Bash

Chaining commands

• a && b: do b if a succeeded
  • not to be confused with a & b
• a || b: do b if a failed
  • these can be chained: command && echo "success" || echo "fail"
• a; b: do b regardless of whether a succeeded

• <: use file on right as stdin for command on left
• |: pipe stdout of command on left to stdin of command on right
• >: write file on right with stdout of command on left
  • write to /dev/null to discard output
• >>: same as >, but appends
• &>: redirect stdout & stderr

Background commands

• command &: run command in the background

  • to avoid spamming the terminal, redirect output to /dev/null: command &>/dev/null &

• Ctrl+z: suspend (pause) the foreground command

  • fg: resume suspended command in foreground
  • bg: resume suspended command in background

• jobs: list all background and suspended commands

• kill %1: kill job #1

• disown: disconnect a background command, so it keeps running when the terminal is closed

Substitution

• $VARNAME: variable substitution

  • happens in double quotes or when unquoted, but not in single quotes

• $(command): command substitution (substitutes the result of the command)

• ${param}: parameter substitution (substitutes the parameter value)

• !!: repeat the last command entered (ex. sudo !!)

• !$: last argument of the last command (ex. in cat foo.txt, foo.txt)

• $?: exit code of the last command

  • 0 means success, non-zero means failure

Function parameters

• access function parameters using $1, $2, etc
• "$@": all parameters
  • use double quotes to avoid issues with spaces or wildcards
• $#: number of parameters given

Control flow & comparisons

• for loop over every file in working directory

for i in *; do echo "$i"; done

• if statement
  • the spaces around the condition are important!
  • 0 is true and 1 is false (this matches exit codes)

if [[ condition ]]; then
	echo "condition 1 passed"
elif [[ condition2 ]]; then
	echo "condition 2 passed"
else
	echo "both conditions failed"
fi

• [[ condition ]]: conditional test (the spaces are important)
  • 0 is true and 1 is false (matches exit codes)
  • =: string comparison
  • -eq: numeric comparison (also: -lt, -le, -gt, -ge, -ne)
  • -f: test if something is a file

Wildcards

• cp m*.txt scifi/
  • */ matches only folders
  • file ^*foo*: exclude files containing foo
    • requires setopt extendedglob
• ? acts as a single character wildcard
  • ex. ls *.?s would list all .js, .ts, etc. files
• ** lists files recursively, ex. ls **/*.json

Hotkeys

• Ctrl+r and Ctrl+s: search command history backwards/forwards
• Ctrl+a: move to beginning of line
• Ctrl+e: move to end of line
• Ctrl+w: delete word
• Ctrl+u: clear from start of prompt to cursor
• Ctrl+k: clear from cursor to end of prompt
• Ctrl+y: undo deleting text
• Ctrl+l: refresh screen

Fish

Basics

• set -x VARIABLE SomeValue: export a variable
  • set -U sets a universal variable (persists across sessions)
• alias --save name="definition": save alias
• >?: redirect output to file, but avoid overwriting existing file
• psub: lets you save stdout to a temporary file
  • ex: diff (sort a.txt | psub) (sort b.txt | psub)
• command substitution: (command) (no $)
• access arguments with $argv[1] instead of $1
• use $status instead of $? for last exit code
• if type -q command ...: check if command exists
• set_color: set the text color
  • supports 3 or 6 digit hex colors, or reserved color names
    • use -c to print reserved color names
    • normal will reset the color
  • -o, -i, -u: set bold/italic/underline
  • -b: set the background color

History recall

• Alt+Up Arrow: recall individual arguments
• Alt+s: recall last command with sudo prefixed

String matching

• string match 'word' 'string1' 'string2'...
  • -a: print all matches
  • -e: print entire string, not just match
  • -i: case insensitive
  • -n: report each match as a 1-based start position and length
  • -r: interpret as Perl-style regex
    • without -r or -e, the pattern must match the entire string
    • if the regex contains capturing groups, multiple items will be reported for each match, one for the entire match and one for each capturing group
    • automatically sets variables for all named capturing groups (?<name>expression)
      • if -a is used, each variable will contain a list of matches
  • -v: only print lines that do not match the pattern

Snippets

Get name of frontmost app (macOS)

set result (string match -r '"LSDisplayName"="(.+)"' (lsappinfo info -only name (lsappinfo front)))
echo $result[2]

