Count like Chuck Norris

Python 2/C (clang), 109 107 100 84 95 88 89 88 87 84 bytes

i=0;
#/*
while 1:i+=1L;print`i`[::-1]
'''*/
a(){for(;;)printf("%i %1$i ",++i);}//'''

Python: Try it online!

C: Try it online!

The L in the Python code is part of the delimiter.

Explanation:

In the C code, It first sets i to 0. Then, it starts a comment (# is valid code in C for #include statements) where the Python code goes. At the end of the comment, it defines a function that forever increments a variable and prints it twice, space-delimited. It then begins a comment.

In the Python code, i=0; sets i to zero. Python ignores the next line because # starts a single-line comment. It then forever increments it and turns it into a long a number and prints its reversed string representation. The 'L' from the long is part of the delimiter. After that, it starts a multi-line string to comment the C code, that ends later.

 

-2 bytes thanks to @LuisMendo. -7 bytes thanks to @ZacharyT. -6 more bytes thanks to @ZacharyT. +11 bytes to fix a bug thanks to @mbomb007. -7 bytes thanks to @Doorknob. +1 byte to fix a bug thanks to @Doorknob. -1 byte thanks to @yoann. -1 more byte thanks to @yoann. -3 bytes thanks to @Cyoce.


05AB1E / Jelly,  14  13 bytes

-1 byte thanks to Adnan (avoid triplicate with non-popping print)

Raw bytes (hexadecimal):

31 5b 3d 3d 3e 5d fc 06 b6 3b 87 08 15

In 05AB1E's code-page:

1[==>]üε¶;‡ηΩ

Try it online!

In Jelly's code-page:

1[==>]‘©Ṛ;⁷®ß

Try it online!

How?

The 05AB1E program prints the double count with each entry separated by newlines:

1[==>]üε¶;‡ηΩ
1             - push 1 onto the stack
 [            - start infinite loop:
  =           - print top of stack
   =          - print top of stack
    >         -   increment the top of the stack (pop(1), x+=1, push)
     ]        - end the infinite loop
      üε¶;‡ηΩ - this code is never executed nor is it parsed

The Jelly program prints the reversed count with each entry separated by newlines.

The parser will treat a valid literal between [ and ] as the enclosed literal, otherwise these bytes are undefined tokens and as such become equivalent to tokens separating the code into lines. ==> does not parse as a literal, so the code is effectively:

1 - link 1 (not executed but parsed)
1 - literal 1

==> - link 2 (not executed but parsed)
=   - left equals right? (vectorises)
 =  - left equals right? (vectorises)
  > - is greater than the right argument? (vectorises)

‘©Ṛ;⁷®ß - Main link: no arguments
‘       - increment (implicit 0 on left increments to 1 on the first pass)
 ©      - copy the result to the register and yield it
  Ṛ     - reverse, this has an implicit decimal list build, e.g. 142 -> [2,4,1]
    ⁷   - a newline character
   ;    - concatenate, e.g. [2,4,1] -> [2,4,1,'\n']
     ®  - recall the number from the register, e.g. 142
        - The chain has reduced to one item of arity 0, causing a pre-evaluation print:
        -     ...and since the list contains a character it gets smashed together
        -     e.g. [2,4,1,'\n'] prints 241 followed by a newline character
      ß - call this link with the same arity, e.g. as a monad with left = 142

Jelly/Pyth, 15 bytes

.V1_`b;"1üÉÉ$

Unprintables are mangled by the SE software, so here's a hexdump:

00000000: 2e56 315f 6062 3b22 7f31 fcc9 c924 0b    .V1_`b;".1...$.

Run with jelly f file and pyth file respectively.

Explanation

First comes the Pyth part. .V runs an infinite loop over the incrementing sequence starting from its input, which is here 1. Then we reverse (_) the stringified (`) loop index (b) and implicitly output it. The ; is there to end the loop, and the " is necessary to treat the rest of the program as a string literal so that the parser doesn't choke on it.

The Jelly part will be explained by first translating the remainder of the program from the Jelly codepage:

¶1‘ṄṄ$¿

The acts as a line feed, effectively ignoring the first part of the program by making it a link that is never called. Then we start at 1 and run a while loop (¿) that uses ṄṄ$ (print twice) as its condition, and increments () the value as the loop body.


Incidentally, replacing the Pyth part with 1[DR,>] would create a valid Jelly/05AB1E submission in 14 bytes, but the current interpeter contains a bug that prevents this.