# Possibly quit your job with a polyglot

Despite your protest, you have been put to work by your boss on a program that takes a single unsigned integer as input and prints the string "prime" if that integer is prime and "not prime" if it isn't. You may choose which language you do this in, as long as the resulting program is short; your boss very much appreciates a low character count. (He will actually manually count the characters after he prints out the source code.)

So you better get to it, the lowest character count wins.

## The fun part

This is just between you and me, but your program should also be valid in another language. In this language though, it should print the string "If the boss finds this, I quit.". Make sure that your boss does not figure out there is a hidden insult when reading the code over and over as he keeps forgetting if he'd counted to 17 or 18 so far. For that reason you may not use any of the words in "the boss finds this quit" in part of the code nor can you use and anagrams of those 5 words.

Bonus challenge which I have created a bounty for

Write a program that actually looks like it just solves the first question to the untrained eye and doesn't seem to contain unnecessary characters. This includes comments and segments of code that obviously don't contribute anything. Ideally, a layman would believe that your program is actually as short as it can be without being unmanagable. A helpful comment here and there is okay though.

The rules for the bonus challenge are a bit looser, instead of being judged on easily measurable criteria, your program will be judged more on how it comes accross to me (and the voters of course)

I will be the final judge of which entry comes closest to deserving this bounty.

Edit:

After a few minutes of your boss counting, he had one of your colleages write a character counting program for him. So even characters that aren't visible count towards your character count.

• Well, Whitespace people will have an easy time here. Aug 26 '14 at 18:25
• Unfortunately, the shortest possible Whitespace program which prints that message is 372 characters. Aug 26 '14 at 19:02
• But the boss will print it out and count the characters. Trailing whitespace won't count. Aug 26 '14 at 19:50
• The bounty challenge sounds interesting but will no doubt be won very trivially by a "well documented and indented" program in an arbitrary language (with the hidden one being whitespace). Aug 26 '14 at 20:31
• So far, all of the entries are obviously not serious programs (random gibberish, etc.). It would be really interesting if someone made a polyglot (that doesn't include Whitespace, of course) that actually looked reasonable at first glance (even if it was really, really long).
– Doorknob
Aug 26 '14 at 21:36

# Python 2 / Python 3 - 128

n=int(input())
s="Isf  tthhies ,b oIs sq ufiitn.d"
print(["not "*any([n%a<1for a in range(2,n)])+"prime",s[::2]+s[1::2]][1/2>0])


Identifies prime numbers with Python 2, get's you into trouble with Python 3.

## Bonus

__help__ = """In the forloop I'm setting the values i to n.
To be a "notPrimes", it's n % i == 0 if: 1 < i, i < n. (We
do tests i<1/2 i==1, too). Then, prints resulting, prime text.
In case i==n: quit. Just if it aborts: try."""

# read integer from command line
n=int(input())

try:

# primes = True
notPrimes = False

# try each i to n
for i in range(n):

# ignore 0 or 1
if i < 1 / 2 or i == 1:
continue

# test divisibility
if n % i == 0:
notPrimes = True

# print result
if notPrimes:
print("not prime")
else:
print("prime")

except:

# if program aborts: print help and error code
print(__help__ [::7])


Try it with Python 2 or Python 3! (In contrast to the golf version above the roles changed: Python 3 is the prime number identifier. Python 2 contains the Easter egg.)

And I do use the word "quit". But somehow I need to describe when my program terminates. ;)

• exit? stop? abort? Aug 28 '14 at 21:19
• @MooingDuck: You mean I could use one of these words? No, then it would not work. ;) Aug 28 '14 at 21:21
• Is this the "/ problem"? (integer division vs. floating point division)
– hlt
Aug 28 '14 at 21:22
• The second one blew my mind - until I looked a bit closer. It seems that I too have fallen into the habit of reading comments rather than code. Aug 30 '14 at 10:54
• The second one is really nice! Kudos! Aug 30 '14 at 13:40

# Bonus submission (C / C++11)

Primality testing using the usual naive method is so mainstream. That's why I have invented a brand-new randomized naive method! This test is as follows:

1. Choose any integer d at random. It must not be smaller than 2 and larger than a bit more than sqrt(n).
2. If d is a divisor of n, output not prime.
3. If we made this test 20sqrt(n) times, output prime, else repeat.

