zsh-cheat-sheet

= Resources =

* zsh         - http://www.zsh.org/
* ZSH-LOVERS  - http://grml.org/zsh/zsh-lovers.html
* manual      - http://zsh.sourceforge.net/Doc/Release/index.html
* oh-my-zsh   - https://github.com/robbyrussell/oh-my-zsh
* prezo       - https://github.com/sorin-ionescu/prezto
* zsh-users   - https://github.com/zsh-users

= Functions =

% print -l $fpath    # the path of the function

= Expansion =

You can always have a look at all supported types of expansions by typing man zshexpn on your console

== Parameter expansion ==

% foo=Hello
% echo "${foo}, world!"
> Hello, world!

== Command substitution ==

% print $(which zsh)
> /usr/local/bin/zsh

more portable

% print `which zsh`
> /usr/local/bin/zsh

== Arithmetic expansion ==

% echo $(( 5 + 4 ))
> 9

% echo $(( (5 + 4) * 3 ))
> 27

% num=5+4
% echo $(( num * 3 ))
> 27

== Brace expansion  ==

% echo picture.jp{eg,g}
> picture.jpeg picture.jpg

% touch log_00{1,2,3}.txt
% ls
> log_001.txt  log_002.txt  log_003.txt

% touch log_{007..011}.nfo
% ls | grep .nfo
> log_007.nfo  log_008.nfo log_009.nfo log_010.nfo  log_011.nfo

# ^ - array expansion operator (man zshoptions - the RC_EXPAND_PARAM section and man zshexpn)
% foo=(A B C)
% bar=(1 2 3)
% echo ${^foo}-${^bar}
> A-1 A-2 A-3 B-1 B-2 B-3 C-1 C-2 C-3


= history =

! - default character to trigger history expansion

setopt NO_HIST_VERIFY
setopt HIST_VERIFY   # force zsh into asking for confirmation every time you bang a command

Word designators
  ^: The first argument.
  $: The last argument.
  %: The most recent match for a given word.
  x-y: A range of words. Negative indexes like -i mean 0-i; thus, -1 would mean "the penultimate entry".
  *: All the arguments. Return null for events with just one word.

== history expansion ==

% ls *.txt
> readme.txt notes.txt
% !!
% ls *.txt
> readme.txt notes.txt

% sh myscript.sh
> myscript.sh: Error: you need to be root to execute this.
% sudo !!
> myscript.sh: executing myscript.sh

% touch log1.txt log2.txt
% vim !^
% vim log1.txt

% touch log1.txt log2.txt
% vim !$
% vim log2.txt

% history | grep ls
> 6596  ls log_008.nfo
> 6597  ls log_007.nfo log_008.nfo
% more !6597$
% more log_008.nfo

== history substitution ==

% ls
> dir1  file.txt
% mv fiel.txt dir1/
mv: rename fiel.txt to dir1/fiel.txt: No such file or directory
% ^fiel^file
% mv file.txt dir1/

= zle =

% bindkey -e                       # switch to emacs modes 
% bindkey -v                       # switch to vi mode
% bindkey -N newmap                # this creates a keybind named 'newmap'
% bindkey -N mycoolmap emacs       # this creates a new keymap based off the existing 'emacs'
% bindkey -A mycoolmap mymacs      # this creates an alias 'mymacs' for 'mycoolmap'
% bindkey -l                       # currently available keymaps
% bindkey -lL                      # format the output as a series of the bindkey commands
% bindkey -L                       # a list of all your current bindings, including those of a built-in keymap, formatted in a way you can use within your scripts

% read -k                          # usefull to figure out the actual escape sequence. For backspace: 
> ^?%

== vi mode ==
<Esc>          - vi mode
<Esc> + :      - run execute prompt, <Ctrl> + g to exit

% hello
execute: ca        # <Tab> key for autocomplete or <Ctrl> + g to exit 
% Hello

