Michael's questions

I'm trying to set some keyboard shortcuts to use single non-standard keys on the keyboard that I don't ever use. Specifically, on my keyboard I have three keys for audio control. I am able to go to Settings > Keyboard shortcuts > and press the corresponding key to set the shortcut, but when I try to use that key it doesn't work.

Keyboard shortcuts calls these keys "Audio previous", "Audio play", and "Audio next". But when I press any of these, even when no shortcut is set on them, I get a square overlay in the lower middle of my screen with a circle with a diagonal line through it. I think this means that some lower level process is intercepting this key like it would with other special keys like screen brightness or volume and is rejected it because there is no audio to deal with, but it is also not passing it on to the system so my shortcut key can use it.

How can I find and disable this so my shortcut will work?

Update: I'm using GNOME and stock Ubuntu 21.10

I've got a bunch of LibreOffice files and I'm trying to find the one that contains specific text. In GNOME 3.38.5, which comes with Ubuntu 20.04, I tried searching from the file manager making sure to to select "Full Text". But nothing comes up, even for text which I know is present. From this I deduce that either GNOME isn't capable of indexing or searching in various file types like some other systems are, or else this capability isn't enabled.

I have found various shell scripts which appear to unzip part of the ODT file and pipe it through grep, however this is somewhat unwieldy and none of the examples I found bothered to actually tell you which file contained the text, which for my purposes make it useless. Also, most of the suggestions I found were quite old leaving the hope that newer versions of Ubuntu have plugged this deficiency.

Is there some way I can enable searching through large numbers of LibreOffice files in 20.04 or newer?

I previously had 18.04 and sound was working. If I plugged in a USB-C headset I could hit the volume controls and it would display an overlay on the screen showing me the volume going up and down.

Now I recently installed 20.10 from scratch (mainly because 18.04 refused to upgrade for many reasons) and I just noticed that when I try to play sound I hear it from the speakers, even when the USB headset is plugged in. Additionally, I cannot turn the volume up or down.

If I plug in a headset to the 2.5mm headset plug, sound no longer comes out of the speakers, but it also doesn't come out of the headset.

When I go to Settings > Sound I am able to move the "System Volume" slider, but it does not affect the volume. Under "Output Device" there are no output devices shown as available.

How can I get sound working properly so I can control the volume through my Fn+F1/2/3 keys as well as working via the USB-C headset?

update: In a comment to another answer I found a suggestion to run pulseaudio -vvv. But what I discovered is that running pulseaudio at all from the command line fixes the issue. But this requires me to run it manually, ties up a terminal, and stops working if pulseaudio is killed. So I suspect that something should be starting/retaining pulseaudio and isn't.

I've got an extra drive that I created an ext4 partition when it was connected to one machine. I plugged it into another machine and now I no longer have access to my files from that machine because my UID on the second machine is not the same as on the first machine. I would like to be able to plug the drive into either machine and be able to access the same files as the same user (name).

Obviously I could change my UID on the second machine, but then I would have to recursively find all files owned by me and change the owner which could take a long time to complete. Previously the drive was NTFS and I reformatted, but at that time I mounted with an option specifying the UID that owned the files, but obviously ext4 has actual the permissions expected by Linux on a per file basis so this would mess things up if any other users had files on the device.

Is there some better way that I can proceed so I can access my files from both machines?

I'm trying to thumbnail an entire image directory consisting of upwards of 100,000 files, a thousand per sub-directory, to minimize delays displaying the thumbnails for each directory. (Making it worse, some take quite some time to compute due to interspersed mp4 files which have been extracted from "picture with motion capture" on my phone). Over time I have noticed that my thumbnails have been mysteriously disappearing which is really frustrating as it takes hours of going to each directory and waiting for all thumbnails to be generated before going to the next one, and so forth.

Recently I discovered that I need to tweak the cache settings to prevent this. I used gconf-editor and in /desktop/gnome/thumbnail_cache I set both maximum_age to -1 (never age out) and maximum size to -1 (no limit). After that I started opening each directory and waiting for it to thumbnail and then on to the next one. I was monitoring the process when suddenly even while thumbnailing was in progress over 15,000 thumbnails suddenly vanished!

