Kobo Start Menu

[Frenzie]

Originally Posted by tshering

How did you call env without first doing some changes via KoboRoot.tgz?

That I did, of course. I used the telnet method described on the wiki. I didn't apply any other patches or edit any other files.

The env variables don't seem to be very helpful though. For example, /dev/ts0 doesn't exist, while cat /dev/input/event1 shows that's obviously the right device. I suppose that's the same on all the other Kobos as well?

Originally Posted by tshering

From the reports I take it that there is no response to touch input at all. Is this correct? Or do at least the menu entries get (sometimes) highlighted?

I am asking because, this case reminds me somehow on a problem we had earlier, cf. the posts starting with #79.

I tried pressing just about every single spot on the screen and best I can tell nothing happened.

[jlynton]

Originally Posted by tshering

From the reports I take it that there is no response to touch input at all. Is this correct? Or do at least the menu entries get (sometimes) highlighted?

In my case there is no response at all, the menu entries do not get highlighted.

[Frenzie]

I looked into the test stuff described in #79. Here's what little I got out of it, from pressing all over the screen:

# cat seltest01.txt 

mode "1440x1080-0"
        # D: 0.010 MHz, H: 0.006 kHz, V: 0.005 Hz
        geometry 1440 1080 1440 2304 16
        timings 96000000 24 267 4 5 24 2
        accel false
        rgba 5/11,6/5,5/0,0/0
endmode

20141009_163415 before

20141009_163415 after

mode "1440x1080-0"
        # D: 0.010 MHz, H: 0.006 kHz, V: 0.005 Hz
        geometry 1440 1080 1440 2304 16
        timings 96000000 24 267 4 5 24 2
        accel false
        rgba 5/11,6/5,5/0,0/0
endmode

20141009_163417 before

[tshering]

Originally Posted by Frenzie

I looked into the test stuff described in #79. Here's what little I got out of it, from pressing all over the screen:

Thank you for testing. All I can see is that not a single touch input was recognized by KSM.
Frankly, I do not know how we can make KSM work for H2O. I guess, one solution could be that sergeyvl12 provides a new touch driver. However, I have no idea whether he is working on this.

[Frenzie]

I see his sources are available. Could you perhaps post your /proc/bus/input/devices and evtest /dev/input/event1 to see if anything's different?

#  cat /proc/bus/input/devices 
I: Bus=0019 Vendor=0001 Product=0001 Version=0100
N: Name="mxckpd"
P: Phys=gpio-keys/input0
S: Sysfs=/devices/platform/mxckpd/input/input0
U: Uniq=
H: Handlers=kbd event0 
B: EV=3
B: KEY=100000 4000000 8000008 0

I: Bus=0018 Vendor=0000 Product=0000 Version=0000
N: Name="zForce-ir-touch"
P: Phys=
S: Sysfs=/devices/virtual/input/input1
U: Uniq=
H: Handlers=event1 
B: EV=b
B: KEY=400 180000 4 0 0 0 0 0 0 0 0
B: ABS=2650000 1030003

# evtest /dev/input/event1
Input driver version is 1.0.0
Input device ID: bus 0x18 vendor 0x0 product 0x0 version 0x0
Input device name: "zForce-ir-touch"
Supported events:
  Event type 0 (Sync)
  Event type 1 (Key)
    Event code 258 (Btn2)
    Event code 307 (BtnX)
    Event code 308 (BtnY)
    Event code 330 (Touch)
  Event type 3 (Absolute)
    Event code 0 (X)
      Value      0
      Min        0
      Max     1440
    Event code 1 (Y)
      Value      0
      Min        0
      Max     1080
    Event code 16 (Hat0X)
      Value      0
      Min        0
      Max     1440
    Event code 17 (Hat0Y)
      Value      0
      Min        0
      Max     1080
    Event code 24 (Pressure)
      Value      0
      Min        0
      Max     2048
    Event code 48 (?)
      Value      0
      Min        0
      Max        2
    Event code 50 (?)
      Value      0
      Min        0
      Max       15
    Event code 53 (?)
      Value      0
      Min        0
      Max     1440
    Event code 54 (?)
      Value      0
      Min        0
      Max     1080
    Event code 57 (?)
      Value      0
      Min        1
      Max        2
Testing ... (interrupt to exit)

