Blog

How to remove all YouTube Shorts from watch history

Sun, 05 May 2024 00:00:00 +0000

Unlike any other solution out there that simply hide the shorts, this one will actually remove shorts from watch history since beginning of time.

By the nature of ever changing YouTube, this solution is likely to break whenever YouTube changes. Check the comments in case someone already posted a fix.

WARNING: I do not take responsibility if something happens to your account, as Google may not like such activity. If you have a billion of shorts watched, removing all may take several hours. So it’s good to take a break once in a while. It is safe to close the page or the browser at any moment and continue later. Also, watch your RAM consumption, loading too much of the watch history may gradually eat all the RAM.

Go to youtube.com/feed/history, then open browser console (usually F12) and put:

(async function() {
  while (window.innerHeight + window.scrollY < document.documentElement.scrollHeight) {
    let days = Array.from(document.querySelectorAll('#contents > .ytd-reel-shelf-renderer'));
    for (let i = 0; i < days.length; ++i) {
      let day = days[i];
      day.scrollIntoView({behavior: 'smooth', block: 'start'});

      let rightArrowButton = day.querySelector('#right-arrow > ytd-button-renderer');
      while (rightArrowButton.checkVisibility()) {
        rightArrowButton.click();
        await new Promise(resolve => setTimeout(resolve, 100));
      }

      let videos = Array.from(day.querySelectorAll('ytd-reel-item-renderer'));
      for (let j = 0; j < videos.length; ++j) {
        console.log(`day ${i + 1}/${days.length} | video ${j + 1}/${videos.length}`);

        while (true) {
          var menuButton = videos[j].querySelector('#button[aria-label="Action menu"]');
          if (!menuButton) {
            console.log('waiting for menu...');
            await new Promise(resolve => setTimeout(resolve, 1000));
          } else {
            break;
          }
        }

        if (!menuButton.checkVisibility()) {
          continue;
        }
        menuButton.click();
        await new Promise(resolve => setTimeout(resolve, 100));

        let removeButton = Array.from(document.querySelectorAll('.style-scope.ytd-menu-service-item-renderer')).
          find(el => el.textContent === 'Remove from watch history');
        if (removeButton.checkVisibility()) {
          removeButton.click();
        }
        await new Promise(resolve => setTimeout(resolve, 100));
      }
      day.remove();
    }

    window.scrollTo(0, document.documentElement.scrollHeight);
    await new Promise(resolve => setTimeout(resolve, 5000));
  }
  console.log('done');
})();

This will go through every visible short on the page, click the three dots menu and then click “Remove from watch history”. After processing all visible shorts, it will scroll the page for more. Once it reaches the end of the history, it will just stop.

How to install Intune in Debian 12

Tue, 26 Mar 2024 00:00:00 +0000

See 2026 UPDATE.

I wish I didn’t have to use this abomination of a software from Microslop, but no luck. After spending countless hours (and even more in 2026) I had to document this, at least for my own sake.

Common misconceptions

There is no need to install openjdk-11-jdk (see 2026 UPDATE). It is not available in Debian 12 but can be installed from unstabe. Don’t do it, it’s not worth the risk of mixing stable with unstable. Instead you can use msopenjdk-11 package from Microslop’s repository (see below) and create a dummy openjdk-11-jre to satisfy the dependency.
There is no need to install Edge browser, unless you need it for some other reason. Intune does not require it.
There is no need to fake /etc/os-release file during enrolment (see below), but it is still needed for compliance check.

General steps

Create and install a dummy openjdk-11-jre package (see 2026 UPDATE).
sudo apt install msopenjdk-11 (comes from Microslop’s repository, which you added at the previous step).

Symlink openjdk-11-jdk to msopenjdk-11:

sudo ln -s \
  /usr/lib/jvm/msopenjdk-11-amd64 \
  /usr/lib/jvm/java-11-openjdk-amd64

sudo update-alternatives --config java and select the one that points to msopenjdk-11.
Follow Ubuntu 22.04 instructions to install Intune.
Finally run intune-portal.

When running Intune for the first time and clicking “Sign up” it will open a separate second window to enter your corporate email. If this new window is blank, see troubleshooting section below.
After entering your email, it should open yet another third window (yes, 3 windows total!) to enter your corporate password. If it did not open another window, see troubleshooting section below.
After authenticating, you should see the last screen with the “Register” button. If you instead see “Get the app” button, see troubleshooting section below.