~/.cache/thumbnails$ ls large/ | wc -l
23021
[about ten minutes later]
~/.cache/thumbnails$ ls large/ | wc -l
7801

Now not only has a bunch of time and flash erase cycles been wasted for nothing, but I can no longer trust my system as apparently at any time, even a few hours after creating the thumbnails it can arbitrary delete them for unknown reasons.

How can I ensure that my thumbnails are permanent and never erased like this?

I have read that dconf-editor is another way to edit these settings, but it is harder to use, offering no hierarchy view, and further there is nothing under /desktop/gnome except "crypto" so I have no idea what the path to thumbnail settings would be or it would even offer anything that gconf-editor does not.

I have recently had an issue with suspending my laptop (GNOME on 18.04 LTS) - when I wake from sleep, sometimes my login session gets killed and I have to log in all over again. But I have found that if I type systemctl suspend before shutting the lid, I do not have this issue at all.

But I am tired of having to type this command every time before I shut the lid, and sometimes I don't have time to unlock the computer and type it, I need to just shut the lid and go. So I was trying to change the behavior of shutting the lid to running this command instead of whatever it is doing now. But apparently the lid action is controlled by /etc/systemd/logind.conf in the line that reads:

HandleLidSwitch=suspend

This doesn't tell me how the system is actually being suspended so I can change it to systemctl suspend! In this file is says to reference the logind.conf man page which itself references the systemd-logind.service man page which doesn't seem to have any useful info. It also has this tidbit:

A different application may disable logind's handling of system power and sleep keys and the lid switch by taking a low-level inhibitor lock ("handle-power-key", "handle-suspend-key", "handle-hibernate-key", "handle-lid-switch"). This is most commonly used by graphical desktop environments to take over suspend and hibernation handling, and to use their own configuration mechanisms.

So it's possible that this is how the system is actually suspending things, which might be why the expected method isn't being used. But it's unclear where I would actually look (taking a lock sounds like something that might be done in source code hidden by compilation to a binary, for example).

How can I set the actual command to run when the lid is closed to systemctl suspend?

I've got some external and network drives and I have noticed that Nautilus keeps a single directory in ~/.cache/thumbnails which it uses to store thumbnails of these external drives right along with my local files. I would really rather have each drive thumbnailed separately, preferably in a folder on that drive itself. This would also keep the thumbnails in a central location, so that from a different machine the thumbnails would not need to be regenerated.

I found this similar question on preventing thumbnails from being created on specific mount points. The answer on that question indicates that is not possible to do, as of over 8 years ago. Right now I'm using a janky system where I manually move what the thumbnail folder points to (via a symlink) but this of course requires that I only try to access one file system at a time. Is it possible in 2020 to make Nautilus keep the thumbnails on the drive containing the actual files being thumbnailed?

A while back I had some issues trying to install "nvidia-driver-440" - after installing the driver, I could no longer log in due to some crash in the X server. As a result, I went with the "nvidia-driver-435" instead. However, today I learned that I must have the 440 driver in order for software that I want to use to work properly.

I want to compare the minor version of the driver that apt-get or ubuntu-drivers would install against the sub-version of the driver that is available if I run sudo add-apt-repository ppa:graphics-drivers/ppa. They all have the same package name, but since they come from different sources I am hoping that latter is a newer minor version that will actually work.

How can I determine the sub-revision of the driver without installing it? If I install the same one as last time my system will be inoperable until I purge nvidia and go back to 435.

So far I tried caching the ubuntu package (sudo apt-file update) and then listing the files in the package (apt-file list nvidia-driver-440). But the package doesn't appear to have any actual driver in it, it's just documentation files in /usr/share/doc/nvidia-driver-440/.

For what it's worth, I need version 440.26.

I'm able to capture from my webcam using ffmpeg, but it's real color. But my webcam is advertised as having IR capability. How do I access this?

