# Specifications

Given a number n, output an ASCII "meme arrow" (greater-than symbol, >) of size n.

n will always be a positive integer, greater than 0.

## Examples

n = 2

\
\
/
/


n = 5

\
\
\
\
\
/
/
/
/
/


# Sample code

Here is a sample program, written in Crystal, that returns the correct results. Run it as ./arrow 10.

arrow.cr:

def f(i)
i.times { |j|
j.times { print ' ' }
puts "\\"
}
i.times { |j|
(i-j-1).times { print ' ' }
puts '/'
}
end

f(ARGV.to_i)


# Rules

• This is . The shortest answer wins. However, I will not select an answer, because the shortest answer may change over time.
• Standard loopholes are not allowed.
• Are you sure this is not a duplicate (that is, did you sandbox this)? I think this is likely to be, but it's slightly nearly impossible to search for. – my pronoun is monicareinstate Jun 9 '19 at 4:46
• Can n be zero? – xnor Jun 9 '19 at 6:05
• I think it's pretty silly to call this common symbol a "meme arrow". They're obviously comedy chevrons. – scatter Jun 10 '19 at 18:04
• @Christian They're actually amusing angles – dkudriavtsev Jun 10 '19 at 20:39
• @ArtemisFowl I thought they were interesting increases – dkudriavtsev Jun 12 '19 at 22:17

# Canvas, 2 bytes

＼═


Try it here!

# C (gcc), 56 bytes

f(n,i){for(i=-n;n;printf("%*c\n",i?++i+n:n--,i?92:47));}


Try it online!

f(n,i){for(i=-n;i;printf("%*c\n",  ++i+n    ,  92   ));     //first print descending '\'s
for(    ;n;printf("%*c\n",        n--,     47));}    // then print returning  '/'s


# Python 2, 54 bytes

f=lambda n,p='':n*'?'and p+'\\\n'+f(n-1,p+' ')+p+'/\n'


Try it online!

Outputs with a trailing newline.

# 05AB1E, 6 bytes

'\3.Λ∊


Try it online!

Explanation

   .Λ    # draw
'\       # the string "\"
# of length input
3      # in the south-eastern direction
∊   # then vertically mirror it


# C# (Visual C# Interactive Compiler), 66 bytes

n=>new int[n*2].Select((a,b)=>"".PadLeft(b<n?b:n*2+~b)+"\\/"[b/n])


Saved a byte thanks to @someone.

Try it online!

# C64Mini/C64 BASIC (and other CBM BASIC variants), 52 tokenized BASIC bytes used

 0INPUTN:N=N-1:FORI=0TON:PRINTTAB(I)"\":NEXT:FORI=NTO0STEP-1:PRINTTAB(I)"/":NEXT


Here is the non-obfuscated version for exaplantion:

 0 INPUT N
1 LET N=N-1
2 FOR I=0 TO N
3  PRINT TAB(I);"\"
4 NEXT I
5 FOR I=N TO 0 STEP -1
6  PRINT TAB(I);"/"
7 NEXT I


What ever number is entered into N in line zero is reduced by one as the TAB command is zero-indexed; The FOR/NEXT loops in lines two through to four and five through to seven then output the upper and lower part if the meme arrow respectively (represented by a shifted M and shifted N in graphics mode source) • Did you know that in Commodore Basic all keywords can be abbreviated? Here's a link: c64-wiki.com/wiki/BASIC_keyword_abbreviation For example, for can be fO (f - shoft o), print is ?, etc. – gaborsch Jun 15 '19 at 17:09
• 52 bytes is misleading in Code Golf, the binaries don't count, just the source code. It shoulf be something ike this: 0inputn:n=n-1:fOi=0ton:?tA(i)"\":nE:fOi=0ton:?tA(i)"/":nE - it takes 57 bytes. – gaborsch Jun 15 '19 at 17:10
• As discussed here -> codegolf.meta.stackexchange.com/questions/11553/… I count the tokens used as this is more representative of how much of the memory is being used. – Shaun Bebbers Jun 17 '19 at 8:39
• Oh, I didn't know this. Is there a decision about this? Even the answer has not been accepted there. – gaborsch Jun 17 '19 at 8:45
• My first computer was a C16, I did a lot of assembly on that, too, so no offence, I love Commodore. C16 had Basic 3,5, 80 chars per line, I also had a book with the ROM listings explained, so I exactly knew how the tokenization and listing worked, – gaborsch Jun 17 '19 at 9:41

# MarioLANG, 719 677 bytes

+
+
+
+
+
+         ((((+)+++++)))<
+>======================"
+)++++++)+++++++++++((-[!)
========================#=-
) ![-    <+;)+++)---)++++)<
)=#======"=================
) >)+)+((!
+ "======#
<))))).(((((((<
========">============"
>)+)+((-[!+))        -[!((((((((.)).))))+(-[!)
"========#=============#====================#<
!)                                          <
#==========================================="
>(((((.)))>
"========<"========
![-)).))).(((((((![-    ))+![-((+)+)<((![<
#================#=========#========"==#="===
>                                   !  >-!
"===================================#  "=#


Try it online!

