0xDEADBEEF is not tasty, but still 0xFEEDBEEF



Inspired by "You had me at hello"


For those who don't know, hexadecimal is a base system like decimal. Contrary to decimal however, it uses 16 (hence the name) digits instead of 10.

Here are the first few numbers while counting up in hexadecimal to maybe give you a feel of things, although you should probably google it for more info.

0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, F, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 1A, 1B, 1C, 1D, 1E, 1F, 20...

From now on I'll refer to hexadecimal as hex.


Take a possibly infinite stream of bytes, and convert it to hex. This should be done as you read since you might get an infinite stream and keep on reading forever.

The twist, is that if a sequence of converted numbers equals the exact hex number DEADBEEF (3735928559 in decimal), you must output FEEDBEEF (4276993775 in decimal) instead.

If no DEADBEEF is found, print the string as it was

Inter-character matches work too, so 2D EA DB EE F0 works and should be converted to 2F EE DB EE F


This might seem a bit vague, so here are 2 step-by-step examples.

Example 1

Input: Hello, World (bytes represented as string)

Step 1: Convert the input to ASCII: 072 101 108 108 111 044 032 087 111 114 108 100

Step 2: Convert the ASCII to hex: 48 65 6C 6C 6F 2C 20 57 6F 72 6C 64

Step 3: Print out everything until DEADBEEF is found. It wasn't found, so just print Hello, World again.

Example 2

Input: Hello,Þ­¾ï,,

Step 1: Convert to ASCII: 72 101 108 108 111 44 222 173 190 239 44 44

Step 2: Convert to hex: 48 65 6C 6C 6F 2C DE AD BE EF 2C 2C

Step 3: Is there DEADBEEF? Yep. so take everything up to it. 48 65 6C 6C 6F 2C

Step 4: Append FEEDBEEF to it. 48 65 6C 6C 6F 2C FE ED BE EF

Step 5: Convert back to a string and print it. Output is Hello,þí¾ï


Try to make a program that does the above in the fewest bytes of source code as possible.


Make sure your program works for finite and infinite streams. The stream can be STDIN, file, function argument (if you can pass streams as arguments), or your language's closest alternative. Output is to STDOUT or function returning (if you can return streams).

More test cases (contain invisible characters)

eeþí¾Þ­¾ïe (6565feedbedeadbeef65) -> eeþí¾þí¾ïe
Þ­¾ï (deadbeef)-> þí¾ï
Þ=êÛîôV -> Þ?îÛîôV


Posted 2018-10-30T20:03:18.060

Reputation: 857

9"Take a ... stream of printable ASCII characters". Þ from your 2nd example isn't a printable ascii character. – nimi – 2018-10-30T23:19:50.340


@Rogem Printable ascii is a defined range of code points

– Jo King – 2018-10-31T13:16:50.540

2If an inter-match is found, like input -êÛîð (2D EA DB EE F0), do we convert it to 2F EE DB EE F0 (resulting in string -êÛîð) or 2F EE DB EE F (resulting in string -êÛî where the last character is the unprintable \x0f)? – Kevin Cruijssen – 2018-10-31T14:29:00.963

Your test cases should be either pasted as TIO links or as hex dumps if they contain unprintable characters. – mbomb007 – 2018-11-01T21:28:11.423

@mbomb007 There are hexdumps of the inputs, the outputs don't contain unprintable characters – FireCubez – 2018-11-01T21:32:46.237

I would format those extra test cases to make that clearer. I think it'd be better to format your test cases as [hex dump] => [hex dump]. – mbomb007 – 2018-11-01T21:34:22.690



Japt, 34 30 bytes

-4 bytes from @Shagy

csG r`Ü%¼ef.+``fe‚¼ef` ò ®nG d

Try it online!

Luis felipe De jesus Munoz

Posted 2018-10-30T20:03:18.060

Reputation: 9 639

Added more test cases – FireCubez – 2018-11-01T17:34:00.757

130 bytes using S.c() instead of S.m() and S.r() instead of S.s(). – Shaggy – 2018-11-01T17:49:22.270

@Shaggy Damn, I dont know how I didnt see those methods :o – Luis felipe De jesus Munoz – 2018-11-01T17:58:51.810


Python 3, 121 bytes

while i:p(end=x[:-4]);x=x[-4:];x!='Þ­¾ï'or+p(end='þí¾ï');i=I(1);x+=i

Try it online!

