Forem: hopsayer

Fix Time Jumps Between Linux and Windows (Dual-Boot)

hopsayer — Fri, 26 Dec 2025 20:34:03 +0000

Synchronizing Time Between Linux and Windows in Dual-Boot Setups

If you run both Linux and Windows on the same machine, you might notice that after rebooting, your system clock jumps by a few hours.

This happens because both systems read the same hardware clock but interpret it differently: Windows treats it as local time, while Linux treats it as UTC.

To fix this, both systems need to use the same time reference. There are two main approaches:

Use UTC for both systems (recommended):

On Linux:

timedatectl set-local-rtc 0 --adjust-system-clock

It's not necessary because Linux uses UTC by default, but this explicitly ensures the setting.

On Windows: Use this command:

reg add "HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\TimeZoneInformation" /v RealTimeIsUniversal /t reg_dword /d 1 /f

It will add a registry key to treat the hardware clock as UTC.
You can save it in a .txt file, rename it to switch-to-utc-time.bat and run as administrator. Alternatively, open a cmd or PowerShell window as administrator, paste the line, and press Enter.
Ater applying the registry setting, restart Windows.

Use local time for both systems (Windows default):

On Linux:

timedatectl set-local-rtc 1 --adjust-system-clock

Then reboot.

After this change, both OS clocks will stay in sync, preventing issues with logs, Git commits, scheduled tasks, or databases caused by time jumps.

What is hardware clock and where is it stored?

The hardware clock, sometimes called RTC or CMOS clock, is built into the motherboard and keeps time even when the computer is powered off.

You’ve probably remembered that small CMOS battery — the one that used to make your clock reset when it died, and maybe the one you’ve even replaced yourself.

Both Linux and Windows read this clock at startup.

Why time jumps feel asymmetric in dual-boot setups

You may notice that switching from one OS to another breaks the clock, but switching back does not always fix it. This asymmetry happens because each OS not only reads the hardware clock differently, but also writes to it on shutdown.

If one system writes UTC and the other later reads it as local time (or vice versa), the offset gets applied twice. Each reboot reinforces the mismatch until both systems agree on how the hardware clock should be interpreted.

The only real fix is consistency: both systems must treat the hardware clock the same way — ideally as UTC.

Why Windows uses local time by default

This behavior is mostly historical. Early PCs, MS-DOS, and early Windows versions assumed single-user, single-location machines. Storing local time in the hardware clock made the system immediately show the “correct” time for the user after boot.

Modern Windows inherited this design for backward compatibility. Linux, on the other hand, adopted UTC as a system default much earlier, especially for servers and networked environments.

UTC vs local time: standard vs convenient simplification

UTC is an international time standard based on atomic clocks. In practice, it acts as the “base unit” for time in computing: it does not change with daylight saving time and does not depend on geography.

Local time, as used by Windows by default, is not a different standard — it is simply UTC plus a timezone offset and daylight saving rules applied for user convenience. This makes it suitable for display, but fragile as a system-level reference.

What about Windows Server? How does it stay synchronized?

Windows Server behaves the same way as desktop Windows: by default, it treats the hardware clock as local time, not UTC. This can add friction in real-world environments, because most servers on the internet — Linux hosts, containers, cloud VMs, CI/CD systems — operate in UTC.

When Windows Server uses local time while the rest of the infrastructure uses UTC, administrators may face:

Misaligned timestamps across systems
Confusing log correlation during incident analysis
Scheduled jobs running at unexpected times
Extra mental overhead when debugging time-related issues

For this reason, Windows Server is often explicitly configured to use UTC by setting the same RealTimeIsUniversal registry key and syncing time via NTP.

Why this matters for users

If your clocks aren’t in sync, you might notice:

Files showing the wrong creation or modification time.
Scheduled tasks running too early or too late.
Confusing timestamps in logs.

Why this matters for developers

For developers, having consistent system time is critical. When dual-boot clocks drift, you may encounter:

Git commits appearing with the wrong timestamp.
Build systems triggering unnecessary rebuilds due to “past” or “future” timestamps.
Scheduled tasks running at unexpected times.
Logs from different OSes or containers being misaligned.

Using UTC is generally better because it avoids timezone-related issues and daylight saving changes. Linux servers, Docker containers, CI/CD pipelines, and cloud environments almost always operate in UTC, so aligning your local machine avoids subtle bugs in development workflows.

Resolving Old AUR App Dependency Hell after Arch Update (ICU 76 78)

hopsayer — Thu, 25 Dec 2025 00:01:31 +0000

Fixing ICU Library Dependency Hell on Arch Linux: A Deep Dive into Shared Library Mismatches

The Problem

I recently encountered a frustrating issue when trying to launch Notepadqq on my Arch-based system (EndeavourOS):

$ notepadqq 
/usr/bin/../lib/notepadqq/notepadqq-bin: error while loading shared libraries: 
libicui18n.so.76: cannot open shared object file: No such file or directory

This is a classic shared library dependency problem that's particularly common on rolling-release distributions like Arch Linux.

Understanding the Root Cause

What is ICU?

ICU (International Components for Unicode) is a mature, widely used set of C/C++ libraries providing Unicode and globalization support. Many applications depend on it, including Qt's WebEngine component.

The Rolling Release Challenge

The issue stemmed from a critical version mismatch between official Arch repositories and AUR packages:

What happened:

Arch upstream had moved forward to ICU 78 in official repositories
Notepadqq (an AUR package) was still compiled against ICU 76
Qt5WebEngine had been removed from official repos but remained on my system, still linked to ICU 76

The core problem: Notepadqq hasn't been actively maintained since 2018. As an AUR package, it doesn't get automatically rebuilt when Arch updates system libraries like ICU. This created a desynchronization between:

Official repository packages (ICU 78)
Outdated AUR binaries (expecting ICU 76)
Orphaned repository packages (Qt5WebEngine with ICU 76 linkage)

This cascading dependency mismatch is a common pitfall when mixing official packages with unmaintained AUR software.

Diagnostic Steps

1. Check Current ICU Version

$ ls /usr/lib/libicui18n.so*
/usr/lib/libicui18n.so
/usr/lib/libicui18n.so.78
/usr/lib/libicui18n.so.78.1

This confirmed I had ICU 78, while the application expected version 76.

2. Identify Package Origin

$ pacman -Ss notepadqq
$                         # Returns empty line
$ yay notepadqq
$ 1 aur/notepadqq 2.0.0beta-3 (+1 0.32) (Installed)

This confirmed notepadqq was installed, but being an AUR package (not from official repos), it wouldn't automatically rebuild when system libraries like ICU were updated.

3. Attempt to Rebuild

When I tried rebuilding Notepadqq from AUR:

$ yay -S notepadqq --rebuild

I hit a deeper issue:

/usr/bin/ld: /usr/lib/libQt5WebEngineCore.so: undefined reference to 
`icu_76::Normalizer::normalize(...)`
/usr/bin/ld: /usr/lib/libQt5WebEngineCore.so: undefined reference to 
`ucnv_open_76'

4. Discover the Real Culprit

$ pacman -Qo /usr/lib/libQt5WebEngineCore.so
/usr/lib/libQt5WebEngineCore.so is owned by qt5-webengine 5.15.19-4

The problem wasn't just Notepadqq—Qt5WebEngine itself was compiled against the old ICU version.

5. Repository Discovery

When trying to update qt5-webengine:

$ sudo pacman -Syu qt5-webengine
error: target not found: qt5-webengine

Critical finding: qt5-webengine had been removed from official Arch repositories but remained installed on my system as an orphaned package.

The Solution

Step 1: Remove Conflicting Packages

$ sudo pacman -Rns qt5-webengine notepadqq

This removed:

qt5-webengine (159.19 MiB)
notepadqq (5.89 MiB)
Related dependencies (qt5-location, qt5-webchannel, qt5-websockets)

Total freed: 174.88 MiB

Step 2: Rebuild Qt5WebEngine from AUR

$ yay -S qt5-webengine

This compiled qt5-webengine from source, linking it against the current ICU 78.

Step 3: Rebuild Notepadqq

$ yay -S notepadqq

With qt5-webengine now properly linked to ICU 78, Notepadqq compiled successfully.

Step 4: Verification

$ ldd /usr/bin/notepadqq | grep icu
        libicui18n.so.78 => /usr/lib/libicui18n.so.78
        libicuuc.so.78 => /usr/lib/libicuuc.so.78

Success! The application now links against the correct ICU version.

Key Lessons Learned

1. AUR Packages Don't Auto-Update with System Libraries

Critical insight: When Arch updates a core library like ICU, official packages are rebuilt automatically. AUR packages are not—they must be manually rebuilt by users or maintainers.

Notepadqq was last updated in 2018, meaning:

It never received rebuilds for ICU 76, 77, or 78
Each ICU update broke the package until manually rebuilt
This is a fundamental difference between official repos and AUR

2. Partial Upgrades Are Dangerous

Even though I was updating my system 2-5 times daily, I hadn't rebooted in 7 days. However, rebooting wasn't the issue — the problem was binary incompatibility on disk.

2. Rolling Release Requires Package Consistency

Arch's rolling release model means:

Libraries like ICU update frequently
All dependent packages must be rebuilt or updated in sync
Orphaned packages from removed repositories can cause hidden issues

3. ABI Compatibility Matters

The ICU version number in the shared library name (libicui18n.so.76 vs libicui18n.so.78) indicates an ABI break. You cannot simply symlink one version to another without risking crashes or undefined behavior.

Recognizing an ABI Break

How did I know this wasn't just a missing symlink issue?