Troubleshooting

When clicking “Sign up”, the new window is blank. The content is there but for some reason it fails to render. One workaround is to blindly navigate to the input field. Click the window in the top left corner and press Tab, it should jump to the input field, then type in your corporate email and press Enter. It should open a new window for entering password, which will have its content visible. Another workaround is to run a separate X server with software renderer:
```
sudo apt install xserver-xephyr
Xephyr -ac -br -noreset -screen 1280x1024 :1 &
# optionaly, run some window manager to be able to change window geometry:
# DISPLAY=:1 openbox &
DISPLAY=:1 intune
```
After entering email there is no separate third window to enter password, instead it asks for password in the same second window. Possible reasons:
- No symlink to msopenjdk-11.
- Intune fails to create a keyring to store the tokens. See Keyring Troubleshooting.
You authenticated but it shows “Get the app” button. After clicking the button nothing happens. See the previous step.
You get error “[1001]”. This can mean anything (classic Microslop) but most likely it one of the following:
- Too many attempts to enrol, the procedure might be temporarily blocked. Try again later.
- Still something is wrong with the keyring. See Keyring Troubleshooting.
If device compliance check asks you to to downgrade to a supported distribution, grab /etc/os-release from Ubuntu 22.04 installation (you may use this one) and keep it updated. Outdated version in os-release may fail the compliance check.

Keyring Troubleshooting

This chapter is also 2026 updated.

To be totally safe, we will create a new keyring setup. You will have to figure out yourself how to merge it later with the old one, or check if someone already commented on it.

Install seahorse (unless already installed):
```
sudo apt install seahorse
```
Move your old keyring setup somewhere else, e.g. rename it to keyrings~:
```
mv -v ~/.local/share/keyrings{,~}
```
Restart gnome-keyring-daemon (this will also recreate the keyring directory in ~/.local/share/keyrings):
```
systemctl --user restart gnome-keyring-daemon.service
```
Re-login, to avoid any running application to try access a not-yet-ready keyring.
Start seahorse. In the UI click the plus button “+”, then select “Password keyring”, name it “Default”, optionally set the password (previously blank password was fine but I’m 100% sure anymore, check if someone already commented on it), finally click “Create”. Right click the “Default” keyring in the side bar and select “Set as default”.
Now start intune-portal and try to enroll again.

2026 UPDATE

Pretty much the old guide still applies with one correction: intune-portal now has a hard dependency on openjdk-11-jre package. Workaround is to create a dummy package:

Install equivs:
```
sudo apt install equivs
```

Create package description template:

mkdir dummy-openjdk-11-jre
cd dummy-openjdk-11-jre
equivs-control openjdk-11-jre

Edit openjdk-11-jre file and replace all the content with following:

Package: openjdk-11-jre
Version: 11.999
Provides: openjdk-11-jre

Create a .deb package:
```
equivs-build openjdk-11-jre
```

Install the newly created package:

sudo dpkg -i ./openjdk-11-jre_11.999_all.deb

Now you can install intune-portal normally.

Merge Git repositories, preserving chronological order of commits

Wed, 22 May 2019 00:00:00 +0000

Sometimes you want to split a huge monorepo into multiple smaller repositories, like if you were migrating from Subversion. What if you want to do the opposite? I did.

I had a set of tiny git repositories that describe how to build Debian packages in OpenSuse Build Service. The repositories had similar directory structure and even the commits in each repository were alike. Eventually I thought that merging all repositories into a single monorepo would be more efficient.

I started searching if somebody had already achieved something like this. I found few attempts but they all differ in a way how the history is preserved.

