Unscramble the Source Code

Question

Note: This challenge is now closed to new cop submissions. This is to ensure that no one can post submissions that only remain uncracked because there aren't enough robbers interested in the challenge anymore.

In this game of cops-and-robbers, each cop will write a simple program to give a single output. They will then make public four things about their program:

1. The language
2. The program length
3. The desired output
4. A scrambled-up version of the source code

Then, the robbers must unscramble the source code so that their program functions like the original.

---

Cop Rules

You are to write a simple program, which the robbers will try to recreate.

Your original program must have a simple functionality: upon execution, it outputs a single string/number and halts. It should give the same output regardless of when/where it is run, and should not depend on extra libraries or the internet.

Your program and output must use printable ASCII (newlines and spaces allowed). The output should be no more than 100 characters long, and the program should take less than about 5 seconds to run on a reasonable machine. You are also not allowed to use hashing (or other cryptographic functions) in your program

Then, you provide a scrambled-up version of the source code and the required output. You can scramble up your source code however you may like, as long as characters are conserved.

Your score is the shortest program you have submitted which hasn't been cracked. After a period of one week, an uncracked submission will become immune. In order to claim this immunity, you should edit your answer to show the correct answer. (Clarification: Until you reveal the answer, you are not immune and can still be cracked.) The lowest score wins.

Simple Example Cop Answers

Perl, 20

ellir"lnto Wo d";prH

Hello World

Or...

Perl, 15

*3i)xp3rn3*x3t(

272727

Robber Rules

Robbers will post their cracking attempts as answers in a separate thread, located here.

You have one attempt at cracking each submission. Your cracking attempt will be an unscrambled version of the source code. If your guess matches the description (same characters, output, and of course language), and you are the first correct guess, then you win a point. It is important to note that your program does not have to exactly match the original, simply use the same characters and have the same functionality. This means there could be more than one correct answer.

The robber with the most points (successful cracks) wins.

Simple Example Robber Answers

Your program was print "Hello World";. (Although print"Hello World" ; could have also worked.)

Your program was print(3**3x3)x3

Safe Submissions

1. ASP/ASP.Net, 14 (Jamie Barker)
2. Befunge-98, 15 (FireFly)
3. GolfScript, 16 (Peter Taylor)
4. CJam, 19 (DLosc)
5. GolfScript, 20 (user23013)
6. Perl, 21 (primo)
7. Python, 23 (mbomb007)
8. Ruby, 27 (histocrat)
9. SAS, 28 (ConMan)
10. Ruby, 29 (histocrat)
11. Python, 30 (mbomb007)
12. JavaScript, 31 (hsl)
13. Ruby, 33 (histocrat)
14. Marbelous, 37 (es1024)
15. Ruby, 43 (histocrat)
16. PHP, 44 (kenorb)
17. Ruby, 45 (histocrat)
18. Marbelous, 45 (es1024)
19. Python 2, 45 (Emil)
20. PHP, 46 (Ismael Miguel)
21. Haskell, 48 (nooodl)
22. Python, 51 (DLosc)
23. Python, 60 (Sp3000)
24. Python 2, 62 (muddyfish)
25. JavaScript, 68 (Jamie Barker)
26. Mathematica, 73 (Arcinde)
27. Haskell, 77 (proudhaskeller)
28. Python, 90 (DLosc)
29. C++, 104 (user23013)
30. ECMAScript 6, 116 (Mateon1)
31. C++11, 121 (es1024)
32. Grass, 134 (user23013)
33. PowerShell, 182 (christopherw)

Unsolved Submissions

In order of time of posting. This list courtesy of many users.

• CoffeeScript, 96 (soktinpk)
• Python 3, 70 (Sp3000)
• TinyMUSH 3.1, 20 (Muqo)
• GolfScript, 32 (Beta Decay)
• Python 2, 101 (Mateon1)
• Lua, 49 (ChipperNickel)
• Python, 61 (imallett)
• Java 6+, 218 (nhahtdh)
• CJam, 51 (Martin Büttner)
• J, 22 (FireFly)
• Marbelous, 106 (es1024)
• Marbelous, 107 (es1024)
• JavaScript, 79 (FireFly)
• CJam, 47 (user23013)
• Rust, 118 + Clojure, 106 + others (Vi.) - version 2
• Marbelous, 144 (es1024)
• Python 2, 80 (MrWonderful)
• Perl, 53 (DLosc)
• Perl, 26 (primo)
• Mathematica, 31 (Arcinde)
• Marbelous, 144 (es1024)
• Assembly, 78 (krzygorz)
• J, 14 (algorithmshark)
• Java 8, 157 (TheBestOne)