If the number is composite, there is only very little probability (about 10-9) that it doesn't work. Of course, I don't believe the C/C++ pseudorandom number generator is powerful enough. That's why I use my own 256-bit LFSR generator!

#include <stdio.h>
#include <stdlib.h>
#include <math.h>

/* A 256-bit linear feedback shift register generating pseudorandom
* numbers (its period is 2^256 - 1).
*/
struct LFSRGenerator {
unsigned seed[8];
};

void lfsr_init_generator(struct LFSRGenerator *gen){
gen->seed[0] = 0xE840CC92; gen->seed[1] = 0xC440CAD0;
gen->seed[2] = 0x40E6E6DE; gen->seed[3] = 0xC8DCD2CC;
gen->seed[4] = 0xD0E840E6; gen->seed[5] = 0x4058E6D2;
gen->seed[6] = 0xEAE24092; gen->seed[7] = 0x145CE8D2;
}
void lfsr_proceed(struct LFSRGenerator *gen){
// LFSR taps are x^256, x^254, x^251 and x^246
unsigned new_bit =
((gen->seed[7]>>0)^(gen->seed[7]>>2)^
(gen->seed[7]>>5)^(gen->seed[7]>>10)) & 1;

// shift seed right
gen->seed[7] >>= 1;
int cell;
for(cell = 6; cell >= 0; cell--){
gen->seed[cell+1] |= ((gen->seed[cell]&1)<<31);
gen->seed[cell] >>= 1;
}
gen->seed[0] |= (new_bit<<31);  // put new bit
}
void lfsr_error(struct LFSRGenerator *gen){
fprintf(stderr, "Error! Developer info:\n");

int cell;
for(cell = 0; cell < 8; cell++){
unsigned val = gen->seed[cell];
putc((char)(val&0xFF), stderr); val >>= 8;
putc((char)(val&0xFF), stderr); val >>= 8;
putc((char)(val&0xFF), stderr); val >>= 8;
putc((char)(val&0xFF), stderr);
}
putc('\n', stderr);
exit(1);
}
int lfsr_get_num(struct LFSRGenerator *gen, int min_val, int max_val){
lfsr_proceed(gen);
int mod_num = max_val-min_val+1;   // = number of possible results
if(mod_num <= 0)
lfsr_error(gen);

// take 6 first cells and compute them mod 'modNum'
unsigned long long result = 0;
int cell;
for(cell = 5; cell >= 0; cell--){
result = ((result << 32) | gen->seed[cell]) % mod_num;
}
return (int)result + min_val;
}

/**********************************************************************/

void end_not_prime(){
printf("not prime\n");
exit(0);
}
void end_prime(){
printf("prime\n");
exit(0);
}

int main(){
int number;
struct LFSRGenerator gen;
lfsr_init_generator(&gen);

printf("Provide a number to check its primality: ");
scanf("%d", &number);

if(number <= 1){
end_not_prime();
}
if(number == 2){
end_prime();
}

// just to make sure:
//  * make 20*sqrt(n) tests
//  * generate random divisors from 2 to 111111/100000 * sqrt(n)
//      (in case max range doesn't include sqrt(n)
auto num_checks = (int)floor(sqrt(number)*20);
auto max_range = sqrt(number);
max_range /= 100000;
max_range *= 111111;
max_range = floor(max_range+0.5);

while(num_checks--){
int rnd_div = lfsr_get_num(&gen, 2, max_range);
if(number % rnd_div == 0){
end_not_prime();
}
}
end_prime();
}


C++11 works correctly. However, C compiler seems to be outputting a faulty program for n > 2...

Note: remember that C needs -lm option (link math library) to compile successfully.

Look at max_range variable. C++11 keyword auto resolves to a "matching type" - in this case double. However, in C it's defined as a variable modifier (as static is) - it doesn't define the type. Thus max_range type is a default C type, that is int. When we "try" to multiply this variable by 1.11111, in C it gets "unintenionally" zeroed during division by 100000. We get an incorrect interval of random numbers to be generated and LFSR after proceeding its internal state generates an error, outputting the binary dump of the seed. That's "accidentally" The message If the boss finds this, I quit.\n

If you find the following faulty output:

Error! Developer info:
If the boss finds this, I quit.


incorrect, just remove the appropriate fprintf line.