Commands

Text search and manipulation

grep

Returns lines that match a word or pattern

• grep 'word' filename
  • filename can also be a directory or pattern, like . or *.js
  • grep -F (or fgrep) is faster but finds only fixed patterns (not regexp)
  • grep -E (or egrep) handles extended regular expressions
  • grep -P uses Perl-style regexps (allows things like look-around assertions)
• -i: case insensitive
• -w: whole word matches
• -v: invert match (match only lines that do not contain the word)
• -R: search directory recursively (ex. grep -R foo ~/bar)
  • -l: show matching filenames
  • -h: do not output file names

Output configuration

• display surrounding lines: -B (before) and -A (after)
• print only what matches pattern: -o
• count number of matches: -c
  • prepend line number: -n

awk

Searches files for pattern matches and performs an action based on them

• format: awk pattern { action }
• the entire program (pattern and action) should be quoted
• lines are split into chunks by whitespace, numbered $1, $2, etc.
  • $0 matches the whole line
  • pass -F regexp to split by regexp instead
• action examples
  • { print $1 }: print the first whitespace-separated chunk
  • length($0) > 72: print lines longer than 72 characters
  • { print $2, $1 }: print first two fields in opposite order
• if pattern is missing it will match every line, if action is missing it will print all matching lines

sed

• sed: modify stdin based on regex
  • ex. sed 's/foo/bar/' file.txt
  • end regex with /g for global search

wc (word count)

• wc: output number of lines, words, and characters (in that order)

cat/less

• cat: print file contents
• less: show file contents in a scrollable view

head/tail

• head -#: output first # lines
• tail -#: output last # lines

sort

• sort: sort alphabetically (doesn't modify original)
  • -r to reverse, -h for numeric sorting

uniq

• uniq: filter out duplicate lines if they are adjacent
  • combine sort and uniq to remove all duplicates

nano

• Ctrl+Shift+6: mark text
  • Ctrl+k: cut/delete marked text

Files & folders

• folder refers to a folder itself, folder/ refers to the contents of the folder
  • cp folder destination will copy folder itself to destination, but cp folder/ destination will copy the contents of folder

ls

• -1: show results in single column
• -a: list hidden files
• -t: sort by modified date

file

• file: show information about file

stat

• stat -f "%Xf" file.txt: if file is hidden, result will be 8000

find

• find dir -type f -name foo*: find all files starting with "foo" in dir (recursive)
  • use -type d for folders (directories)
• find dir -name .DS_Store -delete: delete all .DS_Store files in dir
• find . -name foo* -exec command: execute command on all matching files

fd

• faster and friendlier alternative to [[#find]]
• recursively searches the current directory by default
  • can pass a directory as the second argument
  • can give a regular expression as the pattern
• if called without an argument, lists all files in the current directory recursively
• -e EXT: match files with extension EXT
• -g: match the given pattern exactly
• -H: include hidden files
• -I: include gitignored files
• -E GLOB: exclude files or folders matching GLOB
  • create a .fdignore (like .gitignore) file to always ignore certain files/folders
  • create a global ignore file at ~/.config/fd/ignore
• -p: match against the full path (ex. fd -p 'foo/bar')
• -x COMMAND: run COMMAND against each result in parallel
  • ex. find and unpack all zip files: fd -e zip -x unzip
• -X COMMAND: run COMMAND once with all the results as arguments
  • -X rm: remove all files (not folders) that match
    • run without -X rm first to confirm your choices, or use rm -i

lsof (list open files)

• lsof +d dir: shows open files in dir
• lsof +D dir: shows open files in dir and its subfolders
• lsof -c Finder: list all files Finder has open

Find process using port

• lsof -i :80: find out which processes are using port 80

cp

• move/copy multiple files: cp a.txt b.txt c/

rsync

• uses modification times to sync only changed files
• rsync -av dir1 dir2: copy directory 1 to directory 2
  • -a (archive): sync recursively and preserve symlinks and metadata
  • -v: verbose
• rsync -azP dir1 username@host:/path/to/dir2: copy over SSH
  • -z: use compression
  • -P: show progress bar, and allow resuming transfers
• other options:
  • --exclude=pattern: exclude files matching pattern
  • -n: dry run

ln

• ln -s source target: create a symbolic link from source to target

Processes

ps

• ps -ef: list running processes
  • -e: include processes from other users
  • -f: show more information
• to filter the list while including headers:
  • bash: ps -ef | { head -1; grep Finder; }
  • fish: ps -ef | begin; head -1; grep Finder; end;

top

• get CPU usage of a process
  • -l 2: take 2 samples (CPU usage is based on the delta between samples)
  • for memory usage, replace cpu with mem and omit -l 2

top -stats "command,cpu" -l 2 | grep WindowServer | tail -1

pkill

• send a signal to kill all matching processes by name
• the default signal is -15 which tells the process to exit gracefully
  • send -9 to kill the process immediately
• the name uses regex matching, to match an exact name use ex. "^node$"
  • use the -f flag to match against the full argument list, ex. pkill -f "node index.js"
• use -n or -o to kill the newest or oldest matching process, respectively

pkill -9 "node"