A small tool to verify solutions, courtesy of n̴̖̋h̷͉̃a̷̭̿h̸̡̅ẗ̵̨́d̷̰̀ĥ̷̳

$(function(){function e(){var e=$("#ignore-space").is(":checked");var t=$("#source").val().split("").sort();var n=$("#editor").val().split("").sort();var r,i=0;for(r=0;r<t.length;){if(t[r]==n[i]){t.splice(r,1);n.splice(i,1)}else if(t[r]>n[i]){i++}else{r++}}$("#display").val(t.join(""));n=n.join("");if(e){n=n.replace(/[\r\n\t ]/g,"")}if(n.length!=0){$("#status").addClass("bad").removeClass("good").text("Exceeded quota: "+n)}else{$("#status").addClass("good").removeClass("bad").text("OK")}}$("#source, #editor").on("keyup",function(){e()});$("#ignore-space").on("click",function(){e()});e()})

textarea{width:100%;border:thin solid emboss}#status{width:auto;border:thin solid;padding:.5em;margin:.5em 0}.bad{background-color:#FFF0F0;color:#E00}.good{background-color:#F0FFF0;color:#2C2}

<script src="https://ajax.googleapis.com/ajax/libs/jquery/1.11.1/jquery.min.js"></script>
<h3>Scrambled Source</h3>
<textarea id="source" class="content" rows="10"></textarea>
<h3>Unused Characters</h3>
<textarea id="display" class="content" rows="10" readonly></textarea>
<h3>Your Solution</h3>
<input type="checkbox" id="ignore-space" name="ignore-space"/>
<label for="ignore-space">Ignore space characters</label>
<div id="status" class="good">OK</div>
<textarea id="editor" class="content" rows="10"></textarea>

Answer 1

Python 3 - size 34 [cracked by FireFly]

Scrambled:

print(....__str__)
[[(4,)=={-6}]];

Output:

.

The scrambled code runs fine, but it's not what you want ;).

---

Intended solution

Awkwardly, my red herrings backfired on me, and I only realised a few minutes after posting.
The intended solution was:

_={};_[6.,...]=_ print(str(_)[-4])

If we print out _, we get:

{(6.0, Ellipsis): {...}}

If you add a list or a dictionary to itself in Python, the string representation replaces that part of it with ...! Yes, that means you can do things like:

a=[] a.append(a) print(a[0][0][0][0][0][0][0][0])
:)

Answer 2

Java, size 134 (Cracked)

Code:

"''"{}{}[]()()()()()() <<<<<<< +++-===^ ..;;;; 11113444578999 aaaaabcccccdefgghhhiiiiiiiillmmnnnnnnnnnooopprrrrrrrssssSSStttttttttuuvy

Output:

^74
^376
^2280
^17512
^139368
^1114216
^8913000

Answer 3

Ruby, size 17

Code

$$$$--::<<>>??Wip

Output

1

Answer 4

CJam, size 51

Thought I'd give this a shot!

Code

""%)))))****+++,/::@ABBBBJJSSUWYYYZZ[]____accdeimos

Output

Bond, James Bond.

Answer 5

Ruby, 45

Code

%()[]./ 0123456789_abcdefghijklmnopqrstuvwxyz

Output

9410663329978946297999932

Original Code

p %q[zyfnhvjkwudebgmacsxrl].to_i(36)/51074892

See charredUtensil's hack attempt in the comments for a great explanation. And yes, I generated the code randomly. I hoped the "pangram" approach would paralyze people with the number of options.

Answer 6

Python shell: 15 chars

I had this strange urge not to include any 0's, 3's, or 7's.

11124455566899^

Output (to shell):

339018

Answer 7

PHP, 44 [SAFE]

Code:

""(())46;=BFGHNOWWZ_aaabbcddeeeejjloopssvwyy

Works on PHP 5.3-5.6. A bit long, but I hope it's difficult.

Output

4096

Original code

eval(base64_decode("ZWNobyBwaGFyOjpHWjs=")); which is the same as: eval(base64_decode(base64_encode("echo phar::GZ;")));

Explanation

The code will print predefined constant GZ of Phar class which is just compression constant (0x00001000 = 4096). Then the code is encoded with MIME base64 text format and eval'ed (executed). Note that base64 is data encoding format (not hashing or cryptographic function).