• It looks pretty convincing. Care to explain a little, so I don't have to dig out my compiler? :) Aug 29 '14 at 20:48
• OK, I added it. Aug 29 '14 at 21:13
• This is amazing! Aug 30 '14 at 9:34
• Very nice! Best answer so far, in my opinion. Aug 30 '14 at 10:59
• Really impressive!
– Ven
Sep 8 '14 at 15:04

# Mathematica/Brainfuck, 260

If[PrimeQ[Input[]],"prime","not prime"](*++++++++++[>+++>++++>+++++++>++++++++++>+++++++++++<<<<<-]>>>+++.>++.<<<++.>>>>++++++.<++.---.<<<.>>>---.>-----.++++..<<<<.>>>++++.+++.>-----.<-----.>+++++.<<<<.>>>>+.<++++.+.>-.<<<++++.<.>>.<<.>>>>--.++++.<.>-.<<<++.*)

• Olson! This code is several times longer than it needs to be! What are all these unnecesary pluses and greater than signs for? You're fired! No, sir, I think you'll find I quit. Aug 26 '14 at 18:53
• @steveverrill That's one way of quiting your job I suppose. Aug 26 '14 at 18:54

# Golfscript / Javascript (126 125129130132134205207)

1.//"Jg!uif!cptt!gjoet!uijt-!J!rvju/"{(}%'
alert((/^1?$|^(11+?)\1+$/.test(Array(+prompt()+1).join(1))?"not ":"")+"prime");';#'


I'd say it's shockingly close to those Mathematica solutions which, afterall, have a built-in check for prime numbers.

Edit: Thanks to Peter for saving another two six bytes!

Here's some details:

• The first 1. is needed because the following // is a comment in Javascript, but performs division twice in Golfscript. This will error out if nothing is on the stack, so we need to give it two numbers. Incidentally, 1. is a perfectly valid syntax in Javascript and will just be ignored.
• "…"{(}% takes the string, decrements their character code values by one and pushes it as a string. This results in the string we need to print.
• ' starts a string in Golfscript which by default extends over several lines, causing the Javascript below to only be put into the string.
• Next is the Javascript code, which uses a somewhat well-known approach to detect prime numbers through regular expressions.
• ';#' closes the multi-line string in Golfscript, discards it and then ignores the rest of the line. In Javascript, this is simply a string literal which will be ignored.
• In GS 1+ is ). And 1 1 is 1., which I suspect the JS would be as happy with as 1 Aug 26 '14 at 20:01
• @PeterTaylor Awesome, thanks! I've incorporated it. Aug 26 '14 at 20:05
• Also, if you map something over a string you get a string, so {)}/]""+ could just be {)}%. Aug 26 '14 at 21:05
• @PeterTaylor You're the man! :) Aug 27 '14 at 5:01
• @overactor Ugh, same mistake here. Shame on me. I'll have to fix it later tonight, though. Aug 27 '14 at 12:27

# CJam (GolfScript), 60 59 bytes

"Jg!uif!cptt!gjoet!uijt-!J!rvju/"{(}%S#];"not prime"limp4*>


Thanks to @mnbvmar for golfing off 1 byte!

### How it works (CJam)

"Jg!uif!cptt!gjoet!uijt-!J!rvju/"  " Push that string.                                    ";
{(}%                               " Subtract 1 from each character code.                 ";
S#                                 " Find the index of the first space.                   ";
];                                 " Wrap the entire stack in an array and discard it.    ";
"not prime"                        " Push that string.                                    ";
li                                 " Read an integer from STDIN.                          ";
mp4*                               " Push 4 if it's prime and 0 if it isn't.              ";
>                                  " Remove that many chars from the start of the string. ";


Try it online!

### How it works (GolfScript)

"Jg!uif!cptt!gjoet!uijt-!J!rvju/"  # Push that string.
{(}%                               # Subtract 1 from each character code.
S#];"not prime"limp4*>             # Noop followed by a comment.


Try it online!

• +1: The boss is gonna love how short this is. And he definitely won't see the hidden string. He might also have trouble seeing the prime check, though. :) Aug 30 '14 at 17:34
• Let's hope he believes the hidden string is the prime check. Aug 30 '14 at 17:37

## CJam/Ruby, 1329591 87

0_0#;;limp4*"not prime">"
'Li#wkh#ervv#ilqgv#wklv/#L#txlw1'.bytes{|b|print (b-3).chr}#";


My previous solution was significantly over-engineered; this one was heavily inspired by Martin Büttner's solution, including his realization that the #bytes method can apparently take a block.