And a few lines of the output:

Event: time 1413023927.391625 ------- Config Sync ------ x: 36808 y: 36728 p: 0 ------------
Event: time 1413023927.391634 ------- Report Sync ------ x: 36808 y: 36728 p: 0 ------------
Event: time 1413023927.409451 ------- Config Sync ------ x: 36808 y: 36728 p: 0 ------------
Event: time 1413023927.409459 ------- Report Sync ------ x: 36808 y: 36728 p: 0 ------------

I got annoyed by the busybox limitations and got myself a Debian Jessie chroot environment. The output from real evtest is clearly a lot more useful:

# evtest  
No device specified, trying to scan all of /dev/input/event*
Available devices:
/dev/input/event0:      mxckpd
/dev/input/event1:      zForce-ir-touch
Select the device event number [0-1]: 1
Input driver version is 1.0.0
Input device ID: bus 0x18 vendor 0x0 product 0x0 version 0x0
Input device name: "zForce-ir-touch"
Supported events:
  Event type 0 (EV_SYN)
  Event type 1 (EV_KEY)
    Event code 258 (BTN_2)
    Event code 307 (BTN_NORTH)
    Event code 308 (BTN_WEST)
    Event code 330 (BTN_TOUCH)
  Event type 3 (EV_ABS)
    Event code 0 (ABS_X)
      Value      0
      Min        0
      Max     1440
    Event code 1 (ABS_Y)
      Value      0
      Min        0
      Max     1080
    Event code 16 (ABS_HAT0X)
      Value      0
      Min        0
      Max     1440
    Event code 17 (ABS_HAT0Y)
      Value      0
      Min        0
      Max     1080
    Event code 24 (ABS_PRESSURE)
      Value      0
      Min        0
      Max     2048
    Event code 48 (ABS_MT_TOUCH_MAJOR)
      Value      0
      Min        0
      Max        2
    Event code 50 (ABS_MT_WIDTH_MAJOR)
      Value      0
      Min        0
      Max       15
    Event code 53 (ABS_MT_POSITION_X)
      Value      0
      Min        0
    Event code 54 (ABS_MT_POSITION_Y)
      Value      0
      Min        0
      Max     1080
    Event code 57 (ABS_MT_TRACKING_ID)
      Value      0
      Min        1
      Max        2
Properties:
Testing ... (interrupt to exit)
Event: time 1413025320.241268, type 3 (EV_ABS), code 57 (ABS_MT_TRACKING_ID), value 1
Event: time 1413025320.241307, type 3 (EV_ABS), code 48 (ABS_MT_TOUCH_MAJOR), value 1
Event: time 1413025320.241315, type 3 (EV_ABS), code 50 (ABS_MT_WIDTH_MAJOR), value 1
Event: time 1413025320.241324, type 3 (EV_ABS), code 53 (ABS_MT_POSITION_X), value 799
Event: time 1413025320.241333, type 3 (EV_ABS), code 54 (ABS_MT_POSITION_Y), value 844
Event: time 1413025320.241345, ++++++++++++++ EV_REL ++++++++++++
Event: time 1413025320.241355, -------------- EV_SYN ------------
Event: time 1413025320.330832, type 3 (EV_ABS), code 57 (ABS_MT_TRACKING_ID), value 1
Event: time 1413025320.330870, type 3 (EV_ABS), code 48 (ABS_MT_TOUCH_MAJOR), value 0
Event: time 1413025320.330876, type 3 (EV_ABS), code 50 (ABS_MT_WIDTH_MAJOR), value 0
Event: time 1413025321.621320, type 3 (EV_ABS), code 54 (ABS_MT_POSITION_Y), value 489
Event: time 1413025321.621332, ++++++++++++++ EV_REL ++++++++++++
Event: time 1413025321.621342, -------------- EV_SYN ------------
Event: time 1413025321.720060, type 3 (EV_ABS), code 57 (ABS_MT_TRACKING_ID), value 1
Event: time 1413025321.720097, type 3 (EV_ABS), code 48 (ABS_MT_TOUCH_MAJOR), value 1
Event: time 1413025321.720105, type 3 (EV_ABS), code 50 (ABS_MT_WIDTH_MAJOR), value 1
Event: time 1413025321.720114, type 3 (EV_ABS), code 53 (ABS_MT_POSITION_X), value 805
Event: time 1413025321.720123, type 3 (EV_ABS), code 54 (ABS_MT_POSITION_Y), value 489
Event: time 1413025321.720135, ++++++++++++++ EV_REL ++++++++++++
Event: time 1413025321.720145, -------------- EV_SYN ------------

