# Simple ASCII Gantt

This is a simple one: print an ASCII Gantt chart.

Given tasks' ranges (start-time - end-time Tuples), print a Gantt timeline in the form of - characters for each task duration - each task in a new line.

### Example

Say my tasks ranges are 28->35, 34->40, 39->44, the Gantt will look like this:

-------
------
-----

### Specifications

• You can write a full program, a named function or an anonymous function.
• Your program/function should accept the tasks via STDIN or as arguments.
• Each task should be represented as a string of start->end where start and end are Integers. Tasks are separated by spaces or commas. Alternatively, you may get it as a Tuple of Integers, or as an Array/Collection of 2 Integers. (For example, in JavaScript you can get it as [start,end] - this is allowed).
• Any non-negative number of tasks (arguments) should be supported.
• To make it clear, a single argument of tasks collection is not allowed. You can either parse a single string argument, or support zero-or-more tasks arguments. Where task is a tuple or a collection of size 2.
• You can assume only valid input will be given. That means, each task has a positive duration.
• Return value does not matter, your code must print the timeline on STDOUT.
• Output: per task, start spaces followed by (end-start) dashes and a \n.
• Needless to say, output lines should be ordered correspondingly with the input (tasks) order.
• Trailing spaces before the \n are allowed, if that helps you.

### Test cases

Input:
(empty)

Output:
(empty)

Input:
0->7,5->6,3->6

Output:
-------
-
---

Input:
5->20,5->20,2->10,15->19

Output:
---------------
---------------
--------
----

### Winning

• This is so the least code length (in bytes) wins.
• Traditionally, tie breaker is earlier post.
• "Standard loopholes are no longer funny".

## EDIT

As many of you understood that it is allowed to have a single tasks collection argument, and since there's no much different between that and the original varargs requirement, it is now allowed to have a single collection argument, if you don't want to use the varargs option, or in case your language does not support varargs.

• Point 3 seems clear. But piint 5 (To make it clear...) is not clear at all. – edc65 Jun 29 '15 at 14:37
• Alright, let me rephrase that: You cannot write a function that accepts exactly one argument unless it's a string. If it's a bunch of tuples we're talking about, they may be sent to your function as arguments, not wrapped in a collection. For example, in JavaScript: You may iterate arguments within the function, but you may not assume that arguments[0] is an array of tasks. – Jacob Jun 29 '15 at 14:43
• Why not simply allow input as an array / list / vector / etc. for all languages? Personal preference seems like a pretty weak reason. – Doorknob Jun 29 '15 at 14:58
• Varargs versus an argument list is a purely syntactic distinction and leaves the rules of this question up to an unnecessary and arbitrary degree of interpretation, in my opinion. – JohnE Jun 29 '15 at 15:03
• @Jacob Makes sense. For future challenges, I'd recommend as lax an input spec as possible: Mangling input shouldn't be part of the challenge. – Adám Jan 29 '18 at 12:28

# CJam, 16 14 bytes

q~{S.*~'-e]N}/

This expects a list of lists as input. For example:

[[5 20] [5 20] [2 10] [5 19]]

gives:

---------------
---------------
--------
--------------

How it works

q~                      e# Read the input and parse it as a list of list
{         }/          e# Go over each item in the list in a for loop
S                    e# S is basically this string - " "
.*                  e# Multiply each item of the first list with the corresponding index
e# item of the second list. This basically repeats the space
e# X times where X is the first number of the tuple. The second
e# number remains untouched as the second list was only 1 char long
~                 e# Unwrap the space string and second number containing list
'-               e# Put character '-' on stack
e]             e# Make sure that the space is filled with - to its right so as to
e# make the total length of the string equal to the second number
N            e# Put a newline. After all iterations, the result is printed
e# automatically to STDOUT

Try it online here

# Python 2, 39 Bytes

Straightforward solution using string multiplication :)

for x,y in input():print' '*x+'-'*(y-x)

Accepts input formatted like so:

((5,20),(5,20),(2,10),(15,19))

Check it out here.

## Brainfuck, 120115 111 bytes

At least it's shorter than Java :) The input is a list of bytes, where each pair is a single line in the gantt.

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

Try out

http://copy.sh/brainfuck/

Set end-of-input to char with value \0. Example input: \5\20\5\20\2\10\15\19.