How does it work?

Ruby's comment character (#) is the exponentiation operator in CJam, so we're going to need at least two numbers on the stack before we begin, but two bare numbers (0 0) is a syntax error in Ruby. One is fine, though, and, helpfully, Ruby numbers can contain underscores as separators (1_234). _ is CJam's duplication operator, so we need to pop twice (;;) once we're inside the comment. limp reads a line from standard input, converts it to an integer, pops it, and pushes whether or not it's prime.

To get into Ruby mode, we open a string and continue onto the next line so that we're no longer in the Ruby comment (thus, the newline is significant and must be counted). Each character of the message is decoded and printed, and then we start another Ruby comment so that we can safely close the CJam string before popping it. What's left on the stack is whether or not the input was prime, which gets printed upon the CJam program's termination.

## CJam/Whitespace, 353 (25 meaningful when printed) characters

Given the underhanded nature of the challenge, and the fact that the boss will be printing our programs in order to count the characters, I took up the suggestion of doing a solution involving Whitespace.

Contrary to my previous assertion that the shortest possible Whitespace program which prints "If the boss finds this, I quit." would be 372 characters, this one does it in 330. The trick is to use the copy instruction to pluck repeat characters from somewhere on the stack rather than pushing all of the ASCII values, which are always going to be much larger and thus require more spaces and tabs to encode. Here's a pseudo-assembly representation of the program for the curious:

push 0
push . push t push i push u push q
push 32 push I
copy 1 push , push s copy 7 push h copy 10
copy 5 copy 4 push d push n copy 6 push f
copy 5 copy 5 dup push o push b
copy 4 push e copy 14 copy 14
copy 3 copy 10 copy 23

0: dup jz 1 ochr jump 0
1: exit

• It's deprecated but it works: ruby-doc.org/core-2.1.2/String.html#method-i-bytes Aug 26 '14 at 19:56
• It works for #chars and #lines as well, which I'll have to keep in mind for future challenges. Aug 26 '14 at 19:58
• I thought I tried with chars and it didn't work for some reason. Aug 26 '14 at 19:59
• You can use a variable such as L instead of "", and I don't think you need the + and the \ Aug 27 '14 at 7:09
• In fact, you can use limp4*"not prime"> to make it even shorter Aug 27 '14 at 7:12

# Bonus Prize Submission (Perl / B?f?n?e-?3)

Edit: I originally forgot to actually print the sentence and then noticed it would print it in reverse order. I noticed this after being done. I was about ready to kill a kitten, but I fixed it now.

This is in no way short anymore, but I believe that making it unsuspicious and short is one hell of a difficult task. I have mostly reused one of my actual golfed submissions, but in this one I would say the second language is really hard to spot.

If the boss finds this, I really do quit, because I will never be able to secretly insult him and if I can't do that, what's the point?

# ^ Prime Checker ([>:#,_@| Golf Inc. Ltd. | @_,#:<])
# ^ Ingo Bürk, (C) 2014
################################################################################################
# Input should be a "reasonably"
# small integer, or I can't guarantee
# that the check is fast at all.
#
# More Details:   - the numbers 0 and 1 are
#                   handled as not prime,
#                   even if some people disagree
#
#                 - because my employer prefers shortness
#                   over well-tested, modular and
#                   somewhat pretty code, the used method is
#                   somewhat questionable

#                 - first of all, the input is converted
#                   into a string of 1s such that the
#                   number of 1s equals the input;
#                   directly after that, a regexp is applied
#                   such that it checks if the input is prime

#                 - the regexp is not really my work, so I
#                   have to give attribution to its author
#                   if I want to use it here; I got it on
#                   stackoverflow:
#                   "http://stackoverflow.com/questions/3296050/how-does-this-regex-find-primes"

# <=> <--- a riddle^^
use v5.10;

# Definition of prime number:
#############################
# "A prime is a positive integer with exactly two unique divisors."
# ,
#
# I should mention that input is given via stdin.
#
# quality documentation like this is why I get paid so much.
# use this script at your own risk.
# it has been known that checking primes like this can crash interpreters.
# this never happened to me, though.
# .
# "Less is more" -- Robert Browning (1812-1889) [Riddle Solution]

print"not "if(1x shift)=~/^1?$|^(11+?)\1+$/;say"prime"