Apparently "clicking" the screen would have to be implemented through the ABS_MT_TOUCH_MAJOR value, which generates a 1 while touching and a 0 when not. The driver does not seem to output anything but type 3 events by itself.

So yeah, could anyone tell me if that's the same on other devices or if there's some kind of distinction here? I had expected the driver to simply send a button press when it might think it's appropriate, but then I don't really know anything about touch screens.

[jlynton]

Originally Posted by Frenzie

Apparently "clicking" the screen would have to be implemented through the ABS_MT_TOUCH_MAJOR value, which generates a 1 while touching and a 0 when not. The driver does not seem to output anything but type 3 events by itself.

So yeah, could anyone tell me if that's the same on other devices or if there's some kind of distinction here? I had expected the driver to simply send a button press when it might think it's appropriate, but then I don't really know anything about touch screens.

Frenzie, Sergey, of pbchess and coolreader fame, has started a new thread to address this issue: https://www.mobileread.com/forums/sho...d.php?t=247814

[Frenzie]

Thanks for mentioning it, jlynton. I can't wait to try some third-party reading software (and maybe some other neat things if I can figure out how to update the screen after changing the framebuffer). The display is great, and I knew exactly what I was in for when I got it, but it's just loaded with untapped potential. And frankly after using it for close to a week now I know more surely than ever that even if I'll end up using the device for nothing but reading, I just want good old-fashioned directories without any of this so-called processing which completely ignores my folder structure. I understand that both KOReader and Cool Reader provide just that. Of course I could always return the device, or within the next three weeks anyway, but I really don't want to go back to the '09 Bebook anymore after having used this screen. *grins*

[jlynton]

Originally Posted by Frenzie

Thanks for mentioning it, jlynton. I can't wait to try some third-party reading software (and maybe some other neat things if I can figure out how to update the screen after changing the framebuffer). The display is great, and I knew exactly what I was in for when I got it, but it's just loaded with untapped potential. And frankly after using it for close to a week now I know more surely than ever that even if I'll end up using the device for nothing but reading, I just want good old-fashioned directories without any of this so-called processing which completely ignores my folder structure. I understand that both KOReader and Cool Reader provide just that. Of course I could always return the device, or within the next three weeks anyway, but I really don't want to go back to the '09 Bebook anymore after having used this screen. *grins*

And thank you for all the work that you are putting into this. I also hate the kobo software, processing and database method. With start menu, coolreader for epub and koreader for pdf, and all your books in a hidden directory (I use .books), if you ever do boot into nickel it won't notice them. I use calibre, ignoring the kobo driver, and connect/share to the device, then save the books using {author_sort[0]}/{author_sort}/{series}/{series_index}/{authors} - {title} , this works well for me.

[WS64]

Originally Posted by jlynton

With start menu, coolreader for epub and koreader for pdf

Just curious, what is wrong with Koreader also for Epub? Especially when you use Calibre and this way can use the search function in Koreader...

[jlynton]

Originally Posted by WS64

Just curious, what is wrong with Koreader also for Epub? Especially when you use Calibre and this way can use the search function in Koreader...

I am in the habit of using coolreader, which (correct me if I am wrong) I think has better/easier adjustment of font sizes margins and so on. Don't get me wrong, koreader is great, just a question of personal preference.