Note that setting the end-of-input value to \0 will have the side effect that no more input will be read (and thus stopping the program) when the input contains the number zero. In BF there is no other way of knowing when the input is exhausted.

Explanation*

++++[->++++++++<]>  #Store <space> at index 1
[->+>+<<]           #Move index 1 to index 2 and 3
++++++++++          #Increment index 1 to <newline>
>>                  #Move to index 3
+++++++++++++       #Increment index 3 to <dash>
>                   #Move to (empty) index 4
+                   #Increment to start the main loop
[                   #Main loop
,                   #Read first number to index 4
[->+>+<<]>>[-<<+>>] #Copy index 4 to index 5 (index 5 can now be altered)
,                   #Read second number (the number pair is now stored at index 5 and 6)
<                   #Move to first number (index 5)
[->-<<<<.>>>]       #Decrement index 5 and 6 and print <space> until index 5 equals zero
>                   #move to second input (index 6)
[-<<<.>>>]          #Decrement index 6 and print <dash> until index 6 equals zero
<<<<<.>>>           #Print <newline> and move to index 4 (original first number)
]                   #End of main loop

*(You won't be able to compile/run this due to the comments)

• Brainfuck shorter than Java => world will soon end. – Alex A. Jun 29 '15 at 19:49
• The explanation should actually run fine. The only bf commands in there are < and >, and they're perfectly balanced. – undergroundmonorail Jul 5 '15 at 20:04
• @undergroundmonorail Nice catch, I did not even try to see if they were balanced ;) – Rolf ツ Jul 5 '15 at 21:32

# Pyth, 362219 14 bytes

This is my first Pyth program. Jakube helped golf out 5 bytes!

FNQ<s*V" -"NeN

It expects input in the form [[5,20], [5,20], [2,10], [15,19]].

You can try it online.

# C++14, 69 bytes

[]{int a,b;for(;cin>>a>>b;){cout<<setw(b)<<string(b-a,'-')+'\n';}}();

• Don't you need std:: on cin and cout? – Alex A. Jul 4 '15 at 2:50

# K, 18 bytes

0:" -"@{&x,y-x}.'

Expects a list of pairs as input:

0:" -"@{&x,y-x}.'(0 7;5 6;3 6)
-------
-
---
0:" -"@{&x,y-x}.'(5 20;5 20;2 10; 15 19)
---------------
---------------
--------
----
0:" -"@{&x,y-x}.'()

I unpack each (') tuple using dot-apply (.) so that inside the lambda I have access to the start and end value as x and y, respectively. Then I reassemble these into a (start,length) tuple (x,y-x) and apply "where" (&). This gives me output like so:

{&x,y-x}.'(0 7;5 6;3 6)
(1 1 1 1 1 1 1
0 0 0 0 0 1
0 0 0 1 1 1)

Then I simply have to index into a 2-character array using this ragged matrix (" -"@) and send it all to stdout (0:).

# JavaScript (ES6), 63

Edit 3 byte saved thx @apsillers
63 bytes not counting the assignment to F as an anonymous function is allowed.

A function with a variable number of parameters, as requested.
A function with a list of tasks as a single parameter.

Test running the snippet below (being EcmaScript 6, Firefox only)

F=l=>l.map(t=>console.log(' '.repeat(l=t[0])+'-'.repeat(t[1]-l)))

// TEST

// for this test, redefine console.log to have output inside the snippet
console.log = (...x) => O.innerHTML += x + '\n';

console.log('* Empty'); F([]);
console.log('\n* [0,7],[5,6],[3,6]'); F([[0,7],[5,6],[3,6]])
console.log('\n* [5,20],[5,20],[2,10],[15,19]');F([[5,20],[5,20],[2,10],[15,19]]);
<pre id=O></pre>

• Save one byte by assigning t[0] to a global (or you can safely assign it to l if you don't want to make a global). Also, the spec allows "a named function or an anonymous function" so I think you could omit the F= in your byte count. – apsillers Jun 29 '15 at 16:15
• @apsillers I missed the anonymous think. Thanks – edc65 Jun 29 '15 at 16:23

# Scala, 6763 59 bytes

(r:Seq[(Int,Int)])⇒for((s,e)←r)(println(" "*s+"-"*(e-s)))

Usage: res0() or res0(Seq(28->35, 34->40, 39->44)) etc.