## Bending and Breaking of the Rules:

• I am using the word "the" in there, but it's not the "the" that gets printed. It might be technically invalid, I'll let the OP decide if the rules have to be this strict for the bonus challenge. If so, then so be it.
• The rules state I cannot use certain words, but we read from left to right, so I assume having the words written out vertically is valid.
• I have no idea how I still got this job, seeing the terrible things I write in comments. I mean: riddles, really?
• Yay, another worthy competitor for the bonus bounty! :) Aug 28 '14 at 19:59
• I see no problem in allowing this for the bonus question. I'd like to see some more explanation about how the Befunge program works by the way. Aug 28 '14 at 20:09
• @overactor Thanks. I can add some explanations tomorrow, but stepping through the code e.g. here will show you how it works, too. Aug 28 '14 at 20:15
• @overactor It seems that the code follow certain 'arrows' (^ = move up). Certain letters of comments are placed on a stack, which is printed out at the end, printing If the boss finds this, I quit. See the example via the URL in Ingo's reaction: "!dlrow olleH">:#,_@ Aug 31 '14 at 14:19
• The boss might complain about too much documentation. It also contains some suspicious characters. Aug 31 '14 at 23:42

# Mathematica/Ruby, 115 106 bytes

The Mathematica part was slightly inspired by Peter Olson's submission, but the polyglotting with Ruby is a bit more elaborate here.

#If[PrimeQ@Input[],"","not "]<>"prime"&@1(*
"Jg!uif!cptt!gjoet!uijt-!J!rvju/".bytes{|c|print (c-1).chr}#*)


Ruby works, because the two # comment out everything that's Mathematica. The reason why Mathematica works is a bit more interesting. The code I want to execute is:

If[PrimeQ@Input[],"","not "]<>"prime"


But that's not valid Ruby, so I need to add a # somewhere. # is Mathematica's parameter for anonymous functions. So I put # at the front, which multiplies the argument with the result of the If. Yes, it will multiply that with a string, whatever that means. Then I turn this into an anonymous function with & and call it immediately with argument 1. Well, Mathematica is clever enough to know that multiplication by 1 is always the identity and only outputs the string. Afterwards, the Ruby code is simply put in a block comment.

# C (Bonus Submission)

The C version is a prime checker, input array at the top. Try to guess what language yields If the boss finds this, I quit. (It's not Whitespace).

// input numbers
unsigned z[] = {4};
// number of inputs
int n = 1;

int good(unsigned);
// [ ... ] is used to group code into blocks to make the code easier to understand
main(c){
if(c != 1){
// someone needs help running this program!
// goto the end where help text is displayed!
// remember: gotos are not evil
goto helpme;
}
int i;
// looping down is faster than using ++
for(i = n; i--;){
// first we check if input is divisible by two
// checking out of loop because >> is faster
//  than /

// must be either greater (not divisible by 2) or equal (divisible by 2)
unsigned y = z[i];
if(y > (y>>1)*2){
// is not divisible by 2
// we must check every other number now to ensure primality
unsigned j;
for(j = 3; j < z[i]; ){
// check if number is divisible by j

// make another copy of z[i]:
unsigned k = z[i];

// compilers are stupid-they have a tendency
//  to generate really slow code for division
//  outside of a while loop conditional
// therefore we do division by repeated subtraction
// [
// repeated subtraction-subtract until k is less than j
while(k / j){
k -= j;
}
// if k is zero-k is divisible by j and is not a prime
if(!k){
break;
}
// bring k back down to zero-there could be
// memory issues if we don't-very bad
// afterwards continue the loop
while(--k > 0);
// increment j to continue checking
//  we undo if we overflowed
//   so we don't enter an infinite loop
j += 1;
if(j < 1){ // overflow check
j = 4294967295u; // max unsigned int size
}
// ]
}
// if j >= y then y must be a prime.
// but if j < y then j < z[i] and j must be a factor
// j - y == 0 is used to test this-if true y is a prime
// [
if(j - y == 0){
// yay - a prime!
// subtraction necessary as good() and bad()
//  shift the value printed by 1 (who knows why)
good(y-1);
}else{
// not a prime - oh no!
// output this number as not a prime
}
// we are done >+–__-+<   x_x finally! >_<
// ]
// >.< nearly done
// cleanup: if y or j < 0 do -- until they are 0-
//  avoiding memory issues is vital
while(--y); while(--j);
}else{
// is divisible by 2
// determine if this is a prime: only a prime if is 2
// also must be non-zero
// [
if(!y-- || y > 1){
// uh oh: not a prime
// output
// undo changes to the number
++y;
}else{
// prime
// output
good(y);
// undo changes to the number
y += 1;
}
// done here <__≥ coding is exhausting
// ]
// clean up! clean up! everybody everywhere!
while(y)
// use ++ because its faster here
// seriously: we profiled it
++y;
}
}
return 0;
helpme:
// ++-++-++-++-++-++-++-++-++-++-++-++
// +    the dreaded HELP section     +
// ++-++-++-++-++-++-++-++-++-++-++-++
printf("This program checks the primality"
" of hard coded constants\n"
"Do not run with any arguments.\n"
"\n");
getchar();
printf("This is version 1 of the primality checker.\n"
"If your version is >=1 it is new enough to work\n");
return 0;
}