1. Version in library filename:

libicui18n.so.76 vs libicui18n.so.78
The .76 and .78 are SONAME major versions
Different major versions = intentional ABI incompatibility

2. Versioned symbols in linker errors:

icu_76::Normalizer::normalize ← version baked into symbol name
ucnv_open_76 ← function name includes version
The library exposes icu_78:: symbols, but the binary needs icu_76::
No amount of symlinking can resolve this mismatch

3. Why symlinking would fail:
Even if we created ln -s libicui18n.so.78 libicui18n.so.76:

The binary would load, but call icu_76:: functions
The library only has icu_78:: functions
Result: immediate crash or undefined behavior

4. Dependency Chains Run Deep

What appeared to be a simple Notepadqq issue actually required:

Diagnosing the ICU mismatch
Discovering the Qt5WebEngine dependency
Identifying the orphaned repository package
Rebuilding the entire dependency chain

What NOT to Do

❌ Creating Symlinks

# DON'T DO THIS
$ sudo ln -s /usr/lib/libicui18n.so.78 /usr/lib/libicui18n.so.76

While this might allow the program to start, it can cause:

Runtime crashes
Undefined behavior
Data corruption
Difficult-to-debug issues

❌ Ignoring Update Errors

Regularly running pacman -Syu is good, but you must also:

Check for orphaned packages: pacman -Qtdq
Verify AUR packages after system library updates
Review failed transaction logs

Prevention Strategies

1. Regular Orphan Package Cleanup

$ pacman -Qtdq | pacman -Rns -

2. Monitor AUR Package Build Failures

When a system library updates, rebuild affected AUR packages immediately:

$ yay -S --rebuild <package-name>

Important: Unmaintained AUR packages (like Notepadqq, last updated 2018) will always need manual rebuilds after system library updates. Consider:

Finding actively maintained alternatives
Adopting the AUR package yourself
Using AppImage/Flatpak versions that bundle dependencies

3. Use Binary Package Alternatives

For stability-critical applications, consider -bin packages from AUR that bundle their dependencies, though this increases disk usage.

4. Check Library Dependencies

After major library updates:

$ ldd /usr/bin/<application> | grep "not found"

Conclusion

This issue perfectly demonstrates why Arch Linux is considered an "advanced" distribution. The rolling release model provides cutting-edge packages but requires users to understand:

Shared library versioning (ABI compatibility)
Dependency chains
The relationship between official repositories and AUR
When to rebuild vs. reinstall packages

The solution wasn't a quick fix but rather a systematic diagnosis and rebuild of the dependency chain. While frustrating in the moment, issues like this deepen understanding of how Linux systems work at a fundamental level.

Have you encountered similar dependency issues on Arch? What's your approach to handling library version mismatches? Share your experiences in the comments!

System Info:

Distribution: EndeavourOS (Arch-based)
Package Manager: pacman + yay
Affected Versions: ICU 76 → 78, Qt5WebEngine 5.15.19-4

NVIDIA Drops GTX 900/1000 Support in 590 Driver - Getting a Fix on Arch Linux

hopsayer — Tue, 23 Dec 2025 02:25:28 +0000

Fixing NVIDIA Driver Issues on Arch Linux After the 590 Update

The Problem: How You'll Know Something's Wrong

If you're running Arch Linux with an NVIDIA GTX 900 or 1000 series GPU (Maxwell/Pascal architecture), you might suddenly encounter this prompt when trying to update your system:

$ sudo pacman -Syu
resolving dependencies...
:: Replace nvidia-dkms with extra/nvidia-open-dkms? [Y/n]

Or when trying to install the nvidia driver:

$ sudo pacman -S nvidia-dkms
resolving dependencies...
warning: cannot resolve "nvidia-utils=590.48.01", a dependency of "nvidia-open-dkms"
:: The following package cannot be upgraded due to unresolvable dependencies:
      nvidia-open-dkms

You might have temporarily worked around this by ignoring NVIDIA packages in /etc/pacman.conf, but this isn't a sustainable solution. Your system updates are now incomplete, and you're missing important driver updates.

What Happened: The Context

On December 20, 2025, Arch Linux officially announced that NVIDIA driver version 590 no longer supports Maxwell (GTX 900 series) and Pascal (GTX 1000 series) GPUs.

Here's the timeline of changes:

Before: Arch provided nvidia-dkms and nvidia-utils in official repositories that supported most GPUs
Transition period: Arch began moving to nvidia-open-dkms (open-source kernel modules) for newer GPUs (Turing/Ampere and later)
December 2025: NVIDIA 590 dropped support for Maxwell/Pascal architectures
Current state: Official repos only support newer GPUs; older cards must use legacy drivers from AUR

Who Is Affected?

This affects users with:

GTX 900 series: GTX 950M, 960, 970, 980, 980 Ti, Titan X (Maxwell), GTX 750, 750 Ti, 745, GT 945a
GTX 1000 series: GTX 1050, 1060, 1070, 1080, 1080 Ti, Titan X (Pascal)
Quadro M-series and other Maxwell/Pascal professional cards
Mobile Maxwell GPUs: GTX 980M, 970M, 965M, 960M, 950M, Geforce 945M, 940MX, 940M, 930MX, 930M, 920MX, 920M, 910M, GTX 860M, 850M, 845M

Why Did This Happen?

NVIDIA's proprietary driver development follows this pattern:

Production branch (currently 590+): Supports only recent architectures (Turing and newer)
Legacy branches: Frozen versions for older architectures (580.xx for Maxwell/Pascal, 470.xx for Kepler, etc.)

Arch's decision to remove Maxwell/Pascal from official repos aligns with NVIDIA's EOL (End of Life) for these architectures in the production branch.

The Solution: Migrate to nvidia-580xx from AUR

The 580.xx driver series is the last version that officially supports Maxwell and Pascal GPUs. While it's no longer in Arch's official repositories, it's maintained in the AUR (Arch User Repository).

Step 1: Verify Your GPU Architecture

First, confirm your GPU model:

lspci | grep -i nvidia

For example:

01:00.0 3D controller: NVIDIA Corporation GM107M [GeForce GTX 950M]

Check if your card is Maxwell or Pascal:

Maxwell: GTX 900 series, GTX 700 series (some), Quadro M-series
Pascal: GTX 1000 series, Quadro P-series

Step 2: Back Up Your Settings (Optional but Recommended)

# Back up your NVIDIA settings
cp ~/.nvidia-settings-rc ~/.nvidia-settings-rc.backup

# Back up xorg.conf if it exists
sudo cp /etc/X11/xorg.conf /etc/X11/xorg.conf.backup 2>/dev/null || echo "No xorg.conf found"

Your user settings in ~/.nvidia-settings-rc won't be deleted during the process, but it's good practice to have a backup.

Step 3: Remove Conflicting Official Packages

The old packages conflict with the AUR versions we'll install. We need to remove them, including Steam if it's creating dependency issues:

# Force removal (ignore dependency warnings)
sudo pacman -Rdd nvidia-dkms nvidia-utils nvidia-settings lib32-nvidia-utils

Note: Using -Rdd (remove with dependency override) is necessary here because Steam and other packages depend on these. We'll restore everything immediately after.

You'll see errors about missing NVIDIA modules during initramfs rebuild—this is expected since we just removed the driver.

Step 4: Install nvidia-580xx from AUR

Now install the correct driver packages from AUR. If you're using yay:

yay -S nvidia-580xx-dkms nvidia-580xx-utils nvidia-580xx-settings lib32-nvidia-580xx-utils

If you're using paru:

paru -S nvidia-580xx-dkms nvidia-580xx-utils nvidia-580xx-settings lib32-nvidia-580xx-utils

What each package does:

nvidia-580xx-dkms: Kernel module sources (compiles driver for your kernel)
nvidia-580xx-utils: OpenGL, Vulkan, CUDA libraries and utilities
nvidia-580xx-settings: nvidia-settings GUI configuration tool
lib32-nvidia-580xx-utils: 32-bit support (required for Steam and many games)

During installation, you'll see:

Download of NVIDIA driver archive (~380 MB)
Package compilation (1-3 minutes)
DKMS module compilation (~30-120 seconds)
Successful initramfs generation (no more "module not found" errors!)

Step 5: Reinstall Steam (if needed)

If you had to remove Steam during the process:

sudo pacman -S steam

Step 6: Reboot and Verify

sudo reboot

After reboot, verify everything works:

# Check driver version
nvidia-smi

# Should show something like:
# +-----------------------------------------------------------------------------------------+
# | NVIDIA-SMI 580.119.02             Driver Version: 580.119.02   CUDA Version: 12.4     |
# +-----------------------------------------------------------------------------------------+

# Verify OpenGL
glxinfo | grep "OpenGL renderer"

# Should show your GPU name

What About Future Updates?

Good news: AUR packages update just like official packages!

# Regular system updates will include nvidia-580xx
yay -Syu
# or
paru -Syu

The only difference from official packages:

Initial install: Takes 1-3 minutes to build the package
Updates: Take 1-3 minutes when NVIDIA releases new 580.xx versions (rare)
Kernel updates: DKMS automatically recompiles the module (30-120 seconds, happens automatically)

Troubleshooting

Issue: "conflicting packages" during installation

If you see conflicts with libxnvctrl or similar packages:

# Remove the conflicting package when prompted
# yay will ask: "Remove libxnvctrl? [y/N]"
# Answer: y

Issue: Black screen after reboot

This usually means the module didn't load. Check:

# Check if module is loaded
lsmod | grep nvidia

# Check kernel messages
dmesg | grep -i nvidia

# Manually load module
sudo modprobe nvidia

Issue: "module not found" errors persist

Rebuild initramfs manually:

sudo mkinitcpio -P

Issue: Performance concerns

Q: Will downgrading from 590 to 580 hurt performance?