Basically all the solutions I found merge repositories with --allow-unrelated-histories option. It allows merging branches that don’t share common ancestor commits. But in result the history of each repository is trapped inside merge commits. Here is how it would look for two repositories:

     A  B  C  D       W  X  Y  Z
     •--•--•--•   +   •--•--•--•

                  |
                  v
     A  B  C  D       W  X  Y  Z
     •--•--•--•   +   •--•--•--•
    /                /
•--•-------------•--•-------------•

Which is inconvenient if you still want to reorder or squash some consequent commits across multiple repositories. This is how I wanted the commits, ordered chronologically:

A  W  B  C  X  Y  Z  D
•--•--•--•--•--•--•--•

Here is the approach I went with:

First we clone all repositories somewhere locally.

mkdir -p ~/merge-attempt/repos
cd ~/merge-attempt/repos
git clone git@githost:user/repo1
...
git clone git@githost:user/repo9

Then we do git format-patch inside each repository to extract entire history into a set of patch files. These files will be named like 0000-commit-message.patch, 0001-another-commit.patch and so on. That makes them appear in chronological order, but that not very useful if you try to iterate over the patches from all repositories at once. The obvious way would be to rename the files to include commit date and time in their filenames.

cd repos
for R in *; do
  pushd $R
  git format-patch --root --src-prefix=a/$R/ --dst-prefix=b/$R/
  for P in *.patch; do
    grep '^Date: ' $P | \
      sed 's/^Date: \(.*\)/\1/' | \
      date -Iseconds -f - | \
      sed 's/[:+]/-/g' | \
      xargs -I{} mv $P {}.patch
  done
  mv *.patch ../..
  popd
done

Here --src-prefix and --dst-prefix arguments allow us to preconfigure the patches, so when applied, the files would be created in subdirectories named according to original repository.

We should get something like this:

2018-09-23T23-42-02-03-00.patch
2018-12-10T13-51-56-02-00.patch
2019-04-16T20-05-20-03-00.patch
2019-04-16T20-25-31-03-00.patch
2019-04-16T20-35-56-03-00.patch
2019-04-16T20-38-16-03-00.patch
2019-04-16T20-49-06-03-00.patch

Having all the patches ready, we need to create a new repository and apply them.

mkdir -p ~/merge-attempt/monorepo
cd ~/merge-attempt/monorepo
git init .
git am ../*.patch
git filter-branch --env-filter 'export GIT_COMMITTER_DATE="$GIT_AUTHOR_DATE"'

And that would be it.

How to disable Power Button in Openbox and Debian 8 Jessie

Sun, 09 Oct 2016 00:00:00 +0000

This is something that was bothering me for quite a while. Rarely but I happened to press accidentally the power button and shut down my computer, loosing all unsaved work. These few times I tried to search for solution, it never worked. After another undesired power off I decided to sit down and fix it.

Why I never managed to get it working properly is due to the combination of systemd and Openbox. Both are capable to set what the special buttons like power or sleep do. So fixing it in the either one kept the button misbehaving. And the trick was to configure it in both places.

Let’s first disable the button in Openbox. Edit ~/.config/openbox/rc.xml and find the XF86PowerOff piece and comment it out by wrapping it between :

<!--
<keybind key="XF86PowerOff">
  <action name="Execute">
    <command>systemctl poweroff</command>
  </action>
</keybind>
-->

Reload the Openbox config file:

$ openbox --reconfigure

Time for the systemd part. Open /etc/systemd/logind.conf and find the line with HandlePowerKey in it. By default the value for the power button handler is poweroff. Remove the preceding # symbol and change the value to ignore:

HandlePowerKey=ignore

Lastly, restart logind:

$ sudo systemctl restart systemd-logind

And we’re done. Now the power button is harmless. Of course it will continue working to power on the computer or wake it up from the sleep.

P.S. In case something doesn’t work, check the logs ( -f for real time update):

$ sudo journalctl -f

How to remove all SoundCloud likes at once

Fri, 30 Sep 2016 00:00:00 +0000

How so there is no way to track which tracks I listened? There are various “SoundCloud History” browser extensions, but every other streaming service has that (I know only YouTube) out of the box.

UPDATE Nov 2016: It took only few years of whining to get Listening history. Good job SoundCloud!

So my workaround is to use the Like button. I basically like every track in my stream I listened. And to keep it clean and tidy once in a while I remove all the likes to start fresh. Yet even removing all the likes at once is not possible. I just love the tech support answer:

You can’t do that. Hope that helps.

Helps a lot!

Go to soundcloud.com/you/likes, then open browser console (usually F12) and put:

(async function() {
    while (true) {
        var a = Array.from(document.querySelectorAll(".sc-button-like[title='Unlike']"));
        if (a.length == 0)
            break;
        a.forEach(button => button.click());
        // wait 2 seconds before more tracks get loaded:
        await new Promise(resolve => setTimeout(resolve, 2000));
    }
})();

What it does is that it goes through every Unlike button and clicks it. It may take a while if you have many likes.