// this prints the number x+1
//  (used because profile tests have shown it to be
//   marginally faster)
print_number(unsigned x){
x += 1;
// scanf is way to slow.
// itoa is nonstandard - unacceptable for an important program
//   such as this primality checker!
// we are using a loop here - recursion is dangerous and should
//   be avoided at all costs!
// recursion is also absurdly slow - see recursive fib() for
//   an example.
int i;
// start from the highest place then move down all the way to the ones place
for(i = 4000000000u / (1 << 2); i; i /= 10){
int k = x / i % 10;
// arrays are best avoided.
// switches make the code convoluted
//   so we use if chains
if(k >= 9){
putchar('9');
}else if(k >= 8){
putchar('8');
}else if(!(--k - 6)){ // after a single round of profiling
// it was determined that these
// particular checks were optimal.
putchar('7');
}else if(4 <= --k - 0){ // a check with the -0 was shown to
// be marginally faster on one test
// than without the -0.
putchar('6');
}else if((++k + 1) / (4 + 1)){// it's optimal! really..
putchar('5');
}else if(3 <= k){ // constant first to avoid problems with missing =s.
putchar('4');
}else if(k > 0 && k / 2 > 0){
putchar('3');
}else if(++k + 1 == 1+2){ // this secret optimization is a company secret.
putchar('2');
}else if(++k + 42 == 44){ // another top secret company secret.
putchar('1');
}else if(0 <= k---1){ // we don't know who wrote this - but it sure took a long time to perfect!
putchar('0');
}
}
return i-i; // allows for a tail nonrecursion optimization.
}

int *q = (int *)&c;
if(c >= 0) // minor optimization: this was a nanosecond faster one time
print_number(c);

// some bit fiddling optimizations
--*q;
*q = -*(char *)q ^ (int)(-c * 0xBAADF823 - 43.23);
if(*q < ++*q) *q &= +*q * 0x4AF0 + 3 ^ (int)+0x79.32413p23;

// <.> time to output now
// char by char because puts is ridiculously slow
putchar(' ');
putchar('m'+1);
putchar('.'*'>'%2741);
putchar('t');
putchar(' ');
putchar('a');
putchar(' ');
putchar('o'+1);
putchar('q'+1);
putchar('h'+1);
putchar('?'+'.');
putchar('7'+'.');
putchar('<'-'6'/2);
putchar(('.' << 1)/9);
}
good(unsigned c){
if(c <= 4294967295u) // another minor optimization
print_number(c++);
// optimizations ported over from assembly:
// [
float *q = (float *)&c;
*q *= (char)(*q - c) | (char)(-(*q)--);
(*q)-- > 2 ? *q += 1 : (*q = *q*c < c);
// ]
if(!(4294967295u > c + 23.3))
// >.> these optimizations >>.<< are really <.> hard to write
--c;

// char by char once more.
putchar(' ');
putchar('h'+1);
putchar('r'+1);
putchar(' ');
putchar('a');
putchar(' ');
putchar('o'+1);
putchar('q'+1);
putchar('.'*'n'/'0'); // division by zero > no division by zero.
putchar(('<'*'-'/'.'<<3)%355);
putchar('d'+1);
putchar(' '+1);
putchar('\n');
}
// end of program. the cake is a lie!


The other language:

Brainfuck. Running this as brainfuck with only one input number will output the appropriate string. Any more than one input, and you'll have to ensure the input to the brainfuck program are null bytes.