Thanks gilad for shaving 4 bytes using a for expression!

# Ruby: 35 characters

->*t{t.map{|s,e|puts' '*s+?-*(e-s)}

Sample run:

irb(main):001:0> ->*t{t.map{|s,e|puts' '*s+?-*(e-s)}}.call [0,7], [5,6], [3,6]
-------
-
---

Updated to accept multiple two-element arrays, one for each task to display. (I think that is what the updated requirement expects.)

Javascript(ES6), 61/66 chars

My answer is almost similar to the one posted by @edc65 , but with some improvements. As tasks in single array are not allowed(so function would be called like this: a([3,4], [7,15], [0,14], [10, 15])), correct one would be this(66 chars without name assignment):

a=(...x)=>x.map(([c,d])=>console.log(' '.repeat(c)+'-'.repeat(d-c)))

And if one array argument is allowed(so fn call like this: a([[3,4], [7,15], [0,14], [10, 15]])), then it would be(61 char without assignment):

a=x=>x.map(([c,d])=>console.log(' '.repeat(c)+'-'.repeat(d-c)))

# SWI-Prolog, 55 bytes

a([[A,B]|C]):-tab(A),writef("%r",[-,B-A]),nl,C=[];a(C).

Example: a([[5,20],[5,20],[2,10],[15,19]]). outputs

---------------
---------------
--------
----
• I'm sorry but the input format in this answer does not meet specifications - each task should be represented in one argument, not in two. (Unless I missed something in SWI-Prolog syntax which I'm not familiar with...) – Jacob Jun 29 '15 at 14:13
• @Jacob Yeah when rereading your post I figured that and already changed my code to account for it. – Fatalize Jun 29 '15 at 14:14

(#)=replicate
f i=putStr$g=<<(read$'[':i++"]")
g(s,e)=s#' '++(e-s)#'-'++"\n"

Input format is a string of comma separated tuples, e.g. "(1,2),(3,4)".

Usage examples:

*Main> f "(1,2),(3,4)"
-
-

*Main> f "(0,7),(5,6),(3,6)"
-------
-
---

How it works: for input parsing I enclose the input string in [ and ] and use Haskell's native read function for lists of integer tuples. The rest is easy: for each tuple (s,e) take s spaces followed by e-s dashes followed by a newline and concatenate all into a single string. Print.

with relaxed input format:

(#)=replicate
f=putStr.(g=<<)
g(s,e)=s#' '++(e-s)#'-'++"\n"

Now it takes a list of tuples, e.g f [(0,7),(5,6),(3,6)].

Works as described above, but without input parsing.

# Julia, 44 bytes

x->for t=x a,b=t;println(" "^a*"-"^(b-a))end

This creates an anonymous function that accepts an array of tuples as input and prints to STDOUT.

Ungolfed + explanation:

function f(x)
# Loop over the tasks (tuples) in x
for t in x
# Assign a and b to the two elements of t
a,b = t

# Print a spaces followed by b-a dashes on a line
println(" "^a * "-"^(b-a))
end
end

Examples:

julia> f([(5,20), (5,20), (2,10), (15,19)])
---------------
---------------
--------
----

julia> f([(0,7), (5,6), (3,6)])
-------
-
---

julia> f([])
• Sure. Sorry for the inconvenience. – Jacob Jun 29 '15 at 15:16
• @Jacob: No inconvenience. Nice challenge. :) – Alex A. Jun 29 '15 at 15:17

## JavaScript (ES6), 1068580 68 bytes

As per the updated requirements, a list of tasks is now acceptable

a=>a.reduce((p,v)=>p+=' '.repeat(z=v[0])+'-'.repeat(v[1]-z)+"\n",'')

Takes zero or more arguments: 80 bytes

(...a)=>{s='';a.map(v=>s+=' '[r='repeat'](z=v[0])+'-'[r](v[1]-z)+"\n");return s}

Original attempt, 106 bytes:

(...a)=>{for(i=-1,s='',r='repeat';a.length>++i;){s+=' '[r](a[i][0])+'-'[r](a[i][1]-a[i][0])+"\n"}return s}
• If it's ES6, then why not String.repeat()? – manatwork Jun 29 '15 at 14:55
• @manatwork Thanks for showing me something new!! Unfortunately for code golf it is actually longer to use that – rink.attendant.6 Jun 29 '15 at 14:59
• Indeed, that two dimensional a not really helps. I had in mind something like ()=>{for(i=0,s='';a=arguments[i++];)s+='_'.repeat(a[0])+'-'.repeat(a[1]-a[0])+"\n";return s}. – manatwork Jun 29 '15 at 15:11
• r='repeat' ? ... for 2 times? nah! a=>a.reduce((p,v)=>p+=' '.repeat(z=v[0])+'-'.repeat(v[1]-z)+"\n",'') – edc65 Jun 29 '15 at 16:01
• There is no output. Return value does not matter, your code must print the timeline on STDOUT. (and would be shorter too) – edc65 Jun 29 '15 at 16:03

## C: 108 bytes

void g(int*l){for(int c=0;*l>=0;c=!c,l++){if(!c)l[1]-=*l;while(l[0]-->0)putchar(c?45:32);c?putchar(10):0;}}

Ungolfed:

void gantt(int*l) {
for (int c = 0; *l >= 0; c = !c, l++) {
if (!c) l[1] -= *l;
while (l[0]-- > 0) putchar(c? 45 : 32);
c? putchar(10) : 0;
}
}

Takes as a parameter a list of integers terminated by -1. For example:

int list[] = {
28, 35,
34, 40,
39, 44,
-1
};
gantt(list);

It uses c to toggle between writing spaces and dashes.

• Make c static - you can drop its type (it will be int) and initialization (it will be zero). *l>=0 is the same as *l+1 which is shorter. c&&putchar is shorter than ternary. If you replace c=!c with c^=13 (+1 byte) you can change c?45:32 to 32+c (-3 bytes). Move c flip from for to the end of loop: (c^=13)||putchar(10);. c;void g(int*l){for(;*l+1;l++){l[1]-=c?0:*l;while(l[0]--)putchar(32+c);(c^=13)||putchar(10);}} - 94 bytes. – aragaer Jun 30 '15 at 12:37

# Perl: 42 41 characters

Just to have at least one solution with string parsing too.

s!(\d+)->(\d+),?!$"x$1."-"x($2-$1).$/!ge Sample run: bash-4.3$ perl -pe 's!(\d+)->(\d+),?!$"x$1."-"x($2-$1).$/!ge' <<< '0->7,5->6,3->6' ------- - --- • Actually we already have the straightforward Java answer that parse a string :) Anyway, thanks for this one as well! – Jacob Jun 29 '15 at 14:49 • Yes, but as I understand that expects comma separated numbers, not the format specified in the question. – manatwork Jun 29 '15 at 14:53 # Java 8, 280275246204195185 180 bytes void g(String t){for(String s:t.split(",")){String[]a=s.split("->");s="";Integer i;for(i=0;i<i.valueOf(a[0]);i++)s+=" ";for(;i<i.valueOf(a[1]);i++)s+="-";System.out.println(s);};}; A method that takes a comma-seperated input string and prints the resulting ascii Gantt Chart to stdout. Thanks to durron597 and masterX244 for helping me save 10 bytes • I think you're allowed to use a method instead. – lirtosiast Jun 30 '15 at 3:07 • It is allowed iff this is the (or a) way to create an anonymous function in Java8. Is it? – Jacob Jun 30 '15 at 4:03 • It's the closest thing Java 8 has to such a feature. – SuperJedi224 Jun 30 '15 at 11:47 • If you do Integer i=0; you can do for(;i<i.parseInt;, saving 8 characters. – durron597 Jun 30 '15 at 21:43 • I couldn't get it to compile on Ideone, but it appears that it would not accept empty input, as the rules require (t.split(",") would throw an exception). – Nateowami Jul 1 '15 at 6:01 # Java, 187181197183 101 bytes void g(int[][]g){for(int[]i:g)for(int j=0;j<i[1];System.out.print(j++<i[0]?" ":j==i[1]?"-\n":"-"));} Ungolfed (sort of): void g(int[][] g){ for(int[] i : g) for(int j = 0; j < i[1]; System.out.print(j++ < i[0] ? " " : j == i[1] ? "-\n" : "-")); } Accepts input as 2d array of ints. Thanks to masterX244 for pointing out that this is allowed by the rules. • you can shorten the loops if you use the 3rd bulletpoint of the current question version and varargs for the input – masterX244 Jul 1 '15 at 14:37 • @masterX244 Thanks, I missed that. Seems to me like cheating to have it pre-parsed, but if cheating is allowed... whatever. I'll update it when I have time. – Nateowami Jul 2 '15 at 1:06 # Jelly, 13 9 bytes ạ\⁾ -xµ€Y Try it online! Takes input as [[5, 20], [5, 20], [2, 10], [15, 19]]. -4 bytes thanks to Erik # APL (Dyalog Classic), 12 bytes ↑('-'\⍨≤∘⍳)/ Try it online! APL has no varargs, so the arg here is a single Nx2 matrix. • If you can take two arguments (starts and ends) then ↑'-'\⍨¨≤∘⍳¨ – Adám Jan 29 '18 at 10:32 • challenge author says no – ngn Jan 29 '18 at 15:26 # JavaScript (ES8), 54 bytes a=>a.map(([x,y])=>"-".repeat(y-x).padStart(y)).join  Try it online # PowerShell 3.0, 4836 Bytes$args|%{" "*$_[0]+"-"*($_[1]-$_[0])} Thanks to Mazzy for saving 12 with a better way to pass in the list Old code and explanation: &{param($b="")$b|%{" "*$_[0]+"-"*($_[1]-$_[0])}}