When I run ffmpeg -f v4l2 -list_formats all -i /dev/video0 I get this output:

Compressed:       mjpeg :          Motion-JPEG : 1280x720 640x360 640x480 848x480 960x540
Raw       :     yuyv422 :           YUYV 4:2:2 : 160x120 320x180 320x240 424x240 640x360 640x480

The camera specs are given as video: 1280x720 at 30 fps, IR Video: 340x340 at 60 fps

Also, here is the output of v4l2-ctl --list-devices:

Integrated_Webcam_HD: Integrate (usb-0000:00:14.0-11):
    /dev/video0
    /dev/video1

I thought maybe the Raw stream would contain the Infrared video stream, but when I run this:

ffplay -input_format yuyv422 -video_size 320x240 -i /dev/video0

The output looks identical to the 1280x720 output except smaller.

I also ran ffmpeg -f v4l2 -list_formats all -i /dev/video1 and got this output:

Raw       : Unsupported : 00000032-0002-0010-8000-00aa003 : 640x360

But I don't seem to be able to play anything from this device, I get:

Cannot find a proper format for codec 'none' (id 0), pixel format 'none' (id -1)
Assertion *codec_id != AV_CODEC_ID_NONE failed at src/libavdevice/v4l2.c:811

How do I capture infrared video from the webcam?

I have looked at the default GNOME magnifier, and it either magnifies the whole screen, or half of it. The full screen magnification means I can see only a quarter of the screen at once, and the half screen magnification stops following my pointer and jumps away when stop moving the mouse, making it useless. Even if it did not do that it still hides too much of the screen. I just sometimes need a little help seeing text when it's in a hard to read color or a bit too small, without affecting any of the screen beyond a small around around the pointer.

I have seen magnus, but it appears to put the magnified part of the screen in a totally different part of the screen, meaning I have to look away from where I'm trying to concentrate, as well as having to move the magnification window whenever I need to see what's actually under it - well, at least if I want to see it unmagnified.

I'm picturing something that looks like an actual (or really virtual) rectangular magnifying lens floating over the screen centered on the pointer. Something like this except that that shows the mouse and lets you continue to work as if it's not there. How can I magnify the screen like this? Am I missing a knob to control the built-in GNOME magnifier? Is there a different magnifier I can use?

I'm running speedometer 2.8 which has been recommended in various answers on this site to monitor network utilization. I have read the manual page and have searched on Google but I am not seeing it documented what the numbers on each interface mean.

For instance, when I am receiving data, I might see something like this:

RX: wlp3s0  2.73 MiB/s 2.52 MiB/s  947 KiB/s

A screenshot from the documentation shows:

In addition, above these numbers to the right of the cyan ASCII graph are two vertical bars, one blue and one white which seem to be tracking two of these numbers, although I'm not sure which. This picture shows them circled in yellow.

I assume that maybe these three numbers represent the transfer rate during the last period as well as maybe average, minimum, or maximum values. But it is unclear as to which might be which.

What specifically does each number refer to, and to which of these numbers do the blue and white graphs correspond?

I have read that the XBox One formats external hard drives in NTFS format. My understanding from reading pages such as this one is that NTFS support should work "out of the box". However, when I plug the drive into my Linux machine, nothing gets mounted. In dmesg I see that the device was recognized, but fdisk claims there is no partition table. I have read about a Windows utility that is needed (on Windows, at least) to modify the MBR of the drive to switch it between "XBox Mode" and "PC Mode" so my suspicion is that a similar thing needs to be done in Linux - but I can't find any documentation on exactly what it is doing. Alternately, maybe NTFS simply isn't working out of the box - I don't see ntfs under /proc/filesystems, for instance, and a naive attempt to "modprobe ntfs-3g" failed to find anything.

I'm running (L)Ubuntu 18.04.3 LTS.