Erik the Outgolfer

Posted 2018-10-30T20:03:18.060

Reputation: 38 134

2or+p(end=...) causes a TypeError: bad operand type for unary +: 'NoneType' – mypetlion – 2018-10-30T23:02:40.277

2@mypetlion The error is on purpose, it crashes the program so that it stops outputting. – Erik the Outgolfer – 2018-10-31T12:47:36.183

Oh yes that does make sense. I might have misread the spec. – mypetlion – 2018-10-31T15:52:41.160


Clean, 143 bytes

uses octal escapes here and on TIO to prevent UTF-8 mangling, but it does work with the actual values and so each escape sequence is counted as one byte

This one works with inter-character sequences.

import StdEnv
$l=:[a,'\352\333\356',b:t]|(toInt a)rem 16==13&&(toInt b)>>4==15=[a+'\002':['\356\333\356\015']]=[hd l: $(tl l)]
$[h:t]=[h: $t]

Try it online!

Clean, 38 bytes

uses octal escapes here and on TIO to prevent UTF-8 mangling, but it does work with the actual values and so each escape sequence is counted as one byte

This one does not work with inter-character sequences.

$[h:t]=[h: $t]

Try it online!

Defines the function $ :: [Char] -> [Char] taking and returning a list of characters.
Since Clean is lazily-evaluated, this is equivalent to a stream.


Posted 2018-10-30T20:03:18.060

Reputation: 7 916

Does it work with things like hex(0DEADBEEF9) – FireCubez – 2018-11-02T18:05:28.797

@FireCubez no, as that wasn't clearly required at the one I wrote this. I'll update it today so that it does. – Οurous – 2018-11-02T20:40:54.340

@FireCubez with interchacter matches, what is the second digit of the last F? Do we use 0x0F? 0xF0? – Οurous – 2018-11-03T00:23:49.297

@Ourous 0x0F would be used – FireCubez – 2018-11-03T00:31:39.113


05AB1E, 30 bytes


Try it online or verify a few more test cases.


Ç                               # Convert the (implicit) input-string to codepoint decimals
 h                              # Convert them to hexadecimal
  J                             # Join them together
   l                            # Convert to lowercase
    ’–ìÀÜ’                      # Push compressed string "deadbeef"
          ¡                     # Split the hexadecimal string by this "deadbeef"
           D                    # Duplicate
            н                   # Take the first item of the split-list
             s                  # Swap so the full list is at the top of the stack again
              g1›               # Check if its length is larger than 1
                 ’ŽåÀܒ׫       # If it is: concat "feedbeef" to the head-string
                         2ô     # Split the string into parts of size 2
                           H    # Convert hexadecimal back to decimal
                            ç   # Decimal back to characters
                             J  # Join the characters (and output implicitly)

See this 05AB1E tips of mine (section How to use the dictionary?) to understand why ’–ìÀÜ’ is "deadbeef" and ’ŽåÀÜ’ is "feedbeef".

Kevin Cruijssen

Posted 2018-10-30T20:03:18.060

Reputation: 67 575

Why does it stop output when reaching a DEADBEEF? – l4m2 – 2018-11-01T18:50:17.920

@l4m2 Because that is the challenge.. :) (I.e. in the second example, the two trailing ,, aren't output. It stops as soon as it found the DEADBEEF and appended the FEEDBEEF instead.) – Kevin Cruijssen – 2018-11-01T21:25:25.503

See. The question is inconsistant – l4m2 – 2018-11-02T02:47:27.417

@l4m2 I agree. Although the other answers thus far (Python, Japt, and Clean) interpret it in the same way as me. – Kevin Cruijssen – 2018-11-02T07:26:06.993


C (gcc), 137 bytes

i=16;n;long long u;c;R(k){u=u<<4|c>>k&15,u^=u+'!RA'<<32?0:' @';}main(){for(;n+=!!~(c=getchar(n-=!i&&~putchar(u>>32)));i/=2)R(4),R(0);}

Try it online!


Posted 2018-10-30T20:03:18.060

Reputation: 5 985

@ceilingcat c is used in R, so can't move – l4m2 – 2018-11-03T08:19:44.837