A: No. The 580.xx branch is the last version optimized for Maxwell/Pascal. Versions 590+ don't contain any optimizations for your GPU—they simply removed support for it. You might actually see better stability since 580.xx was specifically designed for your architecture.

Why Not Use nvidia-470xx Instead?

You might see some guides mentioning nvidia-470xx-dkms. Here's the difference:

470.xx: Legacy branch for Kepler architecture (GTX 600/700 series)
580.xx: Legacy branch for Maxwell/Pascal (GTX 900/1000 series)

If you have a GTX 950M or GTX 1060, you want 580.xx, not 470.xx. The 470.xx branch is for even older cards.

Alternative: Using nouveau (Open-Source Driver)

If you don't need CUDA or maximum 3D performance, consider the open-source nouveau driver:

# Remove NVIDIA proprietary driver
sudo pacman -Rdd nvidia-580xx-dkms nvidia-580xx-utils nvidia-580xx-settings lib32-nvidia-580xx-utils

# Install nouveau (if not already installed)
sudo pacman -S xf86-video-nouveau

# Reboot
sudo reboot

Nouveau pros:

Fully open-source
In official repos (no AUR needed)
Better Wayland support
No driver compilation needed

Nouveau cons:

Lower 3D performance (30-50% slower in games)
No CUDA support
Some power management issues on laptops

Summary

The migration to nvidia-580xx-dkms is straightforward:

Remove old official packages: sudo pacman -Rdd nvidia-dkms nvidia-utils nvidia-settings lib32-nvidia-utils
Install AUR packages: yay -S nvidia-580xx-dkms nvidia-580xx-utils nvidia-580xx-settings lib32-nvidia-580xx-utils
Reboot and verify with nvidia-smi

This gives you:

Proper driver support for Maxwell/Pascal GPUs
Automatic updates through pacman/yay
DKMS handles kernel updates automatically
Same performance as before
No more dependency conflicts

The 580.xx branch will continue to work until NVIDIA releases a version that drops your architecture entirely—which likely won't happen for several more years. Community patches in AUR will keep it building on new kernels.

References:

This guide was created based on real successful migration experience with a GTX 950M on Arch Linux (December 2025).

Linux Part 3: AI-Powered Learning, Post-Ubuntu Distros, Actually existing DE's, App Packaging&Delivery Approaches

hopsayer — Thu, 18 Dec 2025 14:49:41 +0000

Linux Part 3: AI-Powered Learning, Post-Ubuntu Distros, Which DE Actually Exist, App Packaging&Delivery Approaches

This is a continuation of the Linux intro series. If you haven't read Parts 1 and 2, start there - this assumes you understand basic Linux concepts like the FHS, package managers, and configuration management.

The Ubuntu Reality Check

Ubuntu has a reputation as the "beginner-friendly" distro. And it was - back in 2010. But here's the uncomfortable truth: Ubuntu's "simplifications" now create more problems than they solve.

You've already experienced this firsthand with your GPG key nightmares and library version conflicts. These aren't random bugs - they're symptoms of Ubuntu's architectural decisions.

The Problems With "User-Friendly" Ubuntu

Snap and Flatpak bloat:

Install 3 applications → They consume 50 GB of disk space
Performance is measurably slower than native packages
Updates happen separately from your system updates
Flatpak's hype died in 2022-2023 - if you need portability, use AppImage instead

Repository fragmentation:

Nearly every application requires manually adding a new repository
Each repository needs its own GPG key (hence your pain)
This creates a tangled web of trust relationships
Updates can break when repos become incompatible

"GUI for everything" isn't unique anymore:

In 2010, avoiding the terminal was Ubuntu's killer feature
In 2025, virtually every major distro has excellent GUI tools
You've already learned the terminal basics - you don't need Ubuntu to protect you from it

The library version hell:

Ubuntu holds packages at specific versions for "stability"
Newer software often needs newer libraries
You end up manually managing dependencies (remember trying to remove Wine libraries?)
This problem is architecturally impossible on rolling-release distros

What You Did Right

Focusing on one distro - Absolutely correct. Distro-hopping teaches you nothing. You learn the system by living with it through problems and solutions.

But you picked Ubuntu because of outdated advice. Let's fix that.

The Real Linux Landscape

Here's the truth: There are only 3 distros. Everything else is a derivative or niche project.

Ubuntu - Debian-based, corporate-backed, "stable" releases
Fedora - Red Hat-based, cutting-edge, 6-month release cycle
Arch - Rolling release, bleeding edge, community-driven

Pick your base. Everything else (Mint, Pop!_OS, Manjaro, etc.) is just someone's pre-configured version of one of these three.

Why Arch (Via EndeavourOS)

I'm recommending EndeavourOS - which is Arch with:

Graphical installer (no command-line installation dance)
Automatic driver detection and installation
Sensible defaults pre-configured
But it's still pure Arch under the hood

The Practical Benefits

Single, unified repository:

No PPA hell
No manual GPG key management
Everything comes from one source (pacman) or the community (AUR)
Your "which repo has this package?" question never comes up

Up-to-date packages:

New software versions arrive within days
Libraries update alongside programs
That Windows program that needs Wine 8.x? You have it already

The Arch Wiki:

Seriously, this alone justifies using Arch
Updated daily by the community
Every problem you'll encounter has a wiki page
Clear, technical, no-nonsense documentation
Example: wiki.archlinux.org - bookmark this

Fast package manager:

pacman is significantly faster than apt
AUR helper (yay) makes community packages feel native
Update your entire system: yay (just that, one command)

The "Arch Breaks on Updates" Myth

Reality: 90% myth, 10% minor inconveniences.

Here's what actually happens:

Small issues occur every few months
Fixes appear within 0-7 days
You check the Arch homepage before major updates
If there's a known issue, you wait a day or two

Your Ubuntu experience was worse:

GPG key failures
Trying to manually manage library dependencies
Repository version mismatches

On Arch, these problems structurally cannot happen because of the rolling release model and unified repository.

The Desktop Environment Question

Forget the 15 different desktop environments. There are only 2 that matter:

GNOME - Opinionated, keyboard-workflow focused, "our way or the highway"
KDE Plasma - Flexible, traditional desktop, infinitely customizable

Everything else (XFCE, Cinnamon, MATE, etc.) is either a fork of these two or a niche project.

You're on KDE already - good choice. It's the Windows-like experience done right. Keep it.

Learning Resources: Stop Googling

You mentioned Google is inefficient for learning Linux. You're absolutely right.

Use AI, Not Search Engines

Instead of Googling "how to install X on Linux":

Ask ChatGPT/Claude/DeepSeek:
"I'm on Arch Linux with KDE. I want to install [program]. 
What's the best way, and what should I know about it?"

You'll get:

Step-by-step instructions
Context about why you're doing each step
Warnings about common pitfalls
Alternative approaches

Google gives you: 10 outdated forum posts from 2015, half for the wrong distro.

Essential Websites (Not Google)

AlternativeTo.net - Finding Linux replacements for Windows programs

Search for your Windows app
Filter by "Linux"
See ratings, reviews, and actual user experiences

appdb.winehq.org - Checking Windows software compatibility with Wine

Search the program
See tested versions and what works/breaks
Community workarounds

DistroWatch.com - The distro leaderboard

Popularity rankings (with context)
Distro family trees
News about releases and changes

portablebspps.com - Finding portable Windows apps

Many installers break under Wine
Portable versions often work flawlessly
Skip the installer entirely

Gaming and Graphics: Real Talk

Forcing GPU Selection

For NVIDIA users (if you have hybrid graphics):

# Force NVIDIA for any program
__GLX_VENDOR_LIBRARY_NAME=nvidia program_name

# Force NVIDIA for Wine/Proton games
__GLX_VENDOR_LIBRARY_NAME=nvidia wine game.exe

This is way simpler than Windows's clunky NVIDIA Control Panel per-application settings.

PortProton Is Unnecessary

You don't need PortProton. Here's why:

What you actually need:

Steam (in the official repos)
Steam Runtime (included with Steam)
That's it

Steam's built-in Proton handles Windows games. PortProton adds a GUI wrapper that doesn't actually improve anything - it's training wheels for people afraid of Steam settings.

For Complex Cases: Lutris

If you need maximum control over Wine/Proton versions, use Lutris:

yay -S lutris

Features:

GUI for managing Wine versions
Per-game Wine/Proton selection
Built-in MangoHud support (FPS overlay)
GPU selection dropdown
Full manual control when you need it

But start with Steam. Only reach for Lutris when Steam Proton doesn't work for a specific game.

Configuration: The Portable Setup

Remember from Part 2: your entire Linux personality lives in ~/.config/.

On Arch, this becomes even more powerful:

# Backup your entire setup
tar -czf my-arch-setup.tar.gz ~/.config ~/.local ~/Documents

# New machine, new install
# After creating the same username:
tar -xzf my-arch-setup.tar.gz -C ~/

Everything returns:

All application settings
Custom keybindings
Menu customizations
Themes and appearances
Even your application menu entries from ~/.local/share/applications/

On Ubuntu, this works mostly. On Arch, it works perfectly because package versions are current and consistent.

The Transition Plan

If you're convinced, here's how to switch without losing your mind:

Phase 1: Preparation (While still on Ubuntu)

Document your installed programs: dpkg --get-selections > my-programs.txt
Backup /home/ entirely
Note any weird customizations you made
Download EndeavourOS ISO

Phase 2: Installation

Boot from USB
Choose "Online install" (gets latest packages)
Select KDE Plasma as desktop
Let it auto-detect drivers
20 minutes later, you're done

Phase 3: Setup

Copy back your ~/.config/ and ~/.local/ folders
Install your programs via pacman and yay
Most of your settings are already restored

Phase 4: Learning

Keep the Arch Wiki open
Ask AI when stuck
Join the EndeavourOS forums (helpful community)

Common Objections Addressed

"But I already spent time learning Ubuntu!"