I'm using VLC ("VideoLAN") on Linux to play 4k video I have recorded. I'm trying to quickly review long segments so I want to play the video faster, at least 5x. But when I exceed 2.9x the video immediately hangs and I am unable to recover. If I reduce the speed, seek, anything I have tried does not fix the issue of video no longer playing. The only thing that fixes it is to exit the program and restart it. If I play slower, above 2.0x but under 2.9x, then it does the same thing after a few seconds. I have played around with the buffering settings to no effect. It doesn't matter how big the window is; I get the same behavior even if I shrink it down substantially.

I have tried Gnome MPV and it doesn't freeze up, but despite the playback speed it doesn't seem to actually be honoring the requested rate. For instance, when I have it set to playback at 5x, it seems to be closer to 2x.

The only way I have found to reliably watch video faster is to upload it to Google Photos, then wait for it to be available at 4k, then play it back. But this requires a lot of waiting (sometimes Google Photos doesn't make my original 4k video available for playing at 4k for several days) and it also requires a fast connection in order to avoid stalling. (My home connection is not fast enough). But when it works, it plays perfectly, smooth video at 5x.

I tried mplayer, but it seems to completely ignore whatever speed I tell it (either via the command line, e.g. "-speed 5 -af scaletempo" or via the keyboard shortcuts, and only ever plays at 1x speed!?!

How can I locally play my video at these playback speeds?

As a workaround I thought maybe I could use FFMPEG to re-encode it into a new video. Using this answer I came up with this command:

ffmpeg -i input.mp4  -filter:v "setpts=PTS/5" -an output.mp4

But this is only encoding about 2 frames per second and my first video alone is about 30 minutes which means I would have to wait 7.5 hours.

I have two batteries for my laptop of differing capacities. I took out the larger battery today and hot swapped it for the smaller battery as I had not used it in a while an I wanted to make sure it didn't discharge too far. (Note: Technically it is probably more accurate to call it a "warm" swap. I basically went to suspend, but mainly for convenience because I have to shut the lid to flip the unit over to swap batteries, and my system is set to suspend when the lid closes.)

But after swapping the battery, the Xfce Power Manager reports strange values. For instance, it said it would take about 60 hours to charge the battery. After a couple of hours it has charged up to 80% and is still saying close to 6 hours are left to charge. Under devices it says that fully charged the battery is 66.2Wh. But if I swap back to the larger battery the battery name (e.g. LG 45NXXXX) doesn't change, and the battery size doesn't change either. This is making me wonder if it thinks the smaller battery is still the larger battery and might try to overcharge it thinking it has the higher capacity.

Is this something I need to worry about? Is there some command I can run to "reset" things so it properly recognizes the new battery?

If I drag a file from one file manager window and drop it in another, the file gets moved from the source folder to the destination folder. If I hold down the control key while doing this, the file gets copied. If I hold down the Shift key while doing this, the file is also moved. Holding down the Alt key causes the window to be moved instead, so I'm ignoring that case.

I would like to change this behavior so that when the Shift key is held down when the file is dropped, a script of my choosing is running which gets the source path/file and destination path. Alternately, it would work if I could copy the path on the source side from the menu or with the keyboard shortcut key (e.g. Ctrl + C) and then have a menu option / shortcut key that performs the action in the target window.

My main (but not only) use for this will be to do things like create hard/soft links. Another thing I want to do is to create a script that, if the file already exists in the destination directory, checks if the files have identical contents, and if they do, it instead deletes the source file.

I'm using PCManFM on Lubuntu 18.04, but if another file manager is the only way to do it that's fine too.

I am trying to configure my laptop running LUbuntu 18.04 to first go to the login screen, and then after the login screen is displayed, the suspend. The reason for this is that the default of simply suspending and then displaying the login screen on resume results in a delay of about 5 seconds before the login screen is displayed upon resume, whereas if the login screen is already up when the machine is suspended, it is available almost instantly on resume.

In any case, I have been researching how to do this, and it seems like my best option is to simply run a script when the lid is closed which first calls dm-tool lock followed by a delay if needed to wait for the login screen to be display, and then systemctl suspend.

