🔄 Linux diff Command

# file1.txt contains:
apple
banana
cherry

# file2.txt contains:
apple
blueberry
cherry

# Compare current running config with saved config
ps aux | diff - saved_processes.txt

# Compare current directory listing with expected
ls -la | diff - expected_files.txt

# Example script usage
if diff -q config.txt config.backup > /dev/null; then
    echo "Config unchanged"
else
    echo "Config was modified!"
    diff -u config.txt config.backup
fi

# Install colordiff
sudo apt install colordiff      # Ubuntu/Debian
sudo dnf install colordiff      # RHEL/Rocky

# Use as drop-in replacement for diff
colordiff -u file1.txt file2.txt

# Exclude multiple patterns
diff -r --exclude="*.log" --exclude="*.tmp" --exclude=".git" dir1/ dir2/

# Exclude from file
diff -r --exclude-from=.diffignore dir1/ dir2/

# Compare current config with backup
sudo diff -u /etc/nginx/nginx.conf /etc/nginx/nginx.conf.backup

# Compare running config with new config before applying
diff -u /etc/mysql/my.cnf /tmp/new_my.cnf

# Show differences ignoring comments and blank lines
diff -u -B -I '^#' /etc/ssh/sshd_config /etc/ssh/sshd_config.default

📚 Additional Information

Understanding diff Output Formats

Normal Format (default):

2c2          # Line 2 changed
< banana     # Old content (from file1)
---          # Separator
> blueberry  # New content (from file2)

# Other codes:
a = added
d = deleted
c = changed

# Examples:
5a6    # After line 5, add line 6
3d2    # Delete line 3 (was line 2)
2,4c2  # Lines 2-4 changed to line 2

Unified Format (-u):

--- file1.txt  # Original file
+++ file2.txt  # Modified file
@@ -1,3 +1,3 @@  # Line numbers (old start,count new start,count)
 apple         # Unchanged (context)
-banana        # Removed from file1
+blueberry     # Added to file2
 cherry        # Unchanged (context)

Context Format (-c):

*** file1.txt  # Original
--- file2.txt  # Modified
***************
*** 1,3 ****    # Lines 1-3 in original
  apple        # Unchanged
! banana       # Changed (! marks changed lines)
  cherry       # Unchanged
--- 1,3 ----   # Lines 1-3 in modified
  apple
! blueberry
  cherry

Common diff Options

Option	Description
-u, --unified	Unified format (most common)
-c, --context	Context format
-y, --side-by-side	Side-by-side comparison
-q, --brief	Report only whether files differ
-r, --recursive	Recursively compare directories
-i, --ignore-case	Ignore case differences
-w, --ignore-all-space	Ignore all white space
-b, --ignore-space-change	Ignore changes in amount of white space
-B, --ignore-blank-lines	Ignore blank lines
-I RE, --ignore-matching-lines=RE	Ignore lines matching regex
--exclude=PAT	Exclude files matching pattern
-N, --new-file	Treat absent files as empty
-a, --text	Treat all files as text

Exit Codes

0 - Files are identical
1 - Files are different
2 - Error occurred

Related Commands

Command	Purpose
diff	Compare files line by line
diff3	Compare three files
sdiff	Side-by-side diff (merge interactively)
cmp	Compare files byte by byte
comm	Compare sorted files line by line
patch	Apply diff output to files
colordiff	Colorized diff output
vimdiff	Visual diff in vim
git diff	Git's diff implementation

Real-World Use Cases

System Administration:

# Compare config before and after changes
sudo cp /etc/nginx/nginx.conf /etc/nginx/nginx.conf.backup
# Make changes...
diff -u /etc/nginx/nginx.conf.backup /etc/nginx/nginx.conf

# Check what changed after system update
diff -r /etc /backup/etc-before-update | less

# Compare user lists
diff <(cat /etc/passwd | sort) <(ssh remote-host 'cat /etc/passwd | sort')

# Find differences in installed packages
diff <(dpkg -l | awk '{print $2}') <(ssh remote 'dpkg -l | awk "{print \$2}"')

Development:

# Create patch for code review
diff -u original.py modified.py > feature.patch

# Compare build outputs
diff -r build-old/ build-new/

# Check if code formatting changed
diff -w -B original.c formatted.c

# Compare API responses
curl -s https://api.example.com/v1 | diff - <(curl -s https://api.example.com/v2)

Backup Verification:

# Verify backup matches original
diff -qr /data /backup/data || echo "Backup differs!"

# Compare directory structures only (not content)
diff <(cd dir1 && find . -type f | sort) <(cd dir2 && find . -type f | sort)

# Check if backup is up to date
if diff -q important.db backup/important.db > /dev/null; then
    echo "Backup is current"
else
    echo "Backup needs updating"
    cp important.db backup/important.db
fi

Working with Patches

Creating Patches:

# Single file patch
diff -u original.txt modified.txt > changes.patch

# Directory patch
diff -urN original-dir/ modified-dir/ > changes.patch

# With better context
diff -U 10 original.c modified.c > changes.patch

Applying Patches:

# Apply patch
patch < changes.patch

# Apply to specific file
patch original.txt < changes.patch

# Dry run (test without applying)
patch --dry-run < changes.patch

# Reverse a patch
patch -R < changes.patch

Advanced Techniques

Process Substitution:

# Compare command outputs
diff <(ls dir1) <(ls dir2)

# Compare sorted files
diff <(sort file1) <(sort file2)

# Compare remote and local
diff local.txt <(ssh server cat remote.txt)

# Compare git branches
diff <(git show branch1:file.txt) <(git show branch2:file.txt)

Filtering diff Output:

# Show only added lines
diff -u file1 file2 | grep '^+'

# Show only removed lines
diff -u file1 file2 | grep '^-'

# Count changes
diff -u file1 file2 | grep '^[+-]' | wc -l

# Get just filenames that differ
diff -qr dir1 dir2 | grep differ | awk '{print $2}'

Ignoring Patterns:

# Create .diffignore file
*.log
*.tmp
.git/
node_modules/
__pycache__/

# Use it
diff -r --exclude-from=.diffignore dir1/ dir2/

# Or inline
diff -r --exclude='*.pyc' --exclude='.git' --exclude='node_modules' src1/ src2/

diff vs Other Tools

Tool	Best For	Comparison Type
diff	Text files, creating patches	Line by line
cmp	Binary files, quick checks	Byte by byte
comm	Sorted text files	Line by line (shows common/unique)
vimdiff	Interactive editing	Visual, side-by-side
git diff	Version control changes	Line by line with git context

Performance Tips

• Large files: Use -q first to check if files differ before full comparison
• Directories: Use --exclude to skip unnecessary files
• Binary files: Use -q or cmp instead of full diff
• Network files: Copy locally first for faster comparison

Common Gotchas

• Line ending differences: Windows (CRLF) vs Unix (LF) can cause every line to show as different
• Trailing whitespace: May cause differences even if content looks the same
• Tab vs spaces: Can make files appear different when they're functionally identical
• File encoding: UTF-8 vs ASCII vs other encodings can cause issues

Useful Aliases

# Add to ~/.bashrc or ~/.bash_aliases
alias diff='diff --color=auto'
alias diffs='diff -y --suppress-common-lines'
alias diffstat='diff -u | diffstat'

# Function for config diff
confdiff() {
    diff -u -B -I '^#' "$1" "$2"
}

Scripts Using diff

Backup Comparison Script:

#!/bin/bash
# Compare production with backup and alert if different

PROD_DIR="/var/www/html"
BACKUP_DIR="/backup/html"
ALERT_EMAIL="admin@example.com"

if ! diff -qr "$PROD_DIR" "$BACKUP_DIR" > /dev/null 2>&1; then
    diff -ur "$PROD_DIR" "$BACKUP_DIR" | mail -s "Backup Differs from Production" "$ALERT_EMAIL"
    echo "Alert sent: Backup differs from production"
else
    echo "Backup matches production"
fi

Config Change Detector:

#!/bin/bash
# Monitor config file for changes

CONFIG="/etc/nginx/nginx.conf"
SNAPSHOT="/var/cache/nginx.conf.snapshot"

if [ ! -f "$SNAPSHOT" ]; then
    cp "$CONFIG" "$SNAPSHOT"
    exit 0
fi

if ! diff -q "$CONFIG" "$SNAPSHOT" > /dev/null; then
    echo "CONFIG CHANGED:"
    diff -u "$SNAPSHOT" "$CONFIG"
    logger "nginx config was modified"
    cp "$CONFIG" "$SNAPSHOT"
fi

Fun Facts

• diff was created in the early 1970s at Bell Labs for Unix
• The algorithm used is called the "Hunt-McIlroy algorithm"
• git's diff is based on the same principles but optimized for version control
• The unified diff format (-u) was created in 1990 for patch compatibility
• diff is so fundamental that nearly every version control system uses its output format