Forums



/* DISPLAY BACKGROUND in VIDEO MODE 3 --------------------------------------------


   I am now trying to beat Marat on making the ugliest code possible


   I hope to have a good chance to win with this one


   Anyone dare to beat me ? */





void Display_Background_3 (void)


{


 int x, y=0, z; byte *p_scr, *pattern=BACK_AREA, *tiles_data;


 for(;y<192;y+=32,pattern+=96) for(x=0;x<256;y-=32,x+=8)


  for(tiles_data=SG_BACK_TILE+(*pattern++<<3),z=0;z<32;z ++)


   {


   p_scr=screenbuffer->line[y+(z&3)]+x;


   p_scr[0]=p_scr[1]=p_scr[2]=p_scr[3]=*tiles_data>>4;


   p_scr[4]=p_scr[5]=p_scr[6]=p_scr[7]=*tiles_data&0x0F;


   if(!(z&3)){y+=4;tiles_data++;}


   }


}

Wow I'm quite proud of this one, it's unreadable - now I can release Meka sources safely ^_^

Zoop
(gotta rework a bit on it now else I won't even be able to maintain the code myself)

; division by 9 written in Z80 assembly code
; originally written by Ricardo Bittencourt
; as a part of a routine to read a sector from msx disk-drives
; enter HL = number from 0 a 1440
; exit A = HL/9
; destroy HL,DE
; clock cycles presented by Z80 and R800 processors
; this is the fastest way to divide a number by 9 in these processors
; without using look-up tables
; ten points to anyone who understand how it works

DIV9: ; Z80 R800
INC HL ; 7 1
LD D,H ; 5 1
LD E,L ; 5 1
ADD HL,HL ; 12 1
ADD HL,HL ; 12 1
ADD HL,HL ; 12 1
SBC HL,DE ; 17 2
LD E,0 ; 8 2
LD D,L ; 5 1
LD A,H ; 5 1
ADD HL,HL ; 12 1
ADD HL,HL ; 12 1
ADD HL,DE ; 12 1
ADC A,E ; 5 1
XOR H ; 5 1
AND 03FH ; 8 2
XOR H ; 5 1
RLCA ; 5 1
RLCA ; 5 1
RET ; total = 157 22

...but if you really wanna see some ugly code, look at the sprite engine in uNESsential... :)

How about this:

c = ap ? a ? a : b : bp ? b ? b : a : a ? a : b;

May not win for ugliness, but certainly could win a prize for
being impossible to maintain or decipher. ;)

It could be ten times more weird if I threw in an additional
set of tests, but I haven't even figured that out yet. :)

[ code example removed ]

Quote

> Wow I'm quite proud of this one, it's unreadable - now I can release Meka sources safely ^_^

That's nothing. Go find the first releases of DarcNES, and look at nes_blit.c. Aside from the unholy mixture of x86 machine code and C structures, it contains the following little gem:

/*
* there are 256 blitters. hence 256 headers and cleanup routines.
* for each pair of blitters x and y where (y = 255 - x), there are
* eight 1 bits and eight 0 bits. taking this data, we come up with
* the following calculation for how much memory we to allocate.
*
* this formula has since been modified for arbitrary length blitters and
* the overhead of clipping the image.
*/
#define SIZECALC(clipside) (((1 << width) * (sizeof(blit_hdr) + sizeof(blit_cleanup))) + ((1 << (width - 1)) * width * (sizeof(blit_0) + sizeof(blit_1))) + ((flags & clipside)? (3 * (1 << width)): 0));
if ((flags & BFL_REVERSED) == 0) {
i = SIZECALC(BFL_LEFTCLIP);
} else {
i = SIZECALC(BFL_RIGHTCLIP);
}
#undef SIZECALC

The function then proceeds to allocate i bytes of memory and, using memcpy() and some direct manipualation, fills the memory with procedures. A procpointer array is then filled with pointers to the procedures. This is done several times.

Of course, the entire approach was Just Plain Wrong, and the code has long since been scrapped. On the other paw, it's things like this that make me confident that I can write a dynarec CPU core whenever I feel like doing so. ^_^

And if you want more unreadable code, see the IOCCC.

Quote

> Zoop
> (gotta rework a bit on it now else I won't even be able to maintain the code myself)

--Nyef

IOCCC

I think you stretched a blood vessel in my brain. I feel so dizzy. Please, all of you should deserve
the trophy for the uglyest programming codes.

Chris :o)

Quote

> I think you stretched a blood vessel in my brain. I feel so dizzy. Please, all of you should deserve
> the trophy for the uglyest programming codes.

The one who thinks optimized code is ugly will never
reach the Zen of programming

(as said by Ricardo Bittencourt, the illuminated one)

Quote

> The one who thinks optimized code is ugly will never
> reach the Zen of programming

What is the Zen? Is it some type of..."charma"? A really..."high" point in programming?

Quote

> (as said by Ricardo Bittencourt, the illuminated one)

Illuminated?? You've gotta show me a JPEG of yourself so I can see the glow!