Network

Simple web server

• if npm is installed:

npx serve@latest out -l 3000

• if Python 3 is installed:

python3 -m http.server 3000

ipconfig

Get list of network adapters

ipconfig getiflist

Get network adapter details

• replace en0 with the network adapter name

ipconfig getsummary en0

Get IP address

• replace en0 with the network adapter name

ipconfig getifaddr en0

curl

• -v: verbose mode, prints connection steps
• -X: set request method (GET, POST, etc)
• -d "body": set the request body
• -H "header": set a header
• -s: silent mode, disables progress bar and error messages
  • -sS: disable progress bar but keep error messages
• -L: follow redirects
• -I: fetch headers only

Send a POST request with a JSON body:

curl -X POST -H "Content-Type: application/json" -d '{"name": "Oscar", "species": "cat"}' https://example.com/api/pets

tldr

• pull up cheat sheets with examples for commands

xargs

• xargs command: run command for each line in stdin
  • -n 2: limit to the first 2 lines

tmux

• tmux: create a new session
• tmux attach: reattach to existing session

Keyboard shortcuts

• Ctrl+b ": split horizontally
• Ctrl+b %: split vertically
• Ctrl+b {arrow}: move focus
• Ctrl+b+{arrow}: resize focused pane
• Ctrl+b x: close pane (y to confirm)
• Ctrl+b d: detach from session

crontab

• crontab -l: print crontab
• crontab -e: edit crontab (auto reloads on exit)
  • format: minute hour dayOfMonth month dayOfWeek command
  • separate multiple times with commas, ex 1,2,3
  • use */# to specify an interval, ex. */20 * * * * command means "every 20 minutes" (without the */ this would mean "on minute 20 of every hour")
  • @reboot command to run on reboot
• using sudo edits a separate crontab for root

fzf

• lets you interactively filter data from stdin and output the selected item to stdout

SSH Agent

• ssh-keygen -l -f <path>: view key fingerprint
  • to view in MD5 format (colon-delimited) add -E MD5
• ssh-add <path>: add key to SSH agent
• ssh-add -l: list keys added to SSH agent
• ssh-add -D: remove all keys from SSH agent

node

• node --watch script.js: re-run script whenever it changes

npm/yarn

NPM and Yarn

ffmpeg

• -c: specify codec
  • use -codecs to get a list of available codecs
  • use -vcodec and -acodec to specify video/audio codecs separately
  • for webm use -vcodec libvpx -acodec libvorbis
  • -c copy avoids re-encoding (useful for trimming, changing metadata, etc)

Trim video

• ffmpeg -i input.mp4 -ss 00:01:40 -to 00:02:16 -c copy output.mp4: trim to between 1m:40s and 2m:16s

Convert video to audio

• ffmpeg -i input.m4a -c:v copy -c:a libmp3lame -q:a 4 output.mp3: convert m4a to mp3 without significant quality loss

Extract frames

• ffmpeg -i input.mp4 -ss 00:00:14.435 -vframes 1 out.png: extract one frame from 0h:0m:14sec:435msec to out.png
• ffmpeg -i input.mp4 frame_%04d.jpg: extract all frames to frame_0001.jpg, frame_0002.jpg, etc

youtube-dl/yt-dlp

• --simulate: dry run
• --match-title keyword: only download items matching keyword
• -F url: list all available formats
  • -f # url: download video using a format number given by -F
• --write-sub --sub-lang en: download subtitles
  • --write-auto-sub: download YouTube automatic subtitles
• --add-metadata: add metadata like chapters
• --embed-thumbnail: embed thumbnail into audio file
• --merge-output-format mp4: set the output format to mp4
• --batch-file='videos.txt': download all files in videos.txt (list of URLs, one per line)
• -f bestaudio[ext=m4a] --add-metadata --embed-thumbnail url: save audio with thumbnail and chapter markers
• -o "%(playlist_index)s-%(title)s.%(ext)s" <playlist_link>: save a playlist with numbered files
  • --playlist-start 2: start at video 2 in the playlist (1-indexed)
  • --playlist-end 5: end at video 5
  • --playlist-items 1,2,5-7: download certain items