Takes arguments as a list of tuples, e.g. (5,20),(5,20),(2,10),(15,19). Had to default $b to a value to take care of the empty string because it somehow entered the foreach block when called with no input. • 36 bytes:$args|%{" "*$_[0]+"-"*($_[1]-$_[0])}. Save as get-asciiGantt.ps1. Test script .\get-asciiGantt.ps1 (5,20) (5,20) (2,10) (15,19) – mazzy Jun 19 '18 at 9:40 # R, 117 90 75 bytes function(y)for(i in 1:ncol(y))cat(" "<y[1,i],"-"<diff(y)[i]," ") "<"=strrep Try it online! Giuseppe golfed at least 29 bytes off my original answer! The idea is straightforward: print as many " " as necessary followed by as many "-" as required. Input is a 2*L matrix with L the number of pairs. The vectorized function diff is used to get the number of "-". • @Giuseppe this is what I get for trying to trying to stick to my original matrix idea while using a for loop... ty! – JayCe Jun 18 '18 at 16:02 • 86 bytes – Giuseppe Jun 18 '18 at 17:34 • @Giuseppe Transposed y to save a few more :) – JayCe Jun 18 '18 at 17:52 • Now 1- indexed would save 4 – JayCe Jun 18 '18 at 17:55 • Nice, use < instead of * and you can get this to 81 bytes – Giuseppe Jun 19 '18 at 20:19 # VBA (Excel), 99 90 bytes Using Immediate Window and [A1] as input eg.0-1,2-5 Thanks to @TaylorSott for cutting some bytes. b=Split([A1]):For x=0To Ubound(b):c=Split(b(x),"-"):?Spc(c(0)-0)String(c(1)-c(0),"-"):Next • If you change the input format to being space delimited rather than being comma delimited, you cang chage the first two clauses from a=[A1]:b=Split(a,",") to b=Split([A1]). Also, you can drop the space before the To in the For loop declaration. – Taylor Scott Feb 15 at 19:34 • Thanks and noted! :D – remoel Feb 18 at 8:32 ## CoffeeScript, 104 82, 65 bytes List of tasks (ES6): 65 bytes (a)->a.map (v)->console.log ' '.repeat(v[0])+'-'.repeat v[1]-v[0] List of tasks (ES5 variant): 82 bytes (a)->a.map (v)->j=-1;s='';s+=(if j<v[0]then' 'else'-') while++j<v[1];console.log s Zero or more arguments: 104 bytes ()->[].slice.call(arguments).map((v)->j=-1;s='';s+=(if j<v[0]then' 'else'-')while++j<v[1];console.log s) Unminified: () -> [].slice.call(arguments).map( # convert to array-like arguments to array and loop (v) -> j = -1 # counter s = '' # initialize string s += (if j < v[0] then ' ' else '-') while ++j < v[1] console.log s # print to STDOUT ) • Not sure from where to where is the JavaScript, CoffeeScript and ECMAScript in your answers, but in ECMAScript you can use Array.from(arguments) instead of [].slice.call(arguments). – manatwork Jun 29 '15 at 16:42 • @manatwork As you can see in my answers (both ES5 and ES6, in CoffeeScript) addressing the changed requirement allowing a list of tasks, I don't need to reference arguments any more. – rink.attendant.6 Jun 29 '15 at 16:56 ## PHP, 94 91 bytes Takes a list of tasks (e.g. [[5,20],[5,20],[2,10],[15,19]]). Thanks @IsmaelMiguel for the reminder of variable function names. function x($a){$r=str_repeat;foreach($a as$v){echo$r(' ',$v[0]).$r('-',$v[1]-$v[0])."\n";}}