This was harder than expected...

# brainfuck, 125 bytes

++++++++++[->+>+++++++++>+++<<<]>>++>++>,[->[->+<<<.>>]<<<.<.>>>>>[-<+>]<+<]<<[--<<+>>]<<+>>>>>[-[-<+<.>>]<<<<<.>.>>>[->+<]>]


Try it online!

++++++++++[->+>+++++++++>+++<<<]>>++>++>    ; Initialize with " \"
,                                           ; Get input
[->                                         ; loop and decrement n
[->+<<<.>>]                             ; output number of spaces, copy n
<<<.                                    ; output \
<.                                      ; output newline
>>>>
>[-<+>]<+                               ; copy copy of n back to original place
<]
<<[--<<+>>]<<+>>                            ; change "\" to "/"
>>>
[                                           ; second loop for bottom half
-                                          ; decrement n
[-<+<.>>]                                  ; output n spaces
<<<<<.>.                                   ; output \ and newline
>>>[->+<]>                                 ; copy n back
]


# Ruby, 111997773686457 56 bytes

-12 bytes thanks to Benjamin Urquhart, -43 thanks to manatwork and -2 bytes thanks to Value Ink.

->i{s=[];puts (0...i).map{|j|s=(p=' '*j)+?/,*s;p+?\\},s}


Try it online!

Explanation:

f=->i{                      # instead of a function, use a lambda
s=[]                      # needs a helper variable *now*, for scope
puts(                     # puts takes arbitrary num of args; \n after each
(0...i).map{|j|         # not from 0 to i but from 0 to i-1 (*three* dots)
s=(
p=' '*j             # p will remain in scope inside of .map,
)
+?/                   # character literal instead of string
,*s                   # essentially appending to the array

p+?\\                 # p is what's returned by .map, not s!

},                      # up until here, 1st arg to display
s                       # NOW, as the *2nd* arg, s is displayed
)
}


# Alternative (but longer) Solutions

A friend read this answer and then tried to come up with a couple more approaches. Putting them here, too, so that they're not lost to the vast interwebs.

## inject and unshift, 72 bytes

->n{puts (0...n).inject([]){|s,i|i=' '*(n-1-i);s.unshift i+?\\;s<<i+?/}}


Try it online!

## downto, inject and unshift, 80 bytes

->n{puts n.downto(1).map{|i|' '*(i-1)}.inject([]){|s,i|s<<i+?/;s.unshift i+?\\}}


Try it online!

## intriguing, two non-nested loops, 127 bytes

->n{
r=->s,c{s[0..-(c+1)],s[-c..-1]=s[c..-1],s[0..c-1];s};
n.times{|i|puts r[' '*n+?\\,n-i]}
n.times{|i|puts r[' '*n+?/,i+1]}
}


Try it online!

# Perl 5-p, 36 bytes

$\=($q=$"x$_)."\\
$\$q/
"while$_--}{  Try it online! # T-SQL code, 80 bytes DECLARE @ INT=3 ,@z INT=0 x:PRINT space(@-abs(@-@z-.5))+char(92-@z/@*45)SET @z+=1IF @z<@*2GOTO x  Try it online # T-SQL query, 96 bytes In order to make this work online i had to make some minor alterations. Spaces in the beginning of a row doesn't display in the online snippet. So I am using ascii 160 instead. When running in management studio, it is possible to change the settings to show result as text, which would result in the correct spaces in this posted script. DECLARE @ INT=3 SELECT space(@-abs(@-number-.5))+char(92-number/@*45)FROM spt_values WHERE number<@*2and'p'=type  Try it online # PowerShell, 44 41 bytes filter f{if($_){'\'
--$_|f|%{"$_"}
'/'}}


Try it online!

# Charcoal, 5 bytes

↘Ｎ‖Ｍ↓


Try it online! Link is to verbose version of code. Explanation:

↘Ｎ


Input a number and print a diagonal line of \s of that length.

‖Ｍ↓


Reflect the line vertically.

# APL(NARS), 40 chars, 80 bytes

{f←{⍺,⍨⍵⍴' '}⋄⊃('\'f¨k),('/'f¨⌽k←¯1+⍳⍵)}


test:

  h←{f←{⍺,⍨⍵⍴' '}⋄⊃('\'f¨k),('/'f¨⌽k←¯1+⍳⍵)}
h 2
\
\
/
/
h 5
\
\
\
\
\
/
/
/
/
/


# Retina 0.8.2, 32 bytes

.+
$* ¶$&$* \G. ¶$\
r.\G
$'/¶  Try it online! Explanation: .+$* ¶$&$*


Generate two lines of n spaces.

\G.
¶$\  Turn the top line into a \ diagonal. r.\G$'/¶


Turn the bottom line into a / diagonal.

# C (gcc), 67 65 bytes

-2 bytes thanks to ceilingcat

f(n,i){for(i=~n;i++<n;)i&&printf("%*c\n",n-abs(i)+1,"/\\"[i<0]);}


Try it online!