You learned Linux fundamentals - file system hierarchy, package management concepts, configuration patterns. That knowledge fully transfers. You're just switching the implementation details (apt → pacman, PPAs → AUR).

"Arch seems harder."

Arch seems harder because it doesn't hide things from you. Ubuntu creates the illusion of simplicity by abstracting problems away - until they break, and then you're lost.

With Arch, you understand what's happening. That feels harder initially, but it's honest difficulty that makes you competent.

"What if I break something?"

You already broke things on Ubuntu (the Wine library incident). The difference on Arch is:

The wiki explains why it broke
The fix is clearer
The community has documented it
You learn the actual system, not Ubuntu-specific workarounds

The Bigger Picture

You're documenting your Linux learning journey in real-time, showing beginner struggles. This is valuable content - the Russian Linux community needs more honest, unglamorified guides.

Switching to Arch would:

Eliminate entire categories of problems you're hitting
Show viewers the difference between "beginner distro" and "actually functional distro"
Give you material for better tutorial content (Arch wiki → translate to Russian)

But more importantly: You'll stop fighting the system and start understanding it.

The Pluralism Zoo

Linux is a zoo of technologies and philosophies. Different package formats (deb, rpm, pkg), different init systems (systemd, OpenRC), different display servers (X11, Wayland), different approaches to everything.

This is not a bug. This is the feature.

It means:

No single point of failure (if Ubuntu dies, Debian continues)
Innovation happens at all levels
You can choose based on your actual needs
The best ideas survive through competition

But as a beginner, this feels overwhelming. That's why you pick one path and master it. You picked Ubuntu. I'm suggesting you pivot to Arch while your fundamentals are still fresh.

Making the Choice

Stay on Ubuntu if:

You're deeply invested in Snap-based workflows
You need enterprise LTS support
Change feels too risky right now

Switch to Arch (EndeavourOS) if:

You're tired of repository and key management
You want bleeding-edge packages
You value learning over hand-holding
The Arch Wiki appeals to your learning style

There's no wrong choice. But there is an informed choice.

Final Thoughts

Your instinct to ask "why does this work this way?" is correct. Ubuntu's complexity isn't deep complexity - it's accumulated cruft from trying to be both Debian and its own thing.

Arch's complexity is honest complexity. The system is what it appears to be. The wiki tells you the truth. The community expects you to think.

You're already thinking. You're already learning. You're already past the point where Ubuntu's "protections" help you.

The question isn't whether you can handle Arch. You already can.

The question is: do you want to keep fighting Ubuntu's architectural decisions, or do you want to use a system designed for people who actually care how things work?

You're making video guides. You're documenting everything. You're not running away from the terminal. You're exactly the person Arch is designed for.

Welcome to the deep end. The water's fine.

Next up: Part 4 could cover customization, themes, and making Linux truly yours - or a Wine/Proton deep-dive for running Windows software. Let me know what's more interesting.

Linux Advanced: Under the Hood - Philosophy And Usage Tips

hopsayer — Thu, 18 Dec 2025 12:01:42 +0000

Linux Advanced: Under the Hood - Deep Dive

This is a continuation of "Linux Quick Intro" - covering the technical details, philosophy, and advanced concepts that were glossed over in the beginner guide.

In this article you will discover:

Why Windows is deliberately different - The historical truth behind drive letters and registry
Advanced Linux directories - /proc, /dev, /sys and the virtual filesystem concept
The FHS philosophy - Why standardization beats flexibility
Package management internals - How AUR transforms .deb into native packages
Configuration portability - Set up Linux once, keep it forever
Symlinks vs Windows shortcuts - Why .lnk is a toy and symlinks are serious tools
Wine demystified - It's not emulation (and why that matters)
Context menu sanity - How Linux saved right-click from bloat
Desktop Environment landscape - KDE, GNOME, and the choice that matters
Driver architecture - Why your Wacom "just works"

Part 1: Why Linux Works This Way (And Why Windows Doesn't)

The Historical Truth

Linux (and UNIX before it) came first. When Microsoft created DOS and later Windows, they deliberately made different design choices - not because they were better, but to avoid looking like UNIX. This isn't conspiracy theory - it's documented history.

The result? Drive letters (C:\, D:) instead of a unified root. Registry instead of plain-text configs. Arbitrary install locations instead of standardized paths. These aren't features - they're deliberately different implementations of solved problems.

The bitter pill: Windows only seems "intuitive" because we're conditioned to its quirks. It's like learning a language with irregular grammar first - you think that's normal until you see a properly designed one.

Why This Matters

Linux's standardized paths mean:

Programs know where to find each other's files
Updates don't break things (if you follow the rules)
You can backup your entire configuration in kilobytes
The system doesn't accumulate cruft over time

Windows's approach means:

Program A can't reliably find Program B's data
Every installer asks where to put things
Uninstallers leave registry debris
After 6 months, you "need" to reinstall for "performance"

Part 2: File System Deep Dive

The Directories You Skipped

/proc/ - Not real files. It's a virtual filesystem exposing kernel and process information. Want to see your CPU info? cat /proc/cpuinfo. Current process list? It's all in /proc/. You rarely touch this directly, but system monitoring tools read from here constantly.

/dev/ - Device files. Your hard drive is /dev/sda, partitions are /dev/sda1, etc. Graphics card, USB devices - everything hardware appears here as a file. "Everything is a file" isn't a metaphor in Linux.

/sys/ - Another virtual filesystem for device management and kernel modules.

Symbolic Links: The Right Way

Remember my warning about recreating Windows structure? Here's why it fails:

Bad approach (what beginners try):

# DON'T DO THIS
mkdir /mnt/games  # "My D: drive"
ln -s /mnt/games/* ~/.local/share/Steam/steamapps/common/

Why it breaks:

Steam updates can delete and recreate directories, breaking links
File permissions get confusing (who owns what?)
You now have TWO sources of truth (actual location vs standard location)
Backup tools might follow links or ignore them unpredictably

Proper approach (when actually needed):

# Move ONE specific huge game, link it back
mv ~/.local/share/Steam/steamapps/common/MassiveGame /mnt/games/
ln -s /mnt/games/MassiveGame ~/.local/share/Steam/steamapps/common/MassiveGame

Only redirect specific items. Keep the standard structure intact.

Part 3: Package Management - How It Really Works

Anatomy of a `.deb` Package

When you download a .deb for Ubuntu, it's just a .tar archive with a specific structure:

package.deb
├── bin/
│   └── program_binary
├── lib/
│   └── libsomething.so
└── usr/
    └── share/
        ├── applications/
        │   └── program.desktop
        └── icons/
            └── program.png

Manual installation (educational purposes only):

# Extract
ar x package.deb
tar xf data.tar.xz

# Copy files to matching system paths
sudo cp bin/* /usr/bin/
sudo cp -r usr/share/* /usr/share/

This works, but it's terrible because:

No dependency tracking
No clean uninstall
Updates won't work
You're now maintaining this manually forever

How AUR Actually Works

When you run yay -S onlyoffice-bin, here's the full process:

Download PKGBUILD script from AUR (stored in ~/.cache/yay/onlyoffice-bin/)
Script executes, which:
- Downloads the .deb from OnlyOffice's official site (URL in script)
- Verifies SHA256 checksum (security)
- Extracts the .deb
- Repackages it as .pkg.tar.zst (Arch format)
- Calls sudo pacman -U package.pkg.tar.zst (install local package)
Pacman tracks this in its database - now you can update/remove it properly

For source packages (no -bin):
The PKGBUILD instead:

Downloads source code
Installs build dependencies
Compiles the program (this is why your CPU spikes for 20+ minutes)
Packages the compiled binaries

This is why -bin versions are preferred when available - they skip compilation.

The `yay` vs `yay -S` distinction

yay package_name      # Search ALL packages, let you choose
yay -S package_name   # Install exactly this package name (must be exact)

Practical example:

yay discord
# Shows:
# 1. discord (official repo)
# 2. discord-bin (AUR)
# 3. discord-canary (AUR)
# 4. discord-ptb (AUR)
# Choose: 1

yay -S discord
# Immediately installs 'discord' from official repo, no questions

Part 4: Configuration Management - The .config Philosophy

Why `/etc/` Edits Get Overwritten

When you install a program, it creates default configs in /etc/. Example:

/etc/
└── grafana/
    └── grafana.ini

If you edit /etc/grafana/grafana.ini and later update Grafana, the package manager overwrites this file with the new default.

The Correct Workflow

# 1. Copy to user config
cp -r /etc/grafana ~/.config/

# 2. Edit YOUR copy
nano ~/.config/grafana/grafana.ini

# 3. Done - this won't be touched by updates

Why this is genius:

Your changes persist across updates
You can version-control ~/.config/ (Git)
Reinstalling the OS? Just copy ~/.config/ and ~/ - all your settings are back
Multiple users on one machine? Each has their own configs

Priority system:

~/.config/program/ (user config) - HIGHEST
/etc/program/ (system config)
Built-in defaults in the program itself

The Portability Dream

Your entire personalized Linux setup in one folder:

~/.config/        # 500 KB - hours of customization
~/.local/share/   # Personal data
~/Documents/      # Your files

Reinstalling? Switching distros? New machine?

# Old system
tar -czf my-linux-life.tar.gz ~/.config ~/.local

# New system (after creating same username)
tar -xzf my-linux-life.tar.gz -C ~/

Everything returns: keybindings, themes, application settings, menu customizations. This is why Linux users rarely "need" to reinstall.