Answer 8

Python, size 60 [safe]

Scrambled:

\n    ((((()))))***,-0123:=____aaaabdegiilmmmnnoopprrrrsttttu

Output:

1177652997443428940314

This probably still falls on the moderately-okay-to-crack side, but I thought the idea might be interesting. \n is a newline.

---

Solution

The number, searchable on OEIS, is the sum of the first 23 factorials. To implement a factorial function without any ifs for the base case we use Python's short circuit evaluation and the quirk that 0**0 evaluates to 1.
t=lambda _:0**_ or _*t(_-1)or i print(sum(map(t,range(23))))

Answer 9

Ruby, 33 [safe]

Code:

&(())*09:==_yogurt puns...perturb

Output:

[0, [0]]
[1, [1]]
[2, [2]]
[3, [3]]
[4, [4]]
[5, [5]]
[6, [6]]
[7, [7]]
[8, [8]]
[9, [9]]

Source:

e=r=p *(0..9).group_by(&:untrust)

Explanation:

e=r= is red herring filler (chosen to create the anagram "puts group_by return"). (0..9).group_by takes a block, passes each number from 0 to 9 into it, and creates a hash mapping distinct outputs of the block to the input. But instead of a literal block, I'm using the & syntax, which calls to_proc on the :untrust symbol to create a proc that "taints" each thing passed into it (which does nothing in this context), and returns the input. So each item in the range gets its own bucket. p * then converts the hash to an array of arguments of the form [key, value_array], and prints each on its own line.

Answer 10

SAS, 17 (Cracked)

%%()1;>>aabelptuv

Outputs 0

Answer 11

Python 3, size 48 (Cracked)

#print(']'2p/')'\xdh))2[:roo,h+(2*e'rw)/r(d'o(()

Output

head

Hopefully this one lasts a bit longer...

Answer 12

GolfScript (13 bytes) (Cracked)

&+,,../224?{}

Expected output:

1234260885218145824583458445854586253225388392299142924142925142926142927142928302928497409497409

Answer 13

Mathematica, 73 72 bytes [Safe]

(I miscounted the bytes when I made the post.)

Code:

""""#&()*,,01>@@@@@[[]]`StrongSolemnNiceLustyTensorSextetNastiesEastCamp

Output:

0.5 + 14.1347 I

Original code:

N@(ToExpression@Last@Select[Names@"System`*",StringCount[#,"eta"]>0&])@1

Answer 14

Marbelous 45 [SAFE]

Chars

^^^^^^^^^^!!&&000004459:<<<>>>\\EE//V{{{{}}}}

Each ^ is a placeholder for a newline.

Output

Z{-8H`

Source

44 {} :{} }0E5 << }0 E9 &0&0 >>\/ >V!! /\{<{0

{} is a board that calculates (((}0 * 3 - 23) mod 256) - 27)/2. This is a Linear Congruential Generator, with its output modified to fit in the printable ascii range more often. The program terminates at the first output that is not a printable ascii character.

JS Marbelous Interpreter

This requires cylindrical boards (i.e. marbles pushed off of the board on the left reappear on the right and vice-versa).

Answer 15

Marbelous, 37 [safe]

Code

^^^^^^^^!!-0002:::<<<<>>>>\F///WW{{}}

^ represents a newline; no ^s are to be used.

Output

lHp`

Original Source

F2 >: :>: }0 <<}0 -W// << >W!! /\{0{>

This passes 0xF2 into the board >:, which finds 2*(3*}0-32) mod 256, checks to make sure the result is greater than 32, and if so, outputs it down and left. The down marble is emitted as a character, and the left character is passed into >: again.

JS Marbelous Interpreter

This requires cylindrical boards (i.e. marbles pushed off of the board on the left reappear on the right and vice-versa).

Answer 16

SAS, 28 (SAFE)

%1(STAR)SPY,&2STUD;BUNNIES,%

Outputs I

(that's capital letter i)

Solution:

%PUT%SUBSTR(&SYSENDIAN,2,1);

Which outputs the second character of the automatic macro variable SYSENDIAN (which takes the values BIG and LITTLE) to the log.