[JimMcLaren]

Originally Posted by WS64

Just curious, what is wrong with Koreader also for Epub? Especially when you use Calibre and this way can use the search function in Koreader...

I use Koreader almost exclusively. BUT, Coolreader can search for text which Koreader can't. I like to be able to search for text so occasionally use Coolreader for that. If only it wasn't so slow to load the first time.

Not everyone uses Calibre. I don't care to search the metadata. I have my books organized so I can always find what I want. Granted, I only have 500 books on my reader, not like some whose library numbers in the thousands. I might change my mind then.

[Ken Maltby]

I have ~7000 books on my AuraHD and I still use a regular file/naming system. I did test the metadata search function, and it is really neat, but I don't use it.

Another new feature in Koreader, the OPDS support is very cool, as well. You can acquire books from OPDS catalogs using Koreader and then read them using Coolreader, if you wanted to.

Luck;
Ken

[tshering]

Edit: Please be aware that currently (latest version is 04 plus updates) there is no package for the H20. If you want to use KSM for H2O, you have to prepare the package yourself and use the touch driver referred to in this post.
Note: You have to do the changes before you install KSM.
Do this only if you know what you are doing. A minor mistake might brick your device!
---------------------

Thanks to the work of Cecill3000, KSM is working on H2O, but

Originally Posted by jlynton

One small bug , When I boot into KSM with Cecell3000's driver on the H2O, start menu starts upside-down. When I tap on any of the menu entries (where they SHOULD be, not where they are when upside-down), the display flips and from then on stays the right way up (until the next boot).
This is only a minor inconvenience, everything otherwise appears to work perfectly.

Since Aura HD shows the screen upside down, when a Qt application is started the first time after booting, I made the KSM application run and exit the first time without any user interaction. (There should be a better way to achieve this, but I have no idea what exactly is changing in the evironment.)

The same fix can be done for the H2O. As shown by Frenzie in this post the code

if [ "$PRODUCT" = "dragon" ]; then
    sh $kbmenudir/kobomenu.sh -autoselectafter=0
fi

should be changed to

if [ "$PRODUCT" == "dahlia" ] || [ "$PRODUCT" == "dragon" ]; then
    sh $kbmenudir/kobomenu.sh -autoselectafter=0
fi

In older versions, this code is in etc\init.d\rcS, in the updated version in .kobo\kbmenu\onstart\ksmhome.sh.

Originally Posted by jlynton

I have just looked into it a bit further, if I exit nickel to KSM, it is upside-down again.
Also a short press on the power button flips it to the right way up.

I quess by "short press" you mean you restart the device. Right?

When returning from nickel to KSM, we have the same problem, but the same solution does not work. Or, it seems to work, but only once. From that point on the screen is shown upside down every second time a Qt application is run.

The solution in the most recent version (I missed it by mistake in version 4) was to reboot after exiting nickel by

if [ "$PRODUCT" == "dragon" ]; then
    reboot
fi

This should now be changed to

if [ "$PRODUCT" == "dahlia" ] || [ "$PRODUCT" == "dragon" ]; then
    reboot
fi

The file is .kobo\kbmenu\onstart\startreader.sh.

If we understood what exactly causes the turning of the screen, we could find a better solution.

[tshering]

Currently the situation for H2O owner who don't want to make changes to the KSM code by themselves is not convenient.
I should do the necessary changes to the current version of KSM (and include also the updates, cf. this link, esp. the spoiler) and upload the repacked packaged. But my current situation (far from home and a very poor internet connection) does not allow this.
The second point is that I have ready the next version of KSM that replaces the current version. I will upload it as soon as I am back home (maybe first or second week of November).
Thank you for interest and help.

[jlynton]

Originally Posted by tshering

Currently the situation for H2O owner who don't want to make changes to the KSM code by themselves is not convenient.

I have made the updates as per your previous post, they work perfectly. Thank you so much for your work, and I look forward to the next version!

Enjoy your travels!