# Python 2, 85848180 75 bytes

def a(n):l="for i in range(n):print' '*";exec l+"i+'\\\\'\n"+l+"(n+~i)+'/'"


Try it online!

param($n)0..--$n|%{' '*$_+'\'}$n..0|%{' '*$_+'/'}  Try it online! Will look into making it so it only goes through the range once. Not bad for the no brain method though. # Twig, 115 bytes Builds the string backwards, "returning" it in the end. Uses a macro to generate all the results. {%macro a(N,s="")%}{%for i in N..1%}{%set s=('%'~i~'s '~s~'%'~i~'s ')|format('\\','/')%}{%endfor%}{{s}}{%endmacro%}  This macro has to be in a file, and imported like this: {% import 'macro.twig' as a %} {{ a.a(<value>) }}  You can try it on https://twigfiddle.com/5hzlpz (click on "Show raw result"). # Haskell, 52 49 bytes -3 bytes thanks to Sriotchilism O'Zaic. unlines.g g 0=[] g n="\\":map(' ':)(g$n-1)++["/"]


Try it online!

• You can save a byte with g$n-1 instead of g(n-1). You also don't need to count the f= since f never gets referenced. – Ad Hoc Garf Hunter Jun 10 '19 at 15:53 # MATL, 1413 12 bytes Xy92*t45-Pvc  1 Byte saved thanks to @LuisMendo Explanation  % Implicitly grab the input as an integer Xy % Create an identity matrix this size 92* % Multiply by 92 (ASCII for '\') t % Duplicate the matrix 45- % Subtract 45 from every element yielding 47 (ASCII for '/') on the diagonal % and -45 everywhere else P % Vertically flip this matrix v % Vertically concatenate the two matrices c % Convert to characters (negative numbers are replaced with a space) % Implicitly display the result  Try it out at MATL Online • @LuisMendo Updated! Thanks! – Suever Jun 11 '19 at 13:39 # Python 3, 90 83 bytes lambda n:'\n'.join([' '*i+'\\'for i in range(n)]+[' '*(n+~i)+'/'for i in range(n)])  Try it online! -7 bytes thanks to @squid • 83 bytes, but I feel like it can still go down – squid Jun 13 '19 at 12:55 • I wish this was possible. – Artemis still doesn't trust SE Jun 13 '19 at 14:45 • Soon... – squid Jun 13 '19 at 14:50 • Oh yeah I forgot about that. Maybe you should submit it! – Artemis still doesn't trust SE Jun 13 '19 at 14:54 # Rockstar, 133 bytes Try it online here! F takes N,S If N is 0 Give back N Say S+"\" Let T be S+" " Let M be N-1 F taking M,T Say S+"/" Listen to X F taking X,""  Since Rockstar is not famous for string operations, it takes relatively lots of code to do it (recursively was even longer). The size of the arrow is taken as input. # PHP, 7963 61 bytes function f($x,$s=''){if($x)echo"$s\\ ",f($x-1,"$s "),"$s/
";}


Try it online!

Recursive in PHP.

### -12 bytes by @manatwork

• Better count down. Try it online!. – manatwork Jun 17 '19 at 19:05
• @manatwork very very nice! – 640KB Jun 17 '19 at 19:11

# \/\/>, 74 bytes

jp100o
-84*}!o:?!x1
@+:q:p=?x:o~$:0(pa"\/"q?$~}}:
x2-:p$1-y$:0(?
.{suh?!;2


Explanation: (lines rotated based on start point)

jp100o                        //setup
~$:0(pa"\/"q?$~}}:@+:q:p=?x:o //add correct slash, go back to loop or switch sides
$:0(?x2-:p$1-y                //flip direction state or continue to print
{suh?!;2.                     //remove extra data and print stack

• \/\/> (pronounced wɜrm) Thanks, I hate it. (jk, I'm looking forward to giving it a try) – Jo King Jul 5 '19 at 6:39
• @JoKing hahaha, gotta wear my inspiration on my sleeve. (thanks!) – torcado Jul 5 '19 at 6:42

# Python 3, 55 bytes

f=lambda n,i="":n and i+"\\\n"+f(n-1,i+" ")+i+"/\n"or""


Try it online!

• – Chas Brown Oct 15 '19 at 0:27

# Erlang (escript), 69 bytes

r(0,_)->"";r(I,X)->X++"\\\n"++r(I-1,X++" ")++X++"/\n".
r(I)->r(I,"").


Try it online!

# Python 2, 63 bytes

f=lambda n,s='':n and f(n-1,~-n*' '+'\\\n'+s+~-n*' '+'/\n')or s


Try it online!

# Stax, 10 bytes

Ç₧¥╗M'gQ9⌂


Run and debug it

# SimpleTemplate, 100 bytes

This was quite a fun challenge, but some bugs in the language made it hard to optimize.

{@set_ argv.0}{@while_}{@callstr_repeat intoS" ",_}{@setO S,"\\\n",O,S,"/\n"}{@incby-1_}{@/}{@echoO}


Basically, cycles throught the values backwards, working the string from the middle out.

{@forfrom argv.0to0step-1}{@callrepeat intoS" ",_}{@setO S,"\\