WARNING: Snap, SoundCloud decided to disable my Liking abilities for who knows how long:

We’ve noticed a high volume of liking coming from your account and we would like to ask that you slow down. If your account is performing many more actions when compared to most others, it loses the human touch. Remember that there is a fine line between promoting yourself and bombarding other users with notifications. We want our community to remain a genuine, positive place for members to interact, so remember to stick to our Community Guidelines. You can continue liking again in 5 hours. The more you hit these limits, the longer the block will last.

How to pair a Low Energy (LE) Bluetooth device in dual boot with Windows & Linux

Thu, 18 Sep 2014 00:00:00 +0000

Those who dualboot know the pain of re-pairing your keyboard or mouse everytime you boot to a different OS. In this tutorial I will show how to pair a LE Bluetooth mouse in both Windows 8 and Debian simultaneously.

First pair the device in Debian, then reboot in Windows and pair the device there too. Yes, this will reset the paring in Debian, just carry on. Now we need to access the pairing keys in Windows. Download psexec.exe and open a command prompt with Administrator rights.

> cd Downloads
> psexec.exe -s -i regedit /e C:\BTKeys.reg HKEY_LOCAL_MACHINE\SYSTEM\ControlSet001\Services\BTHPORT\Parameters\Keys

The keys should be now exported to C:\BTKeys.reg. You should get something like:

Windows Registry Editor Version 5.00

[HKEY_LOCAL_MACHINE\SYSTEM\ControlSet001\Services\BTHPORT\Parameters\Keys\7512a3185b2c\84abd4a25ee1]
"LTK"=hex:6c,54,ee,80,40,47,6c,cb,fc,8e,f3,f1,c6,b2,04,9e
"KeyLength"=dword:00000000
"ERand"=hex(b):1e,12,aa,37,39,cc,af,34
"EDIV"=dword:00003549
"CSRK"=hex:38,d7,aa,c1,42,06,31,25,12,b8,5a,6d,c3,90,98,f2

7512a3185b2c is the MAC address of the Bluetooth adapter, which can also be written in standard format as 75:12:A3:18:5B:2C. 84abd4a25ee1 is the mouse address that was assigned during the pairing. We will need those numbers in the next steps.

Now boot again to Debian. The mouse wont paired automatically, because it is now assigned to a different address and with different keys. Let’s fix it.

$ cd /var/lib/bluetooth/75:12:A3:18:5B:2C/
$ ls
cache 84:AB:D4:A2:5F:E1 settings

If you look closely, the mouse address is not the same as in Windows. In my case only the fifth group is different. We need the device addresses to match, so rename the file.

$ mv 84:AB:D4:A2:{5F,5E}:E1
$ cd 84:AB:D4:A2:5E:E1/
$ ls
attributes gatt info

Now open the info file for editing and update the keys values. The relation between Windows and Bluez keys format is as follows:

LTK goes into Key of the LongTermKey group. Change to uppercase and remove the commas.
KeyLength goes into EncSize. In my case I had to replace the original value 12 with 0.
ERand goes into Rand. This is the tricky part. First you need to write down the ERand value in reverse, so it becomes 34afcc3937aa121e. Then convert it to decimal to get 3796477557015712286.
EDIV goes to EDiv. Convert from HEX to decimal normally, no reverse this time.
CSRK goes to Key of the LocalSignatureKey group. Uppercased and no commas.

At the end you should get this.

[LocalSignatureKey]
Key=38D7AAC14206312512B85A6DC39098F2

[LongTermKey]
Key=6C54EE8040476CCBFC8EF3F1C6B2049E
Authenticated=0
EncSize=0
EDiv=13641
Rand=3796477557015712286

Save the changes and reboot. From now on Debian and Windows will connect the mouse automatically. Awesome!

P.S. Mygod was nice to turn the instructions into a Python script, try it out.

Expanding wildcards with qmake

Tue, 26 Mar 2013 00:00:00 +0000

I was always wondering why doing something like this never worked

SOURCES += *.cpp
SOURCES -= excluded.cpp

excluded.cpp always stayed in SOURCES. The thing is that when you add a new value to the list, it’s just a string. Thus using wildcard in files names is nothing more but an asteriks character. And therefore using -= operator will only remove exact strings. But today I found another undocumented function I’ve been looking for years!

$$files(glob) — Returns a list of files which match the specified glob pattern.