However, how to I link the action of the lid being closed to running this script? From my research it appears that acpid is the way to go. From reading the man page, it appears that there needs to be an executable script in /etc/acpi/events called lidbtn. The directory exists on my system but no such script exists, and the man page is unclear as to exactly what it needs to contain. Looking at the other scripts in this directory it appears that I need to know what to put for the "event=" line, but I haven't found this anywhere.

There is also mention of the window manager handling this event, and it is unclear what I need to do to make sure my script doesn't conflict with that. Under the Xfce Power Manager (Menu -> Preferences -> Power Manager) under the General tab I see "When laptop lid is closed"... Suspend, but the only other choices are "Switch off display" and "Lock screen". There is no choice for "do nothing". What do I need to do here? Does the requested action happen before or after whenever acpid script I might create get executed, or does that completely override this?

I cloned a drive, and now I am trying to change the UUID of the cloned drive. There are various answers on AskUbuntu as to how to do this. I was prompted to run a fresh e2fsck first, which I did. But when I actually ran tune2fs I got this message unexpectedly:

Setting UUID on a checksummed filesystem could take some time.
Proceed anyway (or wait 5 seconds to proceed) ? (y,N)

While I was trying to process what this meant to decide what to do, it went ahead and started. It took a few minutes to complete, but now I am concerned. What is the explanation of why it takes so long? Is this something I should be worried about?

I've got a Lenovo T450s that I'm running LUbuntu 18.04 on. I've got the Synaptics driver and am using Touchpad-Indicator for touchpad control.

I have noticed that the touchpad seems very jerky at times, and when I experimented with it I discovered it is only jerky when I am trying to make arcing or circular motions, for instance when trying to make the motion of signing my name.

In the image below, I watched the touchpad to make my finger move in as close to a circle as possible:

As you can see, instead of being a circle, the cursor wanted to move in straight lines and only change directions at certain intervals.

I tried again, this time to draw a quick circle:

Instead of a circle it is looking more and more like a square with rounded corners!

I tried disabling natural scrolling and circular scrolling under Touchpad Indicator -> Preferences -> Touchpad configuration, but it made no difference.

How can I make the touchpad behave properly? It's practically useless like this!

I'm trying to find the control panel that will let me configure various things about my touch pad such as disabling it when other devices are active and what not. I keep accidentally touching my touchpad while I am typing which causes the cursor to jump somewhere else and then text goes there until I realize what happens and have to try to undo it. I want to be able to have the touchpad ignored when I am using the trackpoint on my computer somehow. The only setting I can find at all related to my touchpad is Preferences -> Keyboard and Mouse, but it is very minimal and doesn't even touch on the touch pad.

In my research I found that there is a package called "gsynaptics" that I need to install after which I should run gpointing-device-settings. However, when I try to install it, it says the package is not found, and when I try to run gpointing-device-settings it doesn't suggest any repositories I could find it in. I have also noticed that most questions about this topic are several years old and in one place it suggested that gsynaptics is no longer used (but then they went on to indicate that they had, in fact, installed it, without saying how).

From what I have read, this package was in Universe, but under "Software and Updates" the "universe" setting is already checked.

Where can I find the package to let me configure my touch pad in LUbuntu 18.04 in this way?

I have tried

sudo touch /forcefsck

Then I reboot, and the file disappears, but no disk check appears to have taken place. The whole boot process takes under 10 seconds!

Then I tried this to force it to reboot every day:

sudo tune2fs -i 1 /dev/sda1

To verify, I ran this:

sudo dumpe2fs -h /dev/sda1 

The relevant lines of output are:

Last checked:             Sun Sep  4 12:15:45 2016
Check interval:           86400 (1 day)
Next check after:         Mon Sep  5 12:15:45 2016

Running "date" I see my clock says it is currently Dec 19, 2016, so fsck should run on the next boot up. But after I reboot it is again really fast and rerunning dumpe2fs I see the same exact text above, suggesting it never even tried to run fsck.

How can I force fsck to run at boot-up?