Show some of your favourite quirky code

[camjocotem]

Javascript:

typeof(NaN)  //"number"

[WereCatf]

C/C++:

uint16_t one = 65536;

 

[Smeltie]

my favorite piece of code i have ever seen was from quake:

float Q_rsqrt( float number )
{
	long i;
	float x2, y;
	const float threehalfs = 1.5F;

	x2 = number * 0.5F;
	y  = number;
	i  = * ( long * ) &y;                       // evil floating point bit level hacking
	i  = 0x5f3759df - ( i >> 1 );               // what the fuck? 
	y  = * ( float * ) &i;
	y  = y * ( threehalfs - ( x2 * y * y ) );   // 1st iteration
//	y  = y * ( threehalfs - ( x2 * y * y ) );   // 2nd iteration, this can be removed

	return y;
}

i like how even the developers where confused about it.

[reniat]

Array(16).join("wat" - 1) + " Batman!"

 

it's a reference to: https://www.destroyallsoftware.com/talks/wat

[Guest]

I once was reading some source code from an 3d shooter someone made and put online, and it checked if a bool was true using this:

bool bool1 = true;
if (bool1 == true && bool1 != false)
{
	
}

When I asked him why he checked the bool twice, he said:"This way I can know for sure if the bool is actually true, because it might be something else then false or true.".

I laughed so hard.

[WereCatf]

2 minutes ago, timl132 said:

When I asked him why he checked the bool twice, he said:"This way I can know for sure if the bool is actually true, because it might be something else then false or true.".

[Mira Yurizaki]

var test

if (test !== 0){
  console.log("Technically it's true, don'tcha know.")
}

This hurts me more often than I care to admit whenever I'm doing something in JavaScript.

[myselfolli]

36 minutes ago, reniat said:

Array(16).join("wat" - 1) + " Batman!"

 

it's a reference to: https://www.destroyallsoftware.com/talks/wat

That is awesome

[reniat]

Some of the code that comes from the The International Obfuscated C Code Contest (https://www.ioccc.org/) is also amazing.

 

For example, this is from this year's entries (0 idea how this will look on mobile):
 

#include<stdio.h>
#include<stdlib.h>
#define M malloc(sizeof(
#define R fread(x,sizeof(float),Z*w,f)
#define E free(

char*G=
"   "
"    '"
"@(*AP(  "
"             "
"                     "
"                              "
"                                        "
"                                                   _"
"0   #@ .  \"C^*/XH\"Q2_U(D&(5*E0C 4B54DV   X    \"!9A@   89H$ !"
"0(/@@4 @(/@@(   #@T  @(\" @(    P, , P# , #Y1244^   ?R 0#%)14,A#9)24$B 1"
"_)!0,$Y145%R Y145$^&!81T! #9)24DV#Y)24DP  ;&P     -C4 $$B% @ !04%!04 @4(D$ #!(1"
"4 @#U574$^'^0D)!_&Z1D9'_$*!@8%^'Z!@8'_(&1D9'_(\"0D)#_%Z1D8%^/\\0$!#_ \"!_X$ /X! 0"
"$&(-$*!#_ $! 0'_/] ,$#_/\\&&&#_'Z!@8%^&\"0D)#_'V\"A8%^&>8D)#_$Z1D9%B(\" _X\" /X"
"! 0'^/ , PSP/P#' /\\,<H$\"C', @'R# ,&AD8F'(&!_P   ,,,,     _X&!! @0\" 0 "
"$! 0$!  @0(   \\5%14\" X1$0G_!$1$1$./\\2$1$. T5%14. \"0D'\\ 'Y)"
"144X \\0$ C_  !7Q     7A$! $2# 3_ $!_H   \\0'Q ? \\0"
"$! ? X1$1$.#A$1\"1_'](1$0X @0$ @? (5%14)"
"  1$?X0!\\\" 0$>!@& 088!X!#@$>"
"!$*! H1#X)!04X!$9%1,1"
"(&!;A 0   _P "
" ! 0;H&!"
" @$\""
"! ("
"\0";                            int s,p,e,c,t,r,a,l;   /* l e a k a g e */
 
