg-buildpackage -rfakeroot -D -us -uc -i -I failed
> bzr: ERROR: The build failed.
>
> -> So debian/patches/0001-Report-touch-down-on-one-finger-entry.patch
> doesn't apply, I think you included it to 1.0.16 from the source. Once
> removed, it builds fine on both natty and maverick. Just a confirmation
> is needed before I sponsor to -proposed.

Yes, utouch-grail/packaging contains the change in source, whereas
utouch-grail/packaging.maverick contains the change as a patch. The patch and
special packaging.maverick branch was made to avoid modifying the main package
version in the SRU. In practice, both branches contain the same change. Please
pick the one you find most suitable, and we can adjust accordingly.

Thanks,
Henrik

Hey Henrik,

ok, but your branch contains the patches where it shouldn't: http://bazaar.launchpad.net/~utouch-team/utouch-grail/packaging/files/head:/debian/patches/

So, I'm removing it and push the removal as rev 65 :)

Thanks for clearing that up, I'm sponsoring now!

> ok, but your branch contains the patches where it shouldn't:

> http://bazaar.launchpad.net/~utouch-team/utouch-
> grail/packaging/files/head:/debian/patches/

Oh! Thanks for the fix, pulled and tested.

Please upload to natty ASAP, too. Thanks!

Accepted utouch-grail into maverick-proposed, the package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

Version 1.0.16-0ubuntu1 seems to behave as expected.

Works well on WeTab

utouch-grail | 1.0.16-0ubuntu1 | maverick-proposed | source

Oops, wrong version. What I meant to say is, please uplaod 1.0.16 to natty ASAP, so that this can progress to maverick-updates. Thanks!

Didier, may we ask for your kind sponsoring once again?

i was about to sponsor the upload, but the branches are a bit messed up now.
https://code.launchpad.net/~utouch-team/utouch-grail/packaging contains a tag and changelog entry for maverick-porposed and there seems to be no plain natty branch atm to pull from for a sponsored upload.
please provide a branch for natty specifically or revert/change the tag and changelog entry in the natty packaging branch.

Thanks for looking into this. A natty branch has been linked here, and a merge request into packaging is pending. Is the change satisfactory? (https://code.launchpad.net/~utouch-team/utouch-grail/packaging.natty)

This bug was fixed in the package utouch-grail - 1.0.16-0ubuntu2

---------------
utouch-grail (1.0.16-0ubuntu2) natty; urgency=low

  * rebuild the -updates maverick version with natty toolchain (LP: #667802)
 -- Didier Roche <email address hidden> Tue, 23 Nov 2010 13:04:23 +0100

This bug was fixed in the package utouch-grail - 1.0.16-0ubuntu1

---------------
utouch-grail (1.0.16-0ubuntu1) maverick-proposed; urgency=low

  * Report touch down on one-finger entry (LP: #667802)
  * New upstream release.
 -- Henrik Rydberg <email address hidden> Mon, 08 Nov 2010 21:02:29 +0100