Original attempt: 94 bytes

function x($a){foreach($a as$v){echo str_repeat(' ',$v[0]).str_repeat('-',$v[1]-$v[0])."\n";}}
• 73 bytes, PHP4: $R=str_repeat;foreach($G as$v)echo$R(' ',$v[0]),$R('-',$v[1]-$v[0]),'\n'; (replace the \n with a real newline). For this to work, you need to send an array on the key $G, over POST/GET/SESSION/COOKIE... – Ismael Miguel Jun 29 '15 at 19:04 • @IsmaelMiguel According to the question, the input needs to come as an argument or from STDIN. – rink.attendant.6 Jun 29 '15 at 19:15 • Does GET parameters count? And I think that GETuses STDIN. – Ismael Miguel Jun 29 '15 at 19:22 # PHP, 89 characters (function body) function gantt($x){array_walk($x,function($a){echo str_pad(str_repeat('-',$a[1]-$a[0]),$a[1],' ',0)."\n";});} I was going to go for reading strings, but as a lot of the entries were taking arrays of integer pairs, I figured I would follow suit for the sake of brevity. For each tuple$a in array $x I echo a string of dashes repeated$a[1] - $a[0] times, padded up to the larger number$a[1] with spaces. Then the obligatory newline.

• You can make your function name just a single letter to save a few bytes. or better yet, if PHP supports anonymous functions, just omit a function name altogether. – Alex A. Jun 29 '15 at 17:43
• Oh I see now what you mean by "function body." You have to count the entire function definition in your score, not just the innards. – Alex A. Jun 29 '15 at 17:47
• printf() seems shorter than echo+str_pad(): function gantt($x){array_map(function($a){printf("%$a[1]s␊",str_repeat('-',$a[1]-$a[0]));},$x);} (The ␊ in the code is for a literal newline: just wrap your code there.) – manatwork Jun 29 '15 at 17:55
• Actually a good old foreach is better: function g($x){foreach($x as$a)printf("%$a[1]s␊",str_repeat('-',$a[1]-$a[0]));} And this is 79 characters including everything. – manatwork Jun 29 '15 at 17:58
• @AlexA. ah, I've seen golfs where people count or discount function headers. I wasn't sure what to go for, hence why I specified what count was what. – JPMC Jun 29 '15 at 20:13

# Gema: 47 characters

<D>-\><D><y>=@left{$1;}@repeat{@sub{$2;$1};-}\n Sample run: bash-4.3$ gema '<D>-\><D><y>=@left{$1;}@repeat{@sub{$2;\$1};-}\n' <<< '0->7,5->6,3->6'
-------
-
---

# PostgreSQL: 160 characters

create function g(int[])returns text as
$$select string_agg(lpad(repeat('-',1[x][2]-1[x][1]),1[x][2]),chr(10))from generate_subscripts(1,1)x$$
language sql;

Sample run:

manatwork=# create function g(int[])returns text as
manatwork-# $$select string_agg(lpad(repeat('-',1[x][2]-1[x][1]),1[x][2]),chr(10))from generate_subscripts(1,1)x$$
manatwork-# language sql;
CREATE FUNCTION

manatwork=# select g(array[[0,7],[5,6],[3,6]]);
-------
-
---

# J, 21 bytes

(' -'#~{.,-~/)"1 ::''

## ungolfed

(' -' #~ {. , -~/)"1 ::''

This is essentially just J's copy verb #, but its we're copying the space character head of list {. number of times, and the hyphen character "2nd list element minus 1st list element" number of times: -~/. Sadly this forces us to have to specify the rank "1 explictly, and we need to use Adverse :: to handle the empty case.

Try it online!