Happy holidays and happy new year,

Chris :o)

Author	Message
Zoop Joined: 24 Jun 1999 Posts: 1732 Location: Paris, France	Ugliest Code Contest ! Posted: Mon Dec 27, 1999 7:53 pm
	/* DISPLAY BACKGROUND in VIDEO MODE 3 -------------------------------------------- I am now trying to beat Marat on making the ugliest code possible I hope to have a good chance to win with this one Anyone dare to beat me ? / void Display_Background_3 (void) { int x, y=0, z; byte p_scr, pattern=BACK_AREA, tiles_data; for(;y<192;y+=32,pattern+=96) for(x=0;x<256;y-=32,x+=8) for(tiles_data=SG_BACK_TILE+(pattern++<<3),z=0;z<32;z ++) { p_scr=screenbuffer->line[y+(z&3)]+x; p_scr[0]=p_scr[1]=p_scr[2]=p_scr[3]=tiles_data>>4; p_scr[4]=p_scr[5]=p_scr[6]=p_scr[7]=*tiles_data&0x0F; if(!(z&3)){y+=4;tiles_data++;} } } Wow I'm quite proud of this one, it's unreadable - now I can release Meka sources safely ^_^ Zoop (gotta rework a bit on it now else I won't even be able to maintain the code myself)

Ricardo Bittencourt Guest	Beat this Posted: Mon Dec 27, 1999 9:36 pm
Ricardo Bittencourt Guest	; division by 9 written in Z80 assembly code ; originally written by Ricardo Bittencourt ; as a part of a routine to read a sector from msx disk-drives ; enter HL = number from 0 a 1440 ; exit A = HL/9 ; destroy HL,DE ; clock cycles presented by Z80 and R800 processors ; this is the fastest way to divide a number by 9 in these processors ; without using look-up tables ; ten points to anyone who understand how it works DIV9: ; Z80 R800 INC HL ; 7 1 LD D,H ; 5 1 LD E,L ; 5 1 ADD HL,HL ; 12 1 ADD HL,HL ; 12 1 ADD HL,HL ; 12 1 SBC HL,DE ; 17 2 LD E,0 ; 8 2 LD D,L ; 5 1 LD A,H ; 5 1 ADD HL,HL ; 12 1 ADD HL,HL ; 12 1 ADD HL,DE ; 12 1 ADC A,E ; 5 1 XOR H ; 5 1 AND 03FH ; 8 2 XOR H ; 5 1 RLCA ; 5 1 RLCA ; 5 1 RET ; total = 157 22

Johannes Holmberg Guest	You know I'd 0wn your ass at this, so I won't even post any code :) Posted: Mon Dec 27, 1999 11:55 pm
Johannes Holmberg Guest	...but if you really wanna see some ugly code, look at the sprite engine in uNESsential... :)

Charles MacDonald Joined: 28 Sep 1999 Posts: 1197	Posted: Tue Dec 28, 1999 4:42 am
Charles MacDonald Joined: 28 Sep 1999 Posts: 1197	How about this: c = ap ? a ? a : b : bp ? b ? b : a : a ? a : b; May not win for ugliness, but certainly could win a prize for being impossible to maintain or decipher. ;) It could be ten times more weird if I threw in an additional set of tests, but I haven't even figured that out yet. :)

Nyef Guest	Posted: Tue Dec 28, 1999 2:19 pm
Nyef Guest	[ code example removed ] Quote > Wow I'm quite proud of this one, it's unreadable - now I can release Meka sources safely ^_^ That's nothing. Go find the first releases of DarcNES, and look at nes_blit.c. Aside from the unholy mixture of x86 machine code and C structures, it contains the following little gem: /* * there are 256 blitters. hence 256 headers and cleanup routines. * for each pair of blitters x and y where (y = 255 - x), there are * eight 1 bits and eight 0 bits. taking this data, we come up with * the following calculation for how much memory we to allocate. * * this formula has since been modified for arbitrary length blitters and * the overhead of clipping the image. / #define SIZECALC(clipside) (((1 << width) (sizeof(blit_hdr) + sizeof(blit_cleanup))) + ((1 << (width - 1)) * width * (sizeof(blit_0) + sizeof(blit_1))) + ((flags & clipside)? (3 * (1 << width)): 0)); if ((flags & BFL_REVERSED) == 0) { i = SIZECALC(BFL_LEFTCLIP); } else { i = SIZECALC(BFL_RIGHTCLIP); } #undef SIZECALC The function then proceeds to allocate i bytes of memory and, using memcpy() and some direct manipualation, fills the memory with procedures. A procpointer array is then filled with pointers to the procedures. This is done several times. Of course, the entire approach was Just Plain Wrong, and the code has long since been scrapped. On the other paw, it's things like this that make me confident that I can write a dynarec CPU core whenever I feel like doing so. ^_^ And if you want more unreadable code, see the IOCCC. Quote > Zoop > (gotta rework a bit on it now else I won't even be able to maintain the code myself) --Nyef IOCCC

Chris Guest	Ahhh!! Posted: Tue Dec 28, 1999 6:18 pm
Chris Guest	I think you stretched a blood vessel in my brain. I feel so dizzy. Please, all of you should deserve the trophy for the uglyest programming codes. Chris :o)

Ricardo Bittencourt Guest	Re: Ahhh!! Posted: Wed Dec 29, 1999 1:36 am
Ricardo Bittencourt Guest	Quote > I think you stretched a blood vessel in my brain. I feel so dizzy. Please, all of you should deserve > the trophy for the uglyest programming codes. The one who thinks optimized code is ugly will never reach the Zen of programming (as said by Ricardo Bittencourt, the illuminated one)

Chris Guest	Enlighten me, Mr. Bittencourt. Posted: Wed Dec 29, 1999 3:52 am
Chris Guest	Quote > The one who thinks optimized code is ugly will never > reach the Zen of programming What is the Zen? Is it some type of..."charma"? A really..."high" point in programming? Quote > (as said by Ricardo Bittencourt, the illuminated one) Illuminated?? You've gotta show me a JPEG of yourself so I can see the glow! Happy holidays and happy new year, Chris :o)

Forums

View topic - Ugliest Code Contest !