Posts by scaesare

Yeah unfortunately, it seems rather random, and takes weeks to appear...

Given it's been discussed here more than once thought I'd see if it was a known issue with either a generally accepted fix/workraound or something on the docket for a future build.

I'll see if I can figure out a way to force the failure so I can capture it...

Thanks.

So, there are a few references I've found related to "text disappearing", however any of the recent (< 2 yrs old) ones are buried inside threads with other main titles and/or are themselves several months old.

I'm getting similar behavior as the other reports: after some time, all text from the GUI will stop displaying. Buttons shapes, icons, etc... will all still be there, just the text on each is gone. This seems to happen after the unit comes out of suspend for me, but I can't confirm that's the only time it happens. I may have had the first instance be after I stooped a video playback. A reboot fixes it.

I've seen reports for a few a couple of different KODI versions on several Intel NUC users and at least one Chromebox. I'm running LE 9.6.2 on an ASUS Chromebox CN60 (Haswell) connected via HDMI.

There seems to be a couple of trains of thought:

1) Some sort of custom EDID configuration... but that seems to have been problematic for a few folks, and it seems odd that only text (and not the rest of the display elements) would go away if the unit didn't like the EDID it got back from the display

2) Changing the default font for the UI to Ariel.

The latter seems like an easy test and/or fix, but thought I'd throw it out here for any other thoughts/ideas.

Thanks.

-Steve

No worries... just wanted to update for anybody looking at the same thing...

Thanks for the help.

Quote from heitbaum

Supposedly : Manage Chrome Browser updates (Linux) - Google Chrome Enterprise Help

To stop Chrome Browser auto-updating, take one of the following actions:

• Create an empty repository before installing Chrome Browser:
  $ sudo touch /etc/default/google-chrome
• Add the following line to /etc/default/google-chrome:
  repo_add_once=false

Does this require adding to the base image? Or will / does it use the addon home?

Well, I tried this (using overlayfs in order to create /etc/default, thanks heitbaum).

Upon re-launch, Chrome 9.x still wanted to auto-update.

Uninstalled the add-in and reinstalled with the repo file still there during installation, and no dice...

Whelp... to talk to myself while learning about LE...

It looks like we are running with / mounted as a RO squashfs with symlinks to /storage/config for configuration stuff

Looks like you can point systemd at /storage when starting a service, but I don't' see any way to create a symlink from /etc/defalt/google-chrome as that other thread suggests without having to extract and modify the squashfs...

So did some poking and sure enough the root filesystem is RO... so I'm not quite sure if the suggestion is implying remounting the filesystem as RW and making the change, or if there is a method of pointing Chrome to a configuration dir under /storage....

Will try and do some more searching....

OK cool, I did see that, but was under the impression that was speaking to LE10.

I've already installed the add-on for LE9, so I suppose could uninstalling and try creating the empty repo dir.

Being somewhat new to LE, it seemed that the config stuff was all being added under /storage/ so wasn't sure messing directly futzing in /etc was recommended...

Thanks for the pointer.

Running 9.2.6 on x86_64 with the Chrome add on. Works great (once I fixed the video tearing issue).

One issue is that a few minutes after every launch Chrome pops up a dialog box complaining that it can't update.

I've not found anyway to disable Chrome from trying to update in it's settings... any pointers on doing so with this add-on?

Thanks.

-sc

So, I finally got all of this working, but seems like I had to jump through several hoops for what I assume would be somewhat default behavior everyone should want... so I'm posting here in hopes that it helps others (the other posts I could find on this are rather old), as well as to see if I just missed something obvious.

Firstly, just plugging the Flirc in to my Chromebox running LibreElec, and setting it up as a Flirc/Kodi device in MyHarmony initially worked pretty seamlessly for control functions, however there were two issues: 1) Any IR command would resume the Chromebox if suspended, and 2) the PowerOff functionality brought up the Kodi power menu, wherein you had to navigate to a selection.