Part 5: Desktop Files and Icons - The Separation Mystery

Why Are They Separate?

You asked why .desktop files and icons aren't bundled like Windows .exe + embedded icon. Two reasons:

Historical: Two competing standards (freedesktop.org vs older X11 conventions)
Practical: Icon themes. Change your system theme, ALL icons update without touching each program

The structure:

/usr/share/applications/     # Menu entries
/usr/share/icons/            # Icon theme system
/usr/share/pixmaps/          # Legacy single icons

Your user overrides:

~/.local/share/applications/ # Custom menu entries
~/.local/share/icons/        # Personal icons
~/.icons/                    # Additional personal icons

Creating Custom Launchers

Scenario: You downloaded a portable AppImage.

# 1. Move it somewhere permanent
mkdir -p ~/Applications
mv ~/Downloads/MyApp.AppImage ~/Applications/
chmod +x ~/Applications/MyApp.AppImage

# 2. Create .desktop file
nano ~/.local/share/applications/myapp.desktop

Contents:

[Desktop Entry]
Type=Application
Name=My Application
Exec=/home/yourname/Applications/MyApp.AppImage
Icon=/home/yourname/Applications/myapp-icon.png
Categories=Utility;

Wait 5 seconds. Open your application menu - it's there. No installers, no registry, just a text file.

Part 6: Gaming Deep Dive - Steam, Proton, Wine

How Steam Actually Launches Games

Native Linux game:

Steam → Steam Runtime → Game binary

Windows game:

Steam → Steam Runtime → Proton (Wine) → Game binary

Steam Runtime is a containerized environment with specific library versions. This ensures games work regardless of your distro's library versions - it's like a mini-OS for games.

Wine Is NOT Emulation

Common misconception: "Wine emulates Windows."

Reality: Wine implements Windows API calls natively on Linux.

Analogy: Imagine a program asking "give me file X from C:\Windows\System32\". Wine catches this request and responds: "Here's the equivalent Linux library." The program never knows it's not on Windows.

No CPU emulation = Native performance (often better than Windows due to lower OS overhead)

Why some games work better on Linux:

Example: GTA IV - 25-30 FPS on Windows, stable 60 FPS on Linux/Proton
Reason: Better GPU detection, less background bloat, more efficient memory management

Checking Compatibility

For games: protondb.com

Platinum: Works flawlessly, zero issues
Gold: Works with minor tweaks (usually just launch options)
Silver: Runs but has bugs
Bronze: Barely playable
Garbage: Doesn't work

For Windows programs: appdb.winehq.org (older, less active)

Pro tip: Ask an AI with web search: "What's the best Photoshop version for Wine on Linux?" - users have collectively tested this stuff extensively.

The Steam Filter "Issue"

You noticed the penguin icon filter doesn't change game counts anymore. This is intentional.

Old behavior:

Penguin icon = show only native Linux games
Game count changes based on filter

New behavior (since Steam Deck):

"Enable Steam Play for all titles" is default ON
Steam assumes EVERYTHING can run via Proton
Penguin filter now meaningless (shows all games)

Valve's logic: Proton compatibility is so good (~95%+ of games work) that distinguishing native vs. Windows games is pointless for users.

For enthusiasts: This is "dumbing down" - removes information. For normies: This is "just works" - they don't need to care.

Part 7: The Templates Feature - Right-Click Menu Control

Windows Hell vs Linux Zen

Windows:

Install Adobe Reader → 5 new right-click menu entries
Install WinRAR → 8 new entries
Install 20 programs → Context menu takes 3 seconds to load
No official way to clean this

Linux:

Programs don't automatically add menu entries
You control what appears via the Templates folder

How It Works

Everything in ~/Templates/ appears in "Create New..."

cd ~/Templates/

# Create a blank text file
touch "New Text Document.txt"

# Create a blank PNG
pinta  # (open Pinta, create 800x600 white canvas, save as "New Image.png")

# Create a spreadsheet template
libreoffice --calc  # (create, save as "Budget Template.ods")

Now right-click anywhere:

"New Text Document.txt" appears
"New Image.png" appears
"Budget Template.ods" appears

Advanced: Make templates actually useful. Put a pre-formatted resume, a blank D&D character sheet, your company's letterhead - whatever you create often.

Desktop Environment Differences

KDE (your system): May have some defaults pre-created
GNOME: Completely empty by default - Dark Souls mode
XFCE/MATE: Usually has a basic text file

Part 8: Hardware and Drivers

The `/lib/modules/` Path

Drivers live in /lib/modules/$(uname -r)/ where $(uname -r) is your kernel version.

You almost never touch this manually. Drivers install/update via:

sudo pacman -S linux-headers   # Build environment for drivers
sudo pacman -S nvidia-dkms      # NVIDIA drivers (example)

Wacom Tablets

You mentioned having a Wacom - it likely works out of the box. Linux has excellent Wacom support built into the kernel. No drivers needed.

Test:

xsetwacom --list devices

If it appears, it works. Pressure sensitivity included.

Configure:

KDE: System Settings → Input Devices → Graphic Tablet
GNOME: Settings → Wacom Tablet

Krita (the painting app I mentioned) has native Wacom support and is genuinely used by professional digital artists.

Part 9: The Mindset - Discipline and Standards

Why Linux Developers Are More Disciplined

You noted: "It's strange that Linux, being decentralized, has better standardization than Microsoft with its central control."

The reason: Open source = public accountability.

Microsoft: Internal teams don't see each other's mess. Politics protects bad code.
Linux: Everything is public. Bad design gets called out. Meritocracy (mostly) wins.

Example: The Filesystem Hierarchy Standard (FHS) is a document anyone can read. Everyone follows it because:

It makes sense
Breaking it makes your software incompatible
The community will roast you

The Learning Curve Investment

Timeframe for comfort:

Smart user, focused learning: 2 months
Casual use, learning as needed: 4-6 months
Just following guides blindly: May never truly understand

What changes when you "get it":

You stop thinking "I wish this was more like Windows"
You start thinking "Why isn't Windows more like this?"
You see the elegance in standardization
You trust the system instead of fighting it

When Manual Compilation Is Actually Hard

I've spent 3-6 hours on a single package compilation, sometimes days when dependencies were tangled.

Early attempts: ~20% success rate
After learning: ~90% success rate

But why bother? AUR already has 60,000+ packages. Someone else has already done this work. Use their PKGBUILD.

Only compile manually when:

Cutting-edge development version not in AUR yet
Specific patch you need
Learning/curiosity

Part 10: Practical Troubleshooting Patterns

"I Need Program X, But There's No Arch Package"

Step 1: Search official repos

pacman -Ss program_name

Step 2: Search AUR

yay program_name  # Will show all AUR results

Step 3: Google "program_name arch linux"

You'll find either:
- AUR package you missed
- Wiki page explaining alternative
- Someone's installation guide

Step 4 (rare): Check official website

Download .deb → Extract manually (last resort, unclean)
Download source → Create your own PKGBUILD (advanced)
Use Flatpak/Snap (containerized, bigger downloads)

When Things Break

Config got messed up?

# Don't reinstall the program - just reset its config
rm -rf ~/.config/program_name
# Restart program - fresh defaults

Want to cleanly remove a program?

sudo pacman -R package_name        # Remove program
sudo pacman -Rs package_name       # Remove + unused dependencies
rm -rf ~/.config/package_name      # Remove user configs
rm -rf ~/.local/share/package_name # Remove user data

System update broke something?

# Downgrade to previous version
sudo pacman -U /var/cache/pacman/pkg/package-old-version.pkg.tar.zst

Closing Thoughts

This guide covered what the Quick Intro skipped - the philosophy, the technical depth, and the "why" behind Linux's design decisions.

Key takeaways:

Linux's design isn't arbitrary - it's intentional and well-reasoned
Standards exist for your benefit, not to constrain you
The learning curve is real, but the payoff is a system you actually understand
"User-friendly" often means "hides complexity" - Linux respects your intelligence

What's next?

Part 3 could cover: Wine deep-dive, running Windows software, Photoshop/Office alternatives in detail
Part 4 could be: Customization - themes, window managers, making Linux truly yours
Part 5: Server use, self-hosting, why Linux dominates infrastructure

You've now gone from "What's a /usr/bin/?" to understanding the entire system architecture. That's not a small achievement. Most Windows users don't understand their own OS at this level.

Final wisdom: The frustrations you feel now (remembering paths, using the terminal) are temporary. The understanding you're building is permanent. In a few months, you won't remember Windows fondly - you'll remember it as the training wheels you've outgrown.

Welcome to the deep end. The water's fine once you stop fighting the current.

Linux Usage Intro: System Folders Explanation, Proton, Apps Install&Remove

hopsayer — Tue, 16 Dec 2025 16:33:20 +0000

Linux Usage Quick Intro: From Confusion to Clarity

If you're new to Linux, especially coming from Windows, the operating system can feel like a completely different world. The lack of C:\ and D:\ drives, the terminal-centric approach, and the different file system hierarchy can be overwhelming at first. This guide aims to demystify Linux usage, focusing on Arch Linux (though most concepts apply universally), and help you understand the why behind the system's design.

The Core Philosophy: Order Over Chaos

Unlike Windows, Linux follows a unified, hierarchical file system that starts from a single root: /. There's no concept of separate drives in the path structure. When you install programs, they go into standardized, predictable locations by default. Why?

Interoperability: Program A can reliably find files from Program B.
Predictability: You (and the system) always know where to look for things.
No manual path selection: You don't need to choose "install locations" for each application.

The rare exception? Massive games or datasets that don't fit on your primary drive. For these, you use symbolic links (advanced "shortcuts") to redirect from standard paths to other locations.