There is a possibility to check whether or not command is bind to a key sequence by putting where-is in execute and
then write a commnad after where-is prompt

% hello
execute: where-is
Where is: capitalize-word_
capitalize-word is not bound to any key

== zle widgets ==

zshzle(1) manpage's STANDARD WIDGETS section

Special variables :

CURSOR                : current position of the cursor on the command line.
BUFFER                : current editing buffer and can span multiple lines.
LBUFFER/RBUFFER       : contents to the left and right of the current cursor, respectively. They too can span multiple lines.
PREBUFFER             : buffer already read when editing a continuation line.
WIDGET                : name of the widget currently in use by the editor.

Loading tetris widget :)

autoload -Uz tetris
zle -N tetris
bindkey '\et' tetris    # <Esc> + t

=== zle widget example ===

function rationalize-dot {
  if [[ $LBUFFER = *.. ]]; then
    LBUFFER+=/..
  else
    LBUFFER+=.
  fi
}
zle -N rationalize-dot
bindkey . rationalize-dot

= globbing =

setopt NULL_GLOB         # discard any pattern without a proper match
setopt NO_NULL_GLOB
setopt NOMATCH
setopt NO_NOMATCH        # any pattern that does not match is passed as a literal argument to the command (like in Bash)
setopt CSH_NULL_GLOB     # mimics the legacy behavior of csh
setopt NO_CSH_NULL_GLOB

% echo *.zip
zsh: no matches found: *.zip

% setopt NULL_GLOB 
% echo *.zip
>

% setopt NO_NOMATCH 
% echo *.zip
> *.zip

% echo .*zsh*
> .zsh_aliases .zsh_funcs .zsh_history .zsh_prompt .zshenv .zshrc

% echo *.??
> script.sh

% ls
> Log.log Main.rb README.md script.sh
% echo [ML]*                  #  any filename that starts with either an uppercase letter M or L
> Log.log Main.rb

% echo *.log_[1-9]
> out.log_1 out.log_2 out.log_3

% echo [1-5M]* # any filename starting with any number from one to five or an uppercase M
> Main.rb

% ls
> bindings.c  bindings.h  bindings.o  main.c  main.o
% echo *.[^o]         # negates character set
> bindings.c bindings.h main.c

== char classes ==

ascii - Anything from the ASCII character set (see man ascii)
lower - Lowercase character
upper - Uppercase character
alpha - Letter
digit - Number
alnum - Alphanumeric character
print - Any printable character
blank - Space or tab character
space - Space character (tab, carriage return, newline and co.)
punct - Anything but an alnum nor a space 

echo [[:alpha:]]*
> awesome

% echo [[:digit:]d]* # all the files that start with either a digit character or the lowercase d letter
> 1.txt deploy

== extending globbing ==

setopt EXTENDED_GLOB         # enable extending globbing
setopt NO_EXTENDED_GLOB

=== recursive searching ===

**/   - recursive search
***/  - follow symbolic links

% echo **/*.md               # perform a recursive search
README.md brew/README.md git/README.md scripts/README.md zsh/README.md

% echo */**/*.md   # without current directory
brew/README.md git/README.md scripts/README.md zsh/README.md

=== alternate patterns ===

% echo [[:upper:]]*.(md|txt)
> README.md README.txt

=== numeric ranges ===

<-> - number special pattern

setopt NUMERIC_GLOB_SORT     #  output a sorted numeric match of any pattern matches 
setopt NO_NUMERIC_GLOB_SORT

% ls
> log.txt   log_002.txt   log_010.txt   log_031.txt   log_001.txt   log_009.txt   log_030.txt

% echo log_<->.txt
> log_001.txt log_002.txt log_009.txt log_010.txt log_030.txt log_031.txt

% echo log_<10->.txt           # grater then 10
> log_010.txt log_030.txt log_031.txt

