Line Break Remover on copy on Wayland

line-break-remover.md

PDF Line Break Remover on KDE Plasma 6 Wayland

A Linux/Wayland equivalent of the classic AutoHotkey trick for cleaning up copied text from PDF viewers. Removes mid-sentence line breaks introduced by PDF column layout, replacing them with spaces — while preserving true paragraph breaks.

Tested on: CachyOS Linux · KDE Plasma 6.6.3 · Wayland

---

How It Works

On Windows, AutoHotkey can intercept Ctrl+C globally and post-process the clipboard.

#Requires AutoHotkey v2.0
#SingleInstance Force

; Create window group
GroupAdd("GroupName", "ahk_class SUMATRA_PDF_FRAME")
GroupAdd("GroupName", "ahk_class PXE:{C5309AD3-73E4-4707-B1E1-2940D8AF3B9D}")
GroupAdd("GroupName", "ahk_class classFoxitPhantom")
GroupAdd("GroupName", "ahk_class classFoxitReader")

#HotIf WinActive("ahk_group GroupName")
^c:: {
    old := ClipboardAll()
    A_Clipboard := ""
    Send("^c")
    if !ClipWait(0.8) {
        A_Clipboard := old
    } else {
        tmp := RegExReplace(A_Clipboard, "(\S.*?)\R(.*?\S)", "$1 $2")
        ; Replace double spaces with single spaces until no more remain
        while InStr(tmp, "  ")
            tmp := StrReplace(tmp, "  ", " ")
        A_Clipboard := tmp
    }
    old := ""
    tmp := ""
}
#HotIf

Wayland intentionally disallows this for security reasons. The workaround is a separate shortcut (Ctrl+Alt+C) that:

1. Sends a real Ctrl+C to the focused window (via ydotool)
2. Polls until the clipboard actually changes (like AHK's ClipWait)
3. Applies a regex to join mid-sentence line breaks
4. Writes the cleaned text back to the clipboard

Because Ctrl+Alt+C is a conscious action, no window class filtering is needed — you simply use it only when you're in a PDF viewer.

---

Step 1 — Install Dependencies

sudo pacman -S ydotool wl-clipboard python

• ydotool — injects keystrokes via /dev/uinput, bypassing the compositor entirely. Works on any Wayland compositor.
• wl-clipboard — provides wl-paste and wl-copy for Wayland clipboard access.
• python — for the regex cleaning step.

---

Step 2 — Add Your User to the input Group

ydotool needs access to /dev/uinput. The package ships a udev rule that grants the input group this access.

sudo usermod -aG input $USER

Then reboot — a simple logout/login may not be sufficient because systemd can preserve the old user session without the new group.

After rebooting, verify:

groups | grep input

input should appear in the output.

---

Step 3 — Enable the ydotool Daemon

ydotool is a client that talks to ydotoold (the daemon). The package ships a systemd user service:

systemctl --user enable --now ydotool.service
systemctl --user status ydotool.service

The service should show Active: active (running). If it fails, check that the reboot in Step 2 was completed and the input group is active.

---

Step 4 — Create the Script

Save the following as ~/.local/bin/pdf-clean-copy.sh:

#!/bin/bash

sleep 0.8

# Save current clipboard to detect change
old=$(wl-paste --no-newline 2>/dev/null)

# Inject Ctrl+C into the focused window (evdev keycodes: 29=Ctrl, 46=C)
# --key-delay 50: 50ms between events prevents stuck modifier keys
ydotool key --key-delay 50 29:1 46:1 46:0 29:0


text="$new"
[ -z "$text" ] && exit 0

# Clean the text using Python regex
# Faithfully ports the original AHK regex: (\S.*?)\R(.*?\S) → $1 $2
# [^\r\n] is used intentionally — re.S/re.DOTALL must NOT be used here,
# as it would allow .* to span multiple lines and cause incorrect merges
cleaned=$(printf '%s' "$text" | python3 -c "
import sys, re
text = sys.stdin.read()
result = re.sub(r'(\S[^\r\n]*?)\r?\n([^\r\n]*?\S)', r'\1 \2', text)
result = re.sub(r' {2,}', ' ', result)
sys.stdout.write(result)
")

# Write back only if text actually changed
[ "$cleaned" != "$text" ] && printf '%s' "$cleaned" | wl-copy

Make it executable:

chmod +x ~/.local/bin/pdf-clean-copy.sh

---

Step 5 — Bind to Ctrl+Alt+C in KDE

1. Open System Settings → Keyboard → Shortcuts
2. Scroll to the bottom → Custom Shortcuts
3. Click Edit → New → Global Shortcut → Command/URL
4. Set:
   • Name: PDF Clean Copy
   • Trigger: Ctrl+Alt+C
   • Action: /home/yourusername/.local/bin/pdf-clean-copy.sh

KDE applies the shortcut immediately — no restart needed.

---

Usage

1. Open any PDF in your viewer (Okular, PDF-XChange via Wine, Evince, etc.)
2. Select text that has broken line endings
3. Press Ctrl+Alt+C instead of Ctrl+C
4. Paste anywhere — line breaks within sentences are replaced with spaces, paragraph breaks are preserved

---

Troubleshooting

ydotool.service fails to start

• Confirm input group is active: groups | grep input
• If missing, reboot (logout alone may not be enough)
• Check udev rule: cat /usr/lib/udev/rules.d/80-uinput.rules — should grant GROUP="input" on uinput

Line breaks not being removed

• The PDF viewer may be slow to populate the clipboard
• increase sleep timer

---

Notes

• Why not Ctrl+C directly? Wayland's security model prevents any application from intercepting or suppressing another app's keyboard shortcuts globally. A separate shortcut is the correct solution, not a workaround.
• Why not wl-paste --watch? Passive clipboard watching applies to all copies everywhere. A deliberate shortcut is cleaner and more predictable.
• Why not wtype? KDE Plasma does not expose the zwp_virtual_keyboard_v1 protocol by default. ydotool bypasses the compositor via the kernel uinput interface instead.