Navigating the File System: Key Directories

Open your file manager, type / in the address bar, and press Enter. Welcome to the root. Here's what matters:

/home/ – Equivalent to C:\Users\. Your personal folder is /home/your_username/. This is typically where your file manager opens.
/etc/ – The configuration hub. Contains config files for most system-wide applications. Crucial: If you edit configs here, they might get overwritten on updates. The proper way is to copy them to your user config directory (more on that later).
/tmp/ – Temporary files. Browser caches, app temp data, etc.
/usr/ – User programs and data (despite the name).
- /usr/bin/ – Main executable binaries for installed programs.
- /usr/share/ – Shared data: icons, .desktop files (menu entries).
/opt/ – Optional/third-party large applications (Steam, some proprietary software).
/var/ – Variable data like logs (/var/log/).
/bin/, /sbin/, /lib/ – Essential system binaries and libraries. Rarely need manual interaction.

Your Home Directory: The Hidden Power

Inside /home/your_username/, you'll find visible folders (Downloads, Documents) and hidden ones starting with a dot (e.g., .config). Enable "Show Hidden Files" in your file manager (Dolphin: Ctrl+H or View menu).

Key hidden directories:

.config – Your personal configuration center. Copy and edit config files from /etc/ here. They won't be overwritten by updates, and they're portable between installations.
.cache – Local cache (browsers, apps). User-specific version of /tmp.
.local/share/ – Your personal AppData\Roaming.
- .local/share/applications/ – Place custom .desktop files here to add programs to your application menu.
Templates – A brilliant feature. Whatever files you put here (e.g., New Image.png, Resume.docx) will appear in your file manager's right-click "Create New" menu. No more cluttered context menus!

Installing Software: Repositories Are Your Friend

Forget downloading .exe files from websites. Linux uses package managers.

Official Repositories (`pacman` in Arch)

Thousands of vetted, maintained packages.

# Search for a package
pacman -Ss package_name

# Install
sudo pacman -S package_name

# Remove
sudo pacman -R package_name

AUR (Arch User Repository) via `yay`

A community-driven repository with virtually every program imaginable. If it's not in the official repos, check AUR.

# Install yay (if not present)
sudo pacman -S yay

# Install from AUR
yay -S package-name-from-aur

Pro tip: yay will show multiple options. Usually pick the first one, especially if it has a -bin suffix (pre-compiled, faster installation).

How AUR works: Community members create PKGBUILD scripts that automatically:

Download the program from its official source (even .deb files for Ubuntu)
Compile or extract it
Place files in correct Linux paths
Create a native Arch package

Gaming on Linux: Steam & Proton

Steam typically installs to ~/.local/share/Steam/ or /opt/.
Proton is a compatibility layer (Wine) optimized for games. It's not emulation – it translates Windows API calls to Linux on the fly.
To enable Proton: Right-click a Windows game → Properties → Compatibility → Enable Steam Play → Choose Proton version.
Check game compatibility at protondb.com.
Steam's new behavior: With "Enable Steam Play for all titles" checked, your library shows ALL games regardless of OS. The Linux filter (penguin icon) won't change counts – Steam assumes everything can run via Proton.

Essential Software Replacements

Graphics:
- sudo pacman -S krita – Digital painting (great with tablets)
- sudo pacman -S gimp – Photo editing (with Photoshop-like extensions)
- sudo pacman -S inkscape – Vector graphics
- sudo pacman -S pinta – Simple image editing
Office Suites:
- sudo pacman -S libreoffice-fresh – Full-featured free suite
- yay -S onlyoffice-bin – Best MS Office format compatibility

The Mindset Shift

Yes, there's a learning curve. You need to remember paths and get comfortable with the terminal. But this is an investment that pays off:

Predictable system behavior
Easy configuration backup (just copy .config)
Clean software management (no registry, no scattered files)
No context menu bloat

Don't try to force Windows patterns onto Linux. Embrace the standardized paths, trust the package manager, and within a couple of months, you'll not only understand the system – you'll appreciate its elegance.

Final tip: When you need software, first search your package manager (pacman -Ss or yay). If it's not there, search "program_name arch linux" – you'll usually find it in the AUR. This approach keeps your system clean and manageable.

Welcome to Linux – where things make sense once you understand the logic!

GTA Vice City Nextgen Edition on Linux/Steam Deck - Guide

hopsayer — Sat, 08 Nov 2025 00:00:31 +0000

Starting GTA Vice City Next Edition on Linux - Complete Guide

Quick Reference for Experienced Users

Install game with GE-Proton10-15
Add to Lutris with Wine runner
First launch: wine-ge-8-26 + d3d9 override (will error but properly creates prefix)
Second launch: wine-10.16-staging + d3d9 override (will work)
Install "Better Handling" and "Sprint Inside Interiors" mods
Set all graphics to Low
Backup saves from: ~/gtavcne/0_pfx/drive_c/users/steamuser/Documents/Rockstar Games/GTA VC/savegames

When to Use This Guide

This guide will help if you're experiencing:

Black screen or instant crash (but splash/menu/loading screens work)
Black screen crash after 5-10 seconds with no splash screen
Instant crash on loading screen
Instant crash during/after intro cut scenes
Splash screen appears but menu crashes instantly
Missing "ENB series" splash screen on startup
Any combination of the above issues
Official instructions are unclear or don't work

Requirements for Successful Launch

Successful Launch Conditions

Official game distribution from developers
NO PATCH needed - V1.2.7 patch can actually cause performance issues
Compatibility layer: Wine and Proton (via Lutris - strongly recommended)
d3d9 override: Must use game's own d3d9.dll via WINEDLLOVERRIDES="d3d9=n"
dsound override NOT needed - Contrary to official instructions, WINEDLLOVERRIDES="dsound=n,b" disables sound completely in my testing
First start with wine-ge-8, second with wine-10.16-staging

Why Lutris is Strongly Recommended

Even if Lutris seems unfamiliar to you, use it:

It's light, easy, GUI, quick, automated, has both Wine and Proton, lets quickly switch between them/their versions
It often has better compatibility and success rates than vanilla Steam Proton
Built-in MangoHud integration (FPS counter, performance monitoring)
Effortless, explicit and efficient choice of used graphics card in GUI game settings

CRITICAL: Wine Version Requirements

WARNING: The game requires a specific two-step Wine configuration:

First launch must use wine-ge-8-26 (or similar wine-8.x version)
- This creates the Wine prefix correctly, while the game gives the error "GTA IV FATAL ERROR: Invalid resource detected - Please re-install the game", it's OK for this time
- If you skip this step, Wine 10 will delete the prefix folder
After successful prefix creation, switch to wine-10.16-staging
- This is needed for the game to actually run
- Launched under Wine 8 again, the game will give the above error endlessly

Why this happens: Unknown RAGE engine quirk. But it works reliably and repeatably.

Proton Does NOT Work for This Game

proton ge 10 gives crash on start with rockstar logo
proton ge 9 gives crash on start without rockstar logo
proton 9.0 beta gives crash on start with rockstar logo shown in wrong aspect ratio
proton experimental hangs on "launching" state
proton 10.0 gives crash on start with rockstar logo shown in wrong aspect ratio
proton hotfix gives crash on start with rockstar logo shown in wrong aspect ratio
GE-proton(latest) gives crash on start with rockstar logo

USE WINE, NOT PROTON. This behavior is consistent whether launching from Lutris or Steam.

Part 1: Setting Up Lutris and Wine

Step 1: Install Lutris

sudo pacman -S lutris

Step 2: Initial Lutris Setup

Launch Lutris
It will auto-install dependencies (you'll see progress bar bottom-left)
- This includes Lutris-patched Wine versions (separate from system Wine)
- Installs to ~/.local/share/lutris/runners by default
Make sure you have both:
- wine-ge-8-26-x86_64 (or similar wine-8.x)
- wine-10.16-staging-amd64-wow64-X86_64 (or similar wine-10.x)

If these aren't installed automatically:

Go to Lutris → Preferences → Runners → Wine
Click the download icon next to the versions you need

Part 2: Installing the Game

Step 1: Prepare Installation Directory

# Create main game folder (choose location with enough space)
mkdir ~/gtavcne

# Create Wine prefix subfolder
mkdir ~/gtavcne/0_pfx

Step 2: Add Installer to Lutris

Click "+" button in Lutris → "Add locally installed game"
Configure installer entry:
- Name: vcne-installer (or whatever you prefer)
- Runner: Wine
- Wine version: GE-Proton10-15 or similar (not clean Wine)
WARNING: clean Wine will give empty/black installer window. Use GE-Proton for installation only, then switch to Wine for gameplay
Switch to "Game Options" tab:
- Executable: Path to vcNE_setup.exe
- Wine prefix: ~/gtavcne/0_pfx
Click "Save"

Step 3: Run Installation

Launch the installer from Lutris
Choose installation target in Windows notation: ~/gtavcne (or your chosen location)
Wait for unpacking to complete
- If installer can't see your disk, install elsewhere then move the folder
DO NOT install the patch (v1.2.7)
- It pulls FusionFix as a dependency which damages performance
DO NOT launch the game from installer - click "Complete" instead
Backup the clean installation:

   cp -r ~/gtavcne ~/gtavcne-backup

This is extremely useful for troubleshooting. Delete later when stable.