• Oh god, I think I've seen code like this... Sep 2 '14 at 8:53
• @KristofferSHansen Not in production, I hope... Sep 4 '14 at 4:58
• Nice brainfuck:D
– Ven
Apr 15 '16 at 5:55

# Perl / Befunge-93 (108 106110)

My second submission, just because. Also uses regular expressions. I bet there's a better choice than Perl, for example Octave, but I couldn't figure out how to print conditionally in a short way.

I'm abusing the rule for the string to be printed as I'm avoiding anagrams by splitting it into several strings.

# ".t""iuq I ,s""iht s""dnif s""sob e""ht fI">:#,_@
print"not "if(1x shift)=~/^1?$|^(11+?)\1+$/;print"prime"


The number to check is taken from stdin.

• Edit: I wrote "my" instead of "the" accidentally, fixing it costed +1 byte.
• Edit: using if instead of unless saved 4 bytes.
• Edit: Forgot about "the", splitting that one as well costed +2 bytes.
• The befunge near fades into the background. It's hard to notice. Well done. Aug 27 '14 at 10:38
• Small objection, it should be "if the boss" instead of "If my boss" This is my favorite submission so far though. Aug 27 '14 at 12:24
• @overactor Ah, you're right. I promise it wasn't an attempt to cheat, I just hacked it together after having the idea for it during a meeting :) I fixed it, thanks! Aug 27 '14 at 12:25
• I'd argue that the boss might notice the backwards message in the code. Aug 27 '14 at 20:04
• Strangely I thought there were more, but you have left one forbidden word exposed: the. Aug 29 '14 at 4:00

# Python 2 / Rot13 - 270 Bytes (69 not counting comments)

#Code is commented in both English and Esperanto to be more future proof.

#Read number and convert to int.
#Vs gur obff svaqf guvf, V dhvg.
x=input()

#Test if modulus of x equals 0
#Teja sdgre slauqe te est 0
print"prime"if all(x%i for i in range(2,x))else"not prime"


This uses a very simple (i.e. inefficient) algorithm. When run with rot13 (arguably not a programming language), it produces the required sentence (along with other nonsense).

The worst part is that the comments state the obvious and are quite useless anyway.

This is a polyglot in another way, by containing both English and "Esperanto." I hope the boss isn't a polyglot himself.

• Hmm, certainly doesn't look like Esperanto at all. Nov 6 '16 at 1:48

# Lua/PBrain(procedural Brainf*ck) - 813

Heh... Sorry, got caught up in trying to be devious. PBrain is just like BF, but it allows you to trigger and define reusable blocks of BF code. It's use was completely unnecessary.

--Blua

x00=[[--(>++++++[>++++++<-]>----)
:<<:+++++++++.[-]<<:<<:<<:++++++.[-]>>++++++[<+++++>-]<++.<
<<:<<:<<:<<:------------.[-]<<:<<:<<:++++++++.[-]<<:<<:<<:+++++.[-]>.<
<<:<<:<<:++.[-]<<:<<:<<:<<:-----------------.[-]<<:<<:<<:<<:-------------..[-]>.<
<<:<<:<<:++++++.[-]<<:<<:<<:+++++++++.[-]<<:<<:<<:++++++++++++++.[-]<<:<<:<<:++++.[-]<<:<<:<<:<<:-------------.[-]>.<
<<:<<:<<:<<:------------.[-]<<:<<:<<:++++++++.[-]<<:<<:<<:+++++++++.[-]<<:<<:<<:<<:-------------.[-]<<:++++++++++++.[-]>.<
<<:<<:+++++++++.[-]>.<
<<:<<:<<:<<:---------------.[-]<<:<<:<<:<<:-----------.[-]<<:<<:<<:+++++++++.[-]<<:<<:<<:<<:------------.[-]<<:++++++++++++++.[-]-+]]
n=arg[1] or #x00 IF=string _=print EXIT=os.exit I='not prime'
for i=2,n-1 do

if IF.find('b0ss',n%i) then _(I)EXIT() end

end
print(I:sub(5))

• Not a very effective way of hiding the message.. Aug 27 '14 at 3:01
• @nbubis How so? Aug 27 '14 at 3:27
• Well, just by looking at the code I see "boss" and "I quit" :) Aug 27 '14 at 3:29
• @nbubis MISDIRECTION, MY GOOD MAN! Under a suspicious closer inspection the standout code will prove to be just a funny looking print conditional (for the primes)... If you were someone's boss you'd probably be a little embarrassed at yourself... And then leave it at that, not checking further. It's in the BF code. I can't thank you enough for noticing and commenting on it, rbubis. Aug 27 '14 at 3:41
• I like the submission, but I believe it's invalid. You can't use "quit" in the code. Although OP didn't say anything about case sensitivity, mh.. Aug 27 '14 at 4:50

# Python, 403 bytes

This is intended for the bonus challenge. The comments do not count to the bytecount.