When I installed the Flirc GUI on a Windows box to program it, it upgraded the firmware on my Flirc, after which it no longer responded to any generic IR event it saw to wake up the Chromebox. And interestingly the ability to toggle that in the Flirc Advanced menu remained grayed out (not sure why).

However, there are "Power On" or "Resume" events in the Harmony Flirc/Kodi profile (Just PowerOff). So even though there's a Flirc "Wake" event (on the Kodi or Full keyboard controller profiles), I had no corresponding Harmony function/button to map to it. I eventually solved this by taking an old IR remote I had from a device I no longer had installed, and creating a new Harmony button I named "Resume" and learning that old remote IR command, and set up the Harmony Flirc/Kodi device with that same code to Power On. I then trained the Flirc by mapping that IR code to "Wake" in the Kodi controller. Once done, the Chromebox only wakes on that specific IR command, and nothing else.

In order to force the Chromebox to suspend, I originally tried to change the default behavior in the Keymap by creating my own /storage/.kody/userdata/keyampas/keymap.xml and setting:

<keymap>
 <global>
  <keyboard>
   <power>XBMC.Suspend()</power>
  </keyboard>
 </global>
</keymap>

That didn't seem to work. So I eventually followed the advice of this Flirc forum post. I took yet another old IR remote for a piece of hardware I no longer had installed, and trained the Flirc to map that remote's power IR command to CTRL-F12, and then used the following keymap.xml:

<keymap>
 <global>
  <keyboard>
   <f12 mod="ctrl">XBMC.Suspend()</f12>
  </keyboard>
 </global>
</keymap>

I then taught the Harmony that same IR command, calling it "Suspend", and added it as the power-off command for my Flirc/Kodi device.

So now I have Kodi suspending and the Chromebox resuming with discrete IR codes. However it seems that if I hadn't had a couple of old IR remotes to learn some IR commands from, I would have had no way to teach the Harmony and Flirc how to do this. It also seems like LibreElec/Kodi might want to to have some way to change the default PowerOff events to directly trigger a power state, without having to bring up the menu first.

Of course, I could have indeed missed something obvious in which case pointers certainly appreciated.

Ah yeah, sorry... yeah it's x86_64 on an Asus Chromebox.

Touchpad seems to be detected as:

Kodi-FR:~ # dmesg | grep Mouse
[    2.762285] input: Microsoft Microsoft® Nano Transceiver v2.0 Mouse as /devices/pci0000:00/0000:00:14.0/usb1/1-6/1-6:1.1/0003:045E:0800.0002/input/input2
[    2.762478] hid-generic 0003:045E:0800.0002: input,hidraw1: USB HID v1.11 Mouse [Microsoft Microsoft® Nano Transceiver v2.0] on usb-0000:00:14.0-6/input1

I've tried to do my homework and see if I could figure our how to disable tap-to-click on the touchpad of my Microsoft wireless keyboard.

The info I've come across entails pointing systemd at /storage/.config/xorg.conf.d/ and adding a 10-synaptics.conf file consisting of:

Section "InputClass"
    Identifier "touchpad catchall"
    Driver "synaptics"
    MatchIsTouchpad "on"
    MatchDevicePath "/dev/input/event*"
    Option "MaxTapTime"             "0"
EndSection

I've also seen suggestion of instead a libinput.conf file similar to:

Section "InputClass"
Identifier "touchpad"
Driver "libinput"
  MatchIsTouchpad "on"
  Option "Tapping" "off"
  Option "NaturalScrolling" "on"
  Option "ClickMethod" "clickfinger"
EndSection

Neither seem to work. Xorg seems to be reading that conf directory successfully as it's picking up changes I made for the video driver (to eliminate Intel HW tearing).

I'm trying to understand if any vendor model touchpad would be covered by the latter? Attempts to use "xinput" to determine what hardware X is using doesn't work, as "xinput" doesn't seem to be installed.

Any pointers appreciated... thanks.