                                                     int
                                                   F(int o
                                                ,int n, int t
                                             ){char*p=(o*(136-o)
                                           >=1260)*7*(o-9)+G; for(
                                          n=!(o=-1); n=*p+++64*n-*G
                                        ,(o+=3)>>2<5-t; ); return 255
                                       &n>>2*o%8; } float L[0x400];int
                                     T(float a,float b,float l,float*e)
                                    {return L[0x3ff]||!(e[1]=b*a+l**e)||
                                   !T(l*a-b**e,b,l,e+1); } float f(int x)
                                  {int w=s<<!!T(1,6.135885e-3,-1.882472e-5
                                 +1,L); x=(x%w+w)%w<<9; return(L[x/s]*(s-x%
                                s)+L[x/s+1]*(x%s))/s; } int C(float*h,float*
                               i,int r,int P){ int x,y,c; for(y=0; h<i; ++y,h
                               +=r){ c=P<1; for(*h=x=0; x<8; c+=(1&P>>x&&P+(*
                              h+=f(e*(l+a*(2*x+1))*y)*f(s/2-e*a*y+e*a*y*y*(t/p
                             )/(s/p)))),++x); float z=f(y*t)*f(y*t); *h=z*(1+3*
                             z+2*z*z)/6**h/c; } return h-i-r+1; } void d(int w,
                            int n,FILE*f){int Z=s/t; int S[5]; float**P=M float*
                            )*c);for(r=0;r<c;++r){ P[r]=M float)*Z); C(P[r],P[r]
                           +Z,1,r); } float*x=M float)*Z*w); for (; ; ){ for(r=0;
                r<5        ; ++r){ int o=1; for(R,a=256; o&&--a; )for(o=t=0; t<Z;        o+=
               P[a][       t]!=x[w*t+n],++t); S[r]=a; } if(R-Z*w)break; for(r=a= -      12;r<
              5&&(a+=     a<11?22:1)<127; )for(r=-1; ++r<5&&F(a,1,r)==S[r]; );if(a     <127&&r
    ==5      )putchar(    a); } for(; --c; E P[c])); E P); fclose(f);}int main(int    q,char**v)      {s=
   44100     ;p=25; e=    55;c=256; t=75; r=q>1?v[1][0]*c+v[1][1]:0; a=7;l=16;if(r    ==11620){d     (atoi
  (v[2]),   atoi(v[3]),  q<5?stdin:fopen(v[4],"r")); return 0; } char*H=malloc(c),*  h=H; sprintf   (H,"\0"
 "Usage\n"  "    %s [-"  "h] [-d #c cid] file1 file2 ... (or stdin)"+(q-1&&r==11624  ),v[0]);if(*  H)q=1;int
 w=2>q?1:q -1; FILE**f=M FILE*)*w); f[0]=stdin; if(q>1)for(r=0; r<w; f[r]=fopen(v[r +1],"r"),++r); int l=s/t
 *w; float *b=M float)*l ),*t=b+l; int*i=M int)*w); for(; ; ){ int x=*h&&1>ungetc(* H++,*f); for(r =w; r--;)
{ i[r]=getc (f[r]); if(i[r]<0)i[r]=!++x; } if(x==w)break; while(++r<6){ while(C(b++,t,w,F(*i++,r< 5,r))); i-=
w; fwrite(b-=w,sizeof(float),l,stdout); } fflush(stdout); } while(fclose(f[--w]),w); E f); E b); E i); E h);}

 

And yes, that code compiles and does something

[camjocotem]

1 hour ago, James Evens said:

for(;;)

{

}

It's like it's saying "for nothing, do nothing, until never and never do it again"

[camjocotem]

function TheQuestion(){
    var 2B;
    return 2B || !2B;
}

[reniat]

9 minutes ago, camjocotem said:

function TheQuestion(){
    var 2B;
    return 2B || !2B;
}

so the answer to that question is just  "yes"?

[camjocotem]

Just now, reniat said:

so the answer to that question is just  "yes"?

Code gives Shakespere his answer

[Mira Yurizaki]

Not really a quirk, but a favorite of mine is the use of gotos in the Linux kernel code. You know that one "rule" that says never use gotos? Linux has examples of where it's actually a useful thing to have.

 

6 hours ago, WereCatf said:

C/C++:

uint16_t one = 65536;

 

Doesn't this equal 0 (assuming it compiles at all)?

[WereCatf]

10 minutes ago, M.Yurizaki said:

Doesn't this equal 0 (assuming it compiles at all)?

No, it equals 1. A 16-bit unsigned integer can have a value of 0 to 65535 and trying to assign more than that into it will cause it to overflow, discarding the 65535 from the value, ie. if you discard 65535 from 65536 you end up with 1.

 

EDIT: I just tested and apparently it actually doesn't work in GCC. It does work in some compilers and the line is actually from Elder Scrolls: Morrowind's code, so it obviously worked for them.

[Mira Yurizaki]

4 minutes ago, WereCatf said:

No, it equals 1. A 16-bit unsigned integer can have a value of 0 to 65535 and trying to assign more than that into it will cause it to overflow, discarding the 65535 from the value, ie. if you discard 65535 from 65536 you end up with 1.

 

EDIT: I just tested and apparently it actually doesn't work in GCC. It does work in some compilers and the line is actually from Elder Scrolls: Morrowind's code, so it obviously worked for them.

I compiled it with GCC using no arguments and the printout said 0.

 

Actually now that I think about it more, it probably depends on the machine. 65536 can be stored either as 0x01 00 00 or 0x00 00 01. Either way, the left most side is likely going to get chopped off.

[Unimportant]

13 hours ago, M.Yurizaki said:

I compiled it with GCC using no arguments and the printout said 0.

 

Actually now that I think about it more, it probably depends on the machine. 65536 can be stored either as 0x01 00 00 or 0x00 00 01. Either way, the left most side is likely going to get chopped off.

 

The literal value "65536" is of the integer type (assuming int is wide enough to hold the value). So this is a narrowing signed to unsigned conversion.

 

The standard says:

Quote

When a value with integer type is converted to another integer type other than _Bool , if the value can be represented by the new type, it is unchanged...

Otherwise, if the new type is unsigned, the value is converted by repeatedly adding or subtracting one more than the maximum value that can be represented in the new type until the value is in the range of the new type.

K&R used a different definition which boils down to the same thing:

Quote

Any integer is converted to a given unsigned type by finding the smallest non negative value that is congruent to that integer,modulo one more than the largest value that can be represented in the unsigned type.

So the answer should always be 0.

 

As a side note, even if bits were to be chopped off, endianness should not play a role. The high order bits are the high order bits, regardless of how the machine arranges them internally.