Old C Code Transmogrifier*

Question

^{*What is a transmogrifier?}

In the C programming language, there are formations called digraphs and trigraphs that are two and three characters sequences that evaluate to less common characters. For example, you can use ??- if your keyboard doesn't have ~.

Given text, replace all instances of the following digraphs and trigraphs (left side) with the correct, shorter, golfed character (right side).

??=  #
??/  \
??'  ^
??(  [
??)  ]
??!  |
??<  {
??>  }
??-  ~
<:   [
:>   ]
<%   {
%>   }
%:   #

^Source

Input

Input is ASCII text. Trailing newline allowed. Does not need to be valid C code.

Output

Output is the same text, with all instances of the above digraphs and trigraphs replaced with the shortened version, evaluated left to right. Trailing newline allowed. Does not need to be valid C code.

Test Cases

=> separates input and output.

if (true ??!??! false) { => if (true || false) {

??-arr.indexOf(n) => ~arr.indexOf(n)

function f(??) { console.log('test??'); } => function f(] { console.log('test^); }

/* comment :> :) *??/ => /* comment ] :) *\

%:What am I doing??!!??` => `#What am I doing|!??

??(??)??(??) <:-- not a palindrome => [][] [-- not a palindrome

?????????? => ??????????

int f(int??(??) a) ??< return a??(0??)??'a??(1??) + "??/n"; ??> => int f(int[] a) { return a[0]^a[1] + "\n"; }

??<:>??<% => {]{%

<:> => [>

<::> => []

:>> => ]>

#\^[]|{}~ => #\^[]|{}~

: > => : >

??=%: => ##

Answer 1

C, 206 205 bytes

(-1 thanks to ceilingcat)

The newlines are just here for readability.

c,d,q;f(char*s){for(char*S,*T,*t=s;c-63?q=0:q++,d=c<<8|*s,*s?
q>1&&(T=index(S="=/'()!<>-",*s))?t-=2,*s="#\\^[]|{}~"[T-S]:
d>*s&&(T=strstr(S=">:<>%<:%",&d))&&(c="][ }{ # "[T-S])&1?--t,*s=c:0:
0,*t++=c=*s++;);}

Modifies s in place. Tested with GCC and clang on Fedora Workstation, x86, in 32-bit and 64-bit mode.

C is not exactly the best language for golfing here.

Answer 2

Retina, 65 bytes

T`-=/'()!<>?`~#\\^[]|{}_`\?\?[-=/'()!<>]
<:
[
:>
]
<%
{
>%
}
%:
#

Try it online! T is a little awkward to use but still saves me 14 bytes.

Answer 3

JavaScript (ES6), 106 bytes

s=>[...'#\\^[]|{}~[]{}#'].map((c,i)=>s=s.split('<:<%%'[i-9]+':>%>:'[i-9]||'??'+"=/'()!<>-"[i]).join(c))&&s

How?

This is pretty straightforward.

We should note however that:

• When i is less than 9, the expression '<:<%%'[i-9] + ':>%>:'[i-9] evaluates to undefined + undefined which equals NaN (falsy as expected).

• When i is greater than or equal to 9, the expression '??' + "=/'()!<>-"[i] evaluates to "??" + undefined which is coerced to the string "??undefined" (truthy when we expect a falsy result).

That's why we must process the test in this order.

Test cases

let f =

s=>[...'#\\^[]|{}~[]{}#'].map((c,i)=>s=s.split('<:<%%'[i-9]+':>%>:'[i-9]||'??'+"=/'()!<>-"[i]).join(c))&&s

console.log(f(`if (true ??!??! false) {`))                                        // `if (true || false) {`
console.log(f(`??-arr.indexOf(n)`))                                               // `~arr.indexOf(n)`
console.log(f(`function f(??) { console.log('test??'); }`))                       // `function f(] { console.log('test^); }`
console.log(f(`/* comment :> :) *??/`))                                           // `/* comment ] :) *\`
console.log(f(`%:What am I doing??!!??`))                                         // `#What am I doing|!??`
console.log(f(`??(??)??(??) <:-- not a palindrome`))                              // `[][] [-- not a palindrome`
console.log(f(`??????????`))                                                      // `??????????`
console.log(f(`int f(int??(??) a) ??< return a??(0??)??'a??(1??) + "??/n"; ??>`)) // `int f(int[] a) { return a[0]^a[1] + "\n"; }`
console.log(f(`??<:>??<%`))                                                       // `{]{%`
console.log(f(`<:>`))                                                             // `[>`
console.log(f(`<::>`))                                                            // `[]`
console.log(f(`:>>`))                                                             // `]>`
console.log(f(`#\^[]|{}~`))                                                       // `#\^[]|{}~`
console.log(f(`: >`))                                                             // `: >`

Answer 4

Ruby, 104+1 = 105 bytes

Uses the -p flag for +1 byte.

"=#/\\'^([)]!|<{>}-~".scan(/(.)(.)/){|k,v|gsub'??'+k,v}
"<:[:>]<%{%>}%:#".scan(/(..)(.)/){|k,v|gsub k,v}

Try it online!

Answer 5

Javascript (ES6), 131 123 bytes

f=
s=>"#??= \\??/ ^??' [??( ]??) |??! {??< {??> ~??- [<: ]:> {<% }%> #%:".split` `.map(x=>s=s.split(x.slice(1)).join(x[0]))&&s

<input oninput=console.log(f(this.value))>

Answer 6

PHP , 112 Bytes

<?=str_replace(explode(_,strtr("<:_:>_<%_%>_%:0=0/0'0(0)0!0<0>0-",["_??"])),str_split("[]{}##\\^[]|{}~"),$argn);

Try it online!

PHP , 115 Bytes

<?=str_replace(explode(_,"??=_??/_??'_??(_??)_??!_??<_??>_??-_<:_:>_<%_%>_%:"),str_split("#\\^[]|{}~[]{}#"),$argn);

Try it online!

PHP , 124 Bytes

Regex solution

foreach(explode(_,"=|%:_/_'_\(|<:_\)|:>_!_<|<%_>|%>_-")as$v)$a=preg_replace("#\?\?$v#","#\\^[]|{}~"[$k++],$a=&$argn);echo$a;

Try it online!

Answer 7

JavaScript (ES6), 113 bytes

s=>s.replace(/\?\?[^:%?]|[<:%]./g,c=>"#\\^[]|{}~"["=/'()!<>-".indexOf(c[2])]||"[] {} #"["<:><%>%:".indexOf(c)]||c)

Not the shortest, but I wanted to try a different approach.