Part 3: Configuring and Launching the Game

The Critical Configuration

Key requirements:

Use Wine (not Proton)
Override d3d9.dll and check if DXVK enabled (should be by default)
Start with wine-ge-8-26 to create prefix
Switch to wine-10.16 for actual gameplay

Step 1: Add Game to Lutris

Click "+" → "Add locally installed game"
Configure game entry:

Game Info Tab:

Name: GTA Vice City Next Edition

Game Options Tab:

Runner: Wine
Executable: ~/gtavcne/LaunchGTAIV.exe
Wine prefix: ~/gtavcne/0_pfx

Runner Options Tab:

Wine version: Select wine-ge-8-26-x86_64 (for first launch)
Enable DXVK: Must be checked, should be ON by default

System Options Tab:

DLL overrides:
1. Click "Add"
2. Key: d3d9
3. Value: n
This forces Wine to use the d3d9.dll inside the game folder
Show MangoHud overlay: Check this for FPS counter (optional)
- If unavailable, install MangoHud: sudo pacman -S mangohud and restart Lutris

For laptops with NVIDIA GPU:

Go to System Options tab
Find "Discrete GPU" → Select your NVIDIA card

Click "Save"

Step 2: First Launch (Wine Prefix Creation)

Launch the game from Lutris with wine-ge-8-26
You'll see: "GTA IV FATAL ERROR: Invalid resource detected - Please re-install the game"

THIS IS EXPECTED! The prefix was created successfully

Close the error

Step 3: Switch to Wine 10

Right-click the game in Lutris → Configure
Go to Runner Options tab
Change Wine version to: wine-10.16-staging-amd64-wow64-X86_64
Save

Step 4: Successful Launch

Run the game from Lutris again
Success indicators:
- You should see "ENB series" splash screen
- Rockstar logo appears
- Initial cutscenes render properly (not black)
- Game graphics display correctly

Note:

Game may crash after helicopter cutscene - just restart if this happens
Skipping cutscenes too quickly can cause crashes - let them play a bit

Essential Mods for Playability

1. Better Handling for GTA VC NE (CRITICAL)

WITHOUT THIS MOD, CARS ARE NEARLY UNCONTROLLABLE.

Install from the mod's folder or website to make vehicles drivable. This mod replaces just one file.

2. Sprint Inside Interiors (HIGHLY RECOMMENDED)

WITHOUT THIS MOD, YOU CANNOT USE TAPE-SAVING FEATURE INSIDE BUILDINGS.

This dramatically improves gameplay quality of life. This mod replaces just 5 files.

Installation

Both mods should have installation instructions in their folders. Generally:

Copy mod files to game directory
Merge with existing files if prompted

Performance Optimization

Enable FPS Counter (MangoHud)

Already covered in Step 1 above - just tick "Show MangoHud overlay" in System Options.

For advanced customization, create/edit ~/.config/MangoHud/MangoHud.conf:

position=top-right
font_size=10
fps
frametime
cpu_temp
gpu_temp
cpu_power
gpu_power
gpu_load
cpu_load
ram
vram
background_alpha=0.0
alpha=0.3
frame_timing=1

Low FPS / Stuttering Issues

Symptoms: 25-35 FPS with severe stuttering (long frametimes), especially on:

Laptops with NVIDIA GPUs below GTX 1000 series
Older/mid-range hardware

Solutions (in order of priority):

Set everything to "Low" in graphics settings - CRITICAL

This game's graphics settings are extremely demanding. Set ALL options to LOW for playable performance.

In-game settings: Graphics → Set everything to Low/Off where possible

Lower Shadow Quality specifically

Shadows are the biggest performance killer. Set to Low (provides huge improvement).

DO NOT install the V1.2.7 patch

The patch can actually worsen performance (FusionFix component). If already installed, restore from your backup.

Lower Resolution

Try 1600x900 or 1280x720 if you're at 1920x1080.
In-game settings: Graphics → Resolution

Note: Lower resolutions cause the game to start in borderless window mode.
Fix: Minimize the game, then click on its window again to stretch fullscreen.

Do NOT use commandline.txt

The commandline.txt file is useless for this instance. Don't modify it.

Expected results after optimization: 45-60+ FPS with smooth frametimes

Troubleshooting

Resolution/Window Issues

Problem: After lowering resolution, game starts in borderless window instead of fullscreen.

Solution:

Minimize the game (Alt+Tab or Super key)
Click on the game window to focus it again
It will stretch to fullscreen

Save Game Backups

Your saves are located at:

~/gtavcne/0_pfx/drive_c/users/steamuser/Documents/Rockstar Games/GTA VC/savegames

Cutscene Crashes

Don't skip cutscenes too quickly
Let them play for at least a few seconds before skipping
If crash occurs, just restart the game

Final Notes

Now your only challenge will be the incompatibility between GTA: Vice City and GTA IV mission playstyles/balance, not launching the game.

Tested Environment

This guide was tested on:

OS: EndeavourOS
Kernel: 6.17.7-arch1-1
Lutris: lutris-0.5.19
Wine (Lutris): wine-ge-8-26-x86_64, wine-10.16-staging-amd64-wow64-X86_64
Proton-GE: GE-Proton10-15, GE-Proton9-27
DXVK: lib32-vulkan-icd-loader 1.4.328.1-1
MangoHud: v0.8.1

To get your system info, run this command:

echo "OS: $(cat /etc/os-release | grep PRETTY_NAME | cut -d'"' -f2)" && \
echo "Kernel: $(uname -r)" && \
echo "Lutris: $(lutris --version 2>/dev/null || echo 'not installed')" && \
echo "Wine (Lutris): $(ls ~/.local/share/lutris/runners/wine/ 2>/dev/null | grep -E 'lutris|wine' | tail -1 || echo 'not found')" && \
echo "Proton-GE: $(ls ~/.steam/root/compatibilitytools.d/ 2>/dev/null | grep -i proton | tail -2 || echo 'not found')" && \
echo "DXVK: $(pacman -Q dxvk-bin 2>/dev/null || pacman -Q lib32-vulkan-icd-loader 2>/dev/null || echo 'check via Lutris')" && \
echo "MangoHud: $(mangohud --version 2>/dev/null || echo 'not installed')"

mpv vs VLC: Making mpv Comfortable to Use

hopsayer — Thu, 17 Jul 2025 12:48:22 +0000

VLC or mpv? That's the question

Both are excellent open-source, cross-platform media players and serve as core components for many other apps. Because they often come bundled as dependencies, especially in Linux, it’s not always obvious which one to pick. Plus, debates about which is better happen a lot.

VLC vs mpv: The Basics

VLC is popular for a reason: it has a full GUI out of the box, supports tons of formats, and can stream YouTube videos, handle DVDs, and almost anything else without fuss.

However, VLC can struggle with modern codecs like HEVC/H.265 — even with codecs installed, playback might lag, fail to decode, or overload your CPU. This problem is worse with Flatpak or Snap versions, where codec support can be limited.

mpv, on the other hand, handled all my HEVC videos flawlessly from the start. It’s lightweight, highly customizable, and perfect for Linux users who like to tweak configs or use CLI tools. By default, mpv has no GUI distractions, but you can add frontends like Celluloid or smplayer if you want.

So I chose using mpv, but it needs some configuring to be comfortable for daily use.

Making mpv Comfortable

To make mpv a daily driver, I customized it via config files:

`~/.config/mpv/mpv.conf`

hwdec=auto              # Enable hardware decoding if available to reduce CPU load  
keep-open=yes           # Keep window open after playback finishes  
loop-file=inf           # Loop current file indefinitely  

cursor-autohide=3000    # Hide mouse cursor after 3 seconds idle  
osd-duration=3000       # Show on-screen messages for 3 seconds

`~/.config/mpv/script-opts/osc.conf`

visibility=auto         # Show control panel (OSC) on mouse movement, then auto-hide  
hidetimeout=2000        # Hide control panel after 2 seconds idle  
fadeduration=500        # Fade-out animation lasts 0.5 seconds

You can download these configs with the full file hierarchy preserved from my GitHub repo.

Final Thoughts

If you want hassle-free playback with maximum control, mpv is the way to go. It has a steeper learning curve than VLC, but a few config tweaks can make it super comfortable and tailored to your workflow.

If you deal with HEVC videos or want a lightweight player, mpv beats VLC hands down.

Try my configs and adjust them to your taste—you might be surprised how slick mpv can be.

Was this article any useful?

Don't be shy to write impressions and suggestions in the comment section!

Casting PC Screen to Android (Linux/Mac/Windows)

hopsayer — Tue, 08 Jul 2025 03:30:34 +0000

Ah. I faced a problem in my 25.

I often watch YouTube via my PC, but unfortunately I need to eat sometimes in my kitchen. This is not a reason to turn the video off - it's rather the opposite.

So there are 2 ways:

Copy the video's link on PC, send it to Android, wait until it fully loads, rewind the video to the moment I stopped at and continue viewing. I manage to finish my food doing this.
Share the PC screen with Android and continue viewing seemlessly. Providing that you have bluetooth headphones. What I do.

The simplest way is Deskreen.

This is an easy-to-use, cross-platform, free and open-source app that lets you share your desktop over Wi-Fi using any browser, including on Android or iPad.

Here’s a guide to start up quickly:

Download and install Deskreen from the official web site on your PC (Windows, Mac, or Linux).
Launch Deskreen on your PC. It will show a QR code and a local IP address.
Scan the QR code with your phone’s camera or manually enter the IP address shown in Deskreen into your Android browser.
Deskreen will ask you to allow sharing your entire screen or a specific application window — choose your preference and confirm.
Your PC screen should now be mirrored on your Android device in real time.
You can connect multiple devices by repeating the above steps.

Tips:

5GHz Wi-Fi can provide lower latency.
Deskreen supports encrypted connections for safer casting.
Adjust the resolution and bandwidth in Deskreen settings to optimize performance and battery usage.