ImageMagick

Input/Output

• To read from stdin or stdout, replace the input or output filename with -

  • Explicit file formats can be provided like this: png:-

• Quick Look result (fish)

qlmanage -c public.image -p (magick convert image.png ... - | psub)

• Select frames from a GIF

magick 'image.gif[0]' image.png
magick 'image.gif[0-3]' image-%d.png # outputs image-0.png thru image-3.png
magick 'image.gif[0,3]' image-%d.png # only frames 0 and 3

Resizing and scaling

• Resize image to fit in a specific size (keeps aspect ratio)
  • Percentages can also be used

magick convert image.png -resize 100x100 resized.png

• Stretch the image to the exact size: -resize 100x100\!
• Only shrink larger images: -resize 100x100\>
• Only enlarge smaller images: -resize 100x100\<
• Resize the smallest dimension to fit: -resize 100x100^
  • Above plus cropping (like image-fit: cover): -resize 100x100^ -gravity center -extent 100x100

• Other scaling algorithms (replace -resize with these):
  • -scale: nearest neighbor
  • -adaptive-resize: can reduce blurring
• To double the image size using Scale2X use -magnify (with no size argument)

jq

• pipe JSON (ex. from curl) into jq to pretty-print it
• jq -s '.[0] * .[1]' --indent 4 old.json patch.json > new.json: recursively merge patch.json into old.json
• live preview filters using [[#fzf]]

echo '' | fzf --preview 'jq {q} < filename.json'

Flags

• -f path: read filter from path
• --indent n: change output indentation
• -r: output strings without quotes

JSON for examples:

{
    "foo": {
        "bar": [
            { "name": "Apple", "color": "Red" },
            { "name": "Orange", "color": "Orange" },
            { "name": "Banana", "color": "Yellow" }
        ]
    }
}

List all values for key in array of objects

• .[] spreads an array

.foo.bar | .[].name

pandoc

pandoc file.docx -o file.html

macOS-specific

caffeinate

• caffeinate: prevent your Mac from sleeping until the process exits
• caffeinate -t N: prevent sleep for N seconds

chflags

Hide/unhide files & folders

See [[#stat]] to check if a file/folder is hidden

chflags hidden file.txt
chflags nohidden file.txt

defaults

Dim Dock icons of hidden apps

defaults write com.apple.Dock showhidden -boolean yes; killall Dock

tccutil

Reset app permissions

• Get the app's Bundle ID:

osascript -e 'id of app "Name of App"'

• Reset permissions
  • you can replace All with a category name, like Photos

sudo tccutil reset All bundle_id

Homebrew

• brew leaves: show packages that aren't dependencies of any other installed package
  • -r: only show leaves that were manually installed
  • -p: only show leaves that were installed as dependencies of another package
    • these can be safely uninstalled, since the package that depended on them is no longer installed
• brew cleanup -s: remove unused and outdated cache files

Moom

• adjust spacing for Stage Manager

defaults read com.manytricks.Moom "Grid Spacing: Apply To Edges: Gaps" {0,0,0,75}

PowerShell

• $profile: print path to PowerShell profile
  • edit profile with ex. code $profile
• explorer .: open Explorer window in current working directory
  • or ii .: short for Invoke-Item .
• taskkill /F /IM name: kill process matching name (wildcard * can be used)
• New-Item -ItemType SymbolicLink -Path "dest" -Target "src": create symbolic link to src at dest
  • must be elevated
  • -ItemType HardLink for hard links
• Get-Content urls.txt | ForEach-Object {Invoke-WebRequest $_ -OutFile $(Split-Path $_ -Leaf)}: download list of URLs
  • Split-Path $_ -Leaf gets the last part of the URL as the file name

Environment Variables

Add or append to environment variable

$Env:VariableName = 'value'
$Env:Path += ';C:\Tools'

Get environment variable

Get-Item Env:VariableName

Snippets

Delete hidden macOS files

$theSource = "E:\"       # <<<<< insert drive here
Get-Childitem $theSource -Include @("._*", ".DS_Store", ".fseventsd", ".TemporaryItems", ".Trashes", ".Spotlight-V100") -Recurse -Force -ErrorAction SilentlyContinue | Remove-Item -Force -Recurse