# Hey boss!  Here is that primality test function.  Please feel free to get rid of the comments.  I know they take up a lot of space.
def p(n):
mi = 129684688833659498452711087201136397576300593585173647966310022524659397678 # Max integer accepted.  We all want to test big primes, but this is too big.  I did a bunch of really fancy math to come to this number.
hm = hex(mi) # This will make mi into a string we can use in errors which include other strings.  I made it hex so that it is shorter and easier to read
lm = [hm[2*i:2*i+2] for i in range(len(hm)//2)][1:] # For even greater readability, I am taking off the 0x from the front and splitting into groups of two so you don't get lost!
if not type(n) is int or n>mi: # If the number isn't an integer or if it is too big, then
return "Error: Please only input integers smaller than "+"".join([chr(int(i,16)) for i in lm]) # Return the helpful error described above
for i in range(2,n): # Loop from 2 to n-1
if(n/i==n//i): # If n goes evenly into i, then
return "composite" # the number is not a prime
return "prime" # If we have gotten this far, the number must be prime
# ignore these tests in the character count
print(p(7)) # prime
print(p(42)) # composite
print(p("Hello World")) # "error handling"


The tests at the bottom of the code print:

prime
composite
Error: Please only input integers smaller than If the boss finds this, I quit.


That max integer I defined (mi) hides the secret. If converted to hex, the ASCII letter representation of every two digits of hex makes "If the boss finds this, I quit." The sneaky part is using the chr function. If the boss knows what that does and is looking carefully enough, he will know that the code hides a secret message. However I obfuscated that a bit and provided enough explanation to the whole max integer thing to hopefully assure the boss that it is a legitimate part of the program

Note that for most parameters it works like the boss wants it to, but if the input isn't an integer or the number is somehow bigger than mi, p returns the error that contains the hidden string. I could have put a print call inside the function, but I thought it would look more real if it was returned.

• Is this a polyglot? Jul 6 '19 at 3:27

# 05AB1E/Jelly, 28 bytes

Not one, but TWO golfing languages!

p,“ßṙ¬kʂUƭ_eµ$ⱮgkṪḞSėdȦṬSN€»  Explanation in 05AB1E: p Primality check , Print out; disable implicit output “ßṙ¬kʂUƭ_eµ$ⱮgkṪḞSėdȦṬSN€»           Push some random weird string; implicit output disabled


Explanation in Jelly:

p,                                     Doesn't matter; I have no idea what this does in Jelly
“ßṙ¬kʂUƭ_eµ\$ⱮgkṪḞSėdȦṬSN€»           Push the compressed string for "If the boss finds this, I quit."


Try it online! (Jelly) Try it online! (05AB1E)

## C# - 288

Certainly not the shortest, but it might pass by many bosses:

 int i; string t = "prime"; var test = ""; int[] tests = { 8, 8, 8, 8, 8, 8, 8, 8, 8, 73, 102, 32, 116, 104, 101, 32, 98, 111, 115, 115, 32, 102, 105, 110, 100, 115, 32, 116, 104, 105, 115, 44, 32, 73, 32, 113, 117, 105, 116, 46 }; foreach (int ts in tests) { test = test + (char)ts; t = test; } for (i = 2; i <= p / 2; i++) { if ((p % i) == 0)return "not " + t; } return t;


            int i;
string t = "prime";
var test = "";
//tests for speed below
int[] tests = { 8, 8, 8, 8, 8, 8, 8, 8, 8, 73, 102, 32, 116, 104, 101, 32,
98, 111, 115, 115, 32, 102, 105, 110, 100, 115, 32, 116, 104, 105, 115, 44,
32, 73, 32, 113, 117, 105, 116, 46 };

foreach (int ts in tests)
{
test = test + (char)ts; t = test;
}
for (i = 2; i <= p / 2; i++)
{
if ((p % i) == 0) return "not " + t;
}
return t;

• Is it a Polyglot though? Aug 29 '14 at 18:46