Additional features:

Mirroring the screen is not the only option. You can also use your another device as a physical second screen. Any of your apps can be mouse-dragged on it. A tablet can be transformed into a monitor. A Virtual Display Adapter is required. Without it, you can only stream an existing screen.
You don’t have to share the whole screen; you can share specific apps — good for privacy and focus.
Multiple connections simultaneously.
Any resolution can be chosen. Auto adjustment can be set.
2 media players are available for better experience and reliability. I prefer the second one, which is not the default.
No any mobile installations.
SmartTV is supported.
Can be used for presentations or video speech recording.
Teleprompter mode: you can flip the streamed image horizontally.

Why I think this could be the best way for screencasting:

User-friendly. Quick to set up, despite the underlying complexity. Clear and simple tutorial on the main page, including video instruction. The product really invites you to try it.
Cross-platform. All major desktop and mobile OS.
Open-source, which allows anyone to take part, just like me. Even if the project is frozen.

Why I actually do not:

Connection can suddenly drop if you leave the Android browser tab for 1+ minutes. In that case the user has to create a new connection, which sometimes ruins the seamless living room to kitchen switching experience. I decided to participate by creating an issue on GitHub, if you're interested: https://github.com/pavlobu/deskreen/issues/280
Deskreen streaming experience is slightly worse on Android Firefox due to how the browser handles background tabs. Picture-in-picture mode is harder to get, and background tabs tend to sleep even while playing video.
No support for trusted devices feature. Each connection from the same mobile device initiates a new session, slightly breaking the flow.
ARM processors are not supported. Building from source doesn’t seem to help.

Still this app is really good, I enjoy using it and appreciate the author's effort.

By the way:

If you use Arch and prefer the .AppImage version of Deskreen instead of the AUR package, you might want to create a .desktop file to launch the app quickly from the application menu:

[Desktop Entry]
Name=Deskreen Screencaster
Exec=/path/to/Deskreen-2.0.4.AppImage
Icon=/path/to/deskreen.png
Terminal=false
Type=Application

Replace /path/to/ with the actual paths on your system.

The deskreen.png icon can be downloaded from the official site by right-clicking (Save As...) or directly from here:

Fix GNOME's Alt+Tab: Always Switch to the Last Used Window (Not All Windows of the App)

hopsayer — Fri, 04 Jul 2025 23:17:22 +0000

If you’re using GNOME (e.g. on Arch Linux) and notice that pressing Alt+Tab switches between applications—not individual windows—you’ve probably experienced this frustrating behavior:

You press Alt+Tab, trying to return to the last window you were using, but instead, GNOME shows all open windows of the application (e.g. all browser windows). This clutters the interface and breaks your flow — even though GNOME’s whole mission seems to be the exact opposite. Let's fix this.

The Problem

By default, GNOME groups windows by application. So if you have multiple windows of the same app, pressing Alt+Tab shows a grouped stack instead of just going to the last active window.

You can install gnome extensions like Coverflow Alt-Tab, but they often overcomplicate things with animations or heavy UI changes.

The Clean Solution

You can fix this behavior with built-in GNOME settings—no extensions required.

Here’s how:

I. Open Settings → Keyboard → View and Customize Shortcuts

II. Find these two entries:
• Switch Applications — this is the default Alt+Tab behavior that groups apps
• Switch Windows — this switches between all windows, regardless of app

III. Do the following:
• Disable “Switch Applications” or unbind its shortcut with the Backspace key.
• Set “Switch Windows” to Alt+Tab

Now, when you press Alt+Tab, it cycles through each window linearly, just like in older desktop environments (Xfce, Mate, etc.). No more app-grouped stacks. The last active window always comes up first.

A little side note:

If only two windows are fully visible (covering everything else), pressing Alt+Tab won’t show the switcher popup—it’ll just toggle focus between those two windows.

Why This Works

GNOME treats app switching and window switching as two separate concepts. The default is designed around grouping, which can be helpful in some workflows, but it often gets in the way when multitasking between many windows.

Rebinding Alt+Tab to “Switch Windows” brings back the classic, no-nonsense window switcher—exactly what most power users expect.

Bonus Tip

Still want to switch between windows within one app (e.g. different VS Code windows)? GNOME has a default shortcut for that:

• Alt+` (tilda key) = switch between windows of the same application

TL;DR

To stop GNOME from grouping windows when pressing Alt+Tab:

• Go to Settings → Keyboard → Shortcuts
• Disable “Switch Applications”
• Set Alt+Tab to “Switch Windows”

Done. Clean, minimal, and works exactly as expected.

Fixing ChatGPT "Lost First Key" Bug in Firefox

hopsayer — Fri, 04 Jul 2025 15:07:08 +0000

Problem

When using ChatGPT in Firefox, the text input field doesn’t properly register the first typed character. The cursor appears focused, but the first keypress is ignored, making it frustrating to start typing. In Chromium browsers, this issue doesn’t occur — the first keystroke goes directly into the input field as expected.

Solution

This happens because ChatGPT’s web app uses React and dynamic rendering, and Firefox handles DOM focus and event propagation slightly differently. React’s internal focus management can interfere, causing the first keystroke to be lost. A workaround is to programmatically focus the textarea when it appears and, if needed, simulate user interaction (like a click) to ensure the input field is truly active.

Instruction

You can create a userscript to fix this in Firefox using Tampermonkey or Violentmonkey. The script observes the page’s DOM for changes, and when the textarea appears, it forces focus on it. Additionally, it simulates a mouse click event in the center of the textarea to activate the input fully.

Here’s a userscript example:

// ==UserScript==
// @name         ChatGPT Firefox Input Focus Fix
// @namespace    http://tampermonkey.net/
// @version      1.0
// @description  Fixes lost first keystroke in ChatGPT input on Firefox by focusing and clicking the textarea
// @match        https://chat.openai.com/*
// @match        https://chatgpt.com/*
// @grant        none
// ==/UserScript==

(function () {
  const simulateClick = (el) => {
    const rect = el.getBoundingClientRect();
    const x = rect.left + rect.width / 2;
    const y = rect.top + rect.height / 2;
    ["mousedown", "mouseup", "click"].forEach(type => {
      el.dispatchEvent(new MouseEvent(type, {
        bubbles: true,
        cancelable: true,
        view: window,
        clientX: x,
        clientY: y,
      }));
    });
  };

  const tryFocusByClick = () => {
    const input = document.querySelector("textarea");
    if (input && document.activeElement !== input) {
      simulateClick(input);
    }
  };

  const observer = new MutationObserver(() => {
    tryFocusByClick();
  });

  observer.observe(document.body, {
    childList: true,
    subtree: true,
  });

  window.addEventListener("load", () => {
    setTimeout(tryFocusByClick, 150);
  });

  document.addEventListener("visibilitychange", tryFocusByClick);
})();

To use it:

• Install Tampermonkey or Violentmonkey in Firefox.
• Create a new userscript and paste the above code.
• Save and reload ChatGPT in Firefox.

Result

After applying this userscript, the textarea in ChatGPT on Firefox will receive proper focus, and the first keypress won’t be lost. You should be able to start typing immediately without needing to press Enter or click the field manually. This significantly improves the input experience on Firefox, bringing it closer to Chromium’s behavior.

Tested on Firefox 135

Removing Firefox context menu items (right-click menu)

hopsayer — Tue, 01 Jul 2025 21:52:35 +0000

I. Problem:
Remove or hide specific items from Firefox context menu (right-click menu), such as:

• "Take Screenshot"
• "Inspect"
• "Inspect Accessibility Properties"
• "View Page Source"
• "Select All"
• "Ask AI"
• "Print Selection"
• "Translate Selection"
• "View Selection Source".

II. Solution:

• Use Firefox's about:config settings to disable certain features entirely
or
• Use userChrome.css to hide individual menu items via CSS rules.

III. Instruction:

• Open Firefox and go to about:profiles.

• Find the user profile which is marked as "This is the profile in use and it cannot be deleted."
Click Open Directory next to Root Directory.

• In the profile folder, create a folder named chrome if it doesn't exist.

• Inside chrome, create or edit a file named userChrome.css.

• To disable context menu items you consider unnecessary, edit the userChrome.css file:

@namespace url("http://www.mozilla.org/keymaster/gatekeeper/there.is.only.xul");

/* Hide "Take Screenshot" */
#context-screenshot {
  display: none !important;
}

/* Hide "Inspect" */
#context-inspect {
  display: none !important;
}

/* Hide "Inspect Accessibility Properties" */
#context-inspect-a11y {
  display: none !important;
}

/* Hide "View Page Source" */
#context-viewsource {
  display: none !important;
}

/* Hide "Select All" */
#context-selectall {
  display: none !important;
}

/* Hide "Print Selection" */
#context-print-selection {
  display: none !important;
}

/* Hide "Translate Selection" */
#context-translate-selection {
  display: none !important;
}

/* Hide "View Selection Source" */
#context-viewpartialsource-selection {
  display: none !important;
}

With this done, you have hidden menu items by adding multiple custom CSS rules.

• In Firefox open new tab with the address about:config.

• Find the toolkit.legacyUserProfileCustomizations.stylesheets flag and set it to true.

• Restart Firefox completely.

IV. Result:

• After applying these settings and restarting Firefox, the specified context menu items will be removed or hidden as requested.

• Disabling features via about:config completely removes their functionality and menu items.

• Using userChrome.css allows selective hiding of menu items without disabling the underlying feature.

• This provides flexible control over Firefox's context menu appearance.

You may also like the second, more tough, way to exclude menu items by disabling entire features, which removes corresponding menu items automatically:

• To disable screenshots feature (removes "Take Screenshot"), change screenshots.browser.component.enabled to false in about:config.

• To disable "Ask ChatGPT" menu item, change browser.ml.chat.enabled and browser.ml.chat.shortcuts to false in about:config.

• To disable accessibility tools (removes "Inspect Accessibility Properties"), change devtools.accessibility.enabled to false.