So now I can do

SOURCES += $$files(*.cpp)
SOURCES -= excluded.cpp

and it actually works.

Shortcut to show folders first in Nautilus

Sat, 24 Nov 2012 00:00:00 +0000

I really miss a menu shortcut in Nautilus to toggle “Show folders before files” option. Going to preferences every time is too much hassle.

But there is other way to toggle it using dconf or gconftool. For Ubuntu 12.04 or newer I suggest using dconf, install it if you don’t have it yet.

Create a script with the following content:

#!/bin/bash

# Version: Sun Sep 8 2013

OLD_VALUE=
LSB_RELEASE=$(lsb_release -sr)

if [ $LSB_RELEASE \< '12.04' ]; then
    OLD_VALUE=$(gconftool --get /apps/nautilus/preferences/sort_directories_first)
else
    OLD_VALUE=$(dconf read /org/gnome/nautilus/preferences/sort-directories-first)
fi

NEW_VALUE='false'
if [ $OLD_VALUE == 'false' ]; then
    NEW_VALUE='true'
fi

if [ $LSB_RELEASE \< '12.04' ]; then
    gconftool --type Boolean --set /apps/nautilus/preferences/sort_directories_first $NEW_VALUE
else
    dconf write /org/gnome/nautilus/preferences/sort-directories-first $NEW_VALUE
fi

~~If you have older Ubuntu, dconf maybe not available. Uncomment gconftool lines and comment dconf ones.~~

UPDATE 2013: the new version of the script works across all versions of Ubuntu.

Save the script somewhere, I prefer to keep mine in ~/.bin. Also give it execute permissions.

The next step may vary depending on Gnome version, I will describe it as it’s in Ubuntu 12.04. Go to keyboard shortcut setting and add a new custom shortcut. The command field set to path to the script we just created, in my case it’s ~/.bin/nautilus-folders-first.sh. You may also try to set full path if relative doesn’t work. Right click on the shortcut and set an actual hotkey combination. Choose something that doesn’t interferer with any other application, the shortcut is going to be system global. Ctrl+Alt+[some key] is good choice, I prefer Ctrl+Alt+F.

Now open Nautilus and browse where you have both files and folders, test the shortcut. If it doesn’t work, try the script in terminal and see if there are any errors.

Using Focusrite Scarlett 18i6 in Linux

Sat, 08 Sep 2012 00:00:00 +0000

With the current 1.0.25 ALSA release you need to do some hacking to get Scarlett working. But the hack is as simple as it can be.

Download the tarball ftp://ftp.alsa-project.org/pub/driver/alsa-driver-1.0.25.tar.bz2 and comment the call to snd_usb_create_mixer() on the line 517 in alsa-kernel/usb/card.c so it would look like:

if (err > 0) {
    /* create normal USB audio interfaces */
    if (snd_usb_create_streams(chip, ifnum) < 0 /*||
        snd_usb_create_mixer(chip, ifnum, ignore_ctl_error) < 0*/) {
        goto __error;
    }
}

Build and install:

$ ./configue
$ make
$ sudo make install

Considering a bad practice to use sudo make install, in our case you don’t need to worry. Just reboot and select Scarlett as a default device in Sound Settings. Voila!

To restore Scarlett after reconnecting it from the USB port use:

$ sudo alsa force-reload
$ alsactl restore 0

Bash auto-completion trick

Tue, 31 Jul 2012 00:00:00 +0000

zhs has a nice feature. If a program supports --help then zsh can do auto-completion of its arguments. The trick is that zsh silently runs the program with --help argument, parses the output and use it for auto-completion. I wish bash could do that. Well, not exactly but it can. Bash has auto-completions too using templates. But if you want to add auto-completion to your program you would need to write a new template and either put it to /etc/bash_completion.d/ or source in .bashrc or .profile.

But you don’t have to do any of it if you can re-use bash templates as is. For example the template for configure scripts does almost the same what zsh does for all programs. It simply parses the output from ./configure --help. Unfortunately it works only for arguments that prefixed with double dash. Better than nothing anyways.