% echo log_<10-20>.txt         # numbers between 10 and 20
> log_010.txt

===  caret operator ===

pattern^other_pattern  - match pattern and avoid match other_pattern

% ls
> README.md  README.txt  bindings.c  bindings.h  bindings.o  main.c  main.o

% echo b^*.o
> bindings.c bindings.h

===  tilde operator ===

% ls
> README.md  README.txt  bindings.c  bindings.h  bindings.o  main.c  main.o
% echo b*~*.o   # match b* and avoid matching *.o
> bindings.c bindings.h

% ls tmp
> delete_me.sh  out.txt
% echo **/*.sh~tmp/*     # match **/*.sh and avoid matching tmp/*
> src/script.sh

=== glob qualifiers ===

(/) - matching directories
(.) - matching files

% echo *zsh 
> xx.zsh zsh

% echo *zsh(/)
> zsh

% echo *zsh(.)
> xx.zsh

common qualifiers:

(N)         : Remove argument if no matches are found, silently ignore errors. Acts as a per-command NO_GLOB option.
(@)         : Symlink qualifier. Used for only selecting symbolic links.
(-@)        : A special variation of the previous one. Use this to find any broken symlinks.
(/)         : Directories only.
(.)         : Files only. Whatever is not either a link, directory, or any of the previous will be selected by this.
(*)         : Executable files. Directories need not apply. Think of this as (.) for those files with +x permissions.
(r)         : File is readable by the current shell user.
(w)         : File is writable by the current shell user.
(x)         : File is executable by the current shell user.
(U)         : File is owned by the current shell user.
(R)         : File is readable by anyone.
(W)         : File is writable by anyone.
(X)         : File is executable by anyone.
(u:root:)   : File is owned by the user root. You can replace the : character with any another pair of symbols such as curly braces: (u{root}). Just refrain from using pipes (|).
(on)        : Sort filenames by name. The echo *(on) construct will be analogous to ls.
(On)        : Reverse-sort filenames by name.
(oL)        : Sort filenames by file size.
(OL)        : Reverse-sort filenames by file size.
(om)        : Sort filenames by modification date.
(Om)        : Reverse-sort filenames by modification date.

% echo *(*r^w)    # print regular files that are readable and executable but not writable by your user
% echo *(@,/)     # print either symlinks or directories

==== timestamp qualifiers ====

m    -  modification
a    -  access
c    -  creation

m        - minute
h        - hour
[empty]  - day
w        - week
M        - month

-   : (now - x, now)        x units of time before now
+   : (inf, now - x)        more than x units time from now   

% echo *(mh-1)     # print files modified in the last hour
% echo *(m-1)      # print files modified in the last day
% echo *(ah-1)     # print files accessed in the last hour
% echo *(ch-1)     # print files created in the last hour

% echo *(mw+3)     # print files modified more than three weeks from today
% echo *(m-5mh+2)  # print files modified between five days and two hours (now - 5d, now - 2h)

==== file size qualifiers ====

(Lm+size)    : The file size is larger than size megabytes. 
(Lm-size)    : The file is smaller than size megabytes. 
(Lk+size)    : The file size is larger than size kilobytes. 
(Lk-size)    : The file is smaller than size kilobytes. 

% echo *(Lm+5)     # larger than five megabytes.
% echo *(Lm-2)     # smaller than two megabytes.
% echo *(Lk+5000)  # larger than 5000 kilobytes.
% echo *(Lm-2000)  # smaller than 2000 kilobytes.

=== zmv function ===

autoload zmv

-f     : Force overwriting of destination files
-i     : Interactive prompt for each operation
-n     : No execution, just print what happens
-v     : Verbose—print a line as it is executed
-w     : Implicitly add parenthesis to wildcards in the pattern
-W     : Like -w, but turn wildcards in replacement patterns into references

= misc =

#returns n th “word” of that output
echo "year =" ${$(date)[4]}

This is great!! Thank you so much.