Answer 17

J - 14 char

Going for the gold. Edit: I have been recently made aware that this code performs differently depending on whether you are using a 32-bit or 64-bit version of the latest J (j803). If this is grounds for disqualification, let me know: otherwise, I would concede a crack as an unscrambling that gives the correct answer on at least one of the two versions.

Scrambled:

7*8..?56!/x\+/

32-bit output:

2704156

64-bit output:

2722721

Answer 18

CJam, 19 bytes (SAFE)

Code:

%\@_@){;_0}*__*)\15

(A few smileys survived the factory explosion...)

Output:

10120125854501259106171741725012526265263117252614371738261720339343012

Solution:

1_{_@)_@_*)\%}50*\;

Calculates and concatenates together the first 51 terms of the following sequence:

a[1] = 1
a[n] = (a[n-1]**2+1)%n

The sequence increases rapidly, but every time a[n] gets above n, it gets knocked down again by the modulo operation. Here's the same thing with spaces inserted:

1 0 1 2 0 1 2 5 8 5 4 5 0 1 2 5 9 10 6 17 17 4 17 2 5 0 1 2 5 26 26 5 26 31 17 2 5 26 14 37 17 38 26 17 20 33 9 34 30 1 2

Answer 19

Perl - 21 bytes (safe)

~$is^*peter**+working

(the above evaluates to 1, btw)

Ouput

65534

---

Original

s**$~^wkirgo*ee+print

Explanation

The code performs a substitution which is evaluated, and then the result of that is evaluated (ee). The variable $~ starts its life as STDOUT, so that $~^wkirgo evaluates to $?-=2;.

The variable $? is also special, in that it is stored internally as an unsigned short. Decrementing it by two wraps around to 65534.

Answer 20

ASP/ASP.Net - 14 [SAFE]

Code

<%=Len('12')%>

Output

1

Answer

<%=1)''Len(2%> (Basically the equivalent of writing <%=1%>)

Answer 21

CoffeeScript, size 25 [Cracked]

Code

the terse tales // [.]

Output

false

You can try it out here in the "Try CoffeeScript" tab.

Answer 22

JavaScript, 26 (cracked)

later(103>(8-9)&80>4|10>9)

Output

999

Answer 23

PHP, Size 14 (Cracked)

aaaaaa:::::??;

Prints nothing when notices are turned off (no warning or fatal error messages).

Answer 24

JavaScript, 61 (Cracked)

Code

x=next facile tortuously inept encrypter.{}()()()()(4-1);;ffo

Output

ring

Answer 25

Python 3, size 37 (Cracked)

Code:

int(22 ** 2 ** 2+range(irr)[]__npof_)

Output (space separated):

0 1 4 9 16 25 36 49 64 81

Answer 26

Python 2, 101

Code

"rne) n(sxx+ai(cpi)% ( *2n+f 1 )corii\.5n /r +5xeini+"h(5te4437f3)xjxo2]2or[)%(#"g r"+"lf/r2 # 1xn"h5

Output

'4
[#
;4
E4
##
Y4
c4

Answer 27

Ruby - 58 (Cracked)

puts 00000000111222344455556677888889^^^^aacdddddefffxxxxx

Yeah you need to xor the hexadecimal literals in this one...

Output:

1436068433

Hope this lasts more than my last one...

Answer 28

Brainfuck, 118 chars [cracked by feersum]

Scrambled:

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

Output:

q$V4yQ!

This won't win, but there just had to be one. Assumes no cell wrapping, and works even if the tape has no cells to the left. I don't know how BF is scored, so I've put down chars.

^{For a BF program, 118's not that long...}

---

Explanation

The intended solution was ++++[>++++<-]>[>+>>>+>>>>+<<<<<<<<-]>>+++++++>+++++>+>+++++[>++++++++<-]+++>>++>[[<[>>+>+<<<-]>>[<<+>>-]<-]<]>>>>[+.>]. The ASCII values are [113, 36, 86, 52, 121, 81, 33], from which we can subtract 1 to get [112, 35, 85, 51, 120, 80, 32]. This can be achieved by building the numbers [16, 7, 5, 17, 3, 40, 2, 16] then multiplying pairwise.
However, as feersum points out, this is completely irrelevant in BF as long as you can golf your code down to be short enough.

Answer 29

C, 30 (cracked by feersum)

Code

_L%"p;at(_N()m_Enin{_}diIf)",r

Output

1

Answer 30

ECMAScript, 26

Run in a Chrome browser console

Code

$( convenient )({"awful"})

Output

$ {}