Looking for the source code for /usr/games/maze

smbaker · Aug 22, 2022

The binary was found in a Unix Sys V distribution circa 1986, and I've seen references in a few manpages and even a PDF manual. Extracting strings from the binary I can see that it was compiled from "maze.c". But I can't find this particular maze.c anywhere. I have found other game source from that era, such as hangman, wumpus, adventure, etc., at https://github.com/msharov/bsd-games. But no maze.

Anyone have any possible leads for me to follow up?

Thanks

Chuck(G) · Aug 22, 2022

I've got this one from an early 80's BSD tape:

Code:

main(argc,argv)
int argc; 
char *argv[];
{
	int a,c,e,f,g,i,j,k,q,r,s,t,u,v;
	int d[4],h[4],m[4],n[10000],o[4],o2[4],p[4],z[5000],w[4];
	char x,y;
	if (argc == 1) {
		printf("syntax error\n"); 
		exit();
	}
	if (argc != 3) {
		printf("length error\n"); 
		exit();
	}
	r=fii(argv[1])*vi(argv[1]); 
	c=fii(argv[2])*vi(argv[2]);
	if (r*c == 0) {
		printf("domain error\n"); 
		exit();
	}
	i=r*c; 
	e=j=v=0; 
	for(s=0; s<i; ++s) z[s]=0;  
	for(s=0; s<i+i; ++s) n[s]=0;
	o2[3]=z[0]=u=2; 
	z[n[1]=i-1]=24; 
	for(s=0; s<4; ++s) o[s]=1<<s;
	o2[0]=4; 
	o2[1]=8; 
	o2[2]=1; 
	m[0]=1; 
	m[1]=(-c); 
	m[2]=(-1); 
	m[3]=c; 
	goto lb;
la:
	s=1+roll(k); 
	t=0; 
	for(a=0; a<4; ++a) {
		t+=p[a]&&h[a]; 
		if(s==t) break;
	}
	q=z[e=d[a]]; 
	j=(q>j)?q:j; 
	z[g]=o[a]+f;
	z[e]=o2[a]+q+(j||f<16 ?0:16); 
	if (k!=1) n[u++]=e;
lb:
	for(s=k=0; s<4; ++s) {
		t=d[s]=e+m[s]; 
		if (t<0 || t>=i) h[s]=0; 
		else {
			switch (s) {
			case 0:
			case 2: 
				h[s]=(t/c == e/c)?1:0; 
				break;
			case 1:
			case 3: 
				h[s]=1;  
			}
			f=z[e]; 
			t=z[t]; 
			k+=p[s]=h[s]&&(t==0||j<(f<16!=t<16));
		} 
	}
	g=e; 
	if (k!=0) goto la; 
	if (u==v) goto lc; 
	e=n[v++]; 
	goto lb;
lc:
	printf("  "); 
	for(s=2; s<=c; ++s) printf(" _"); 
	printf("\n");
	for(s=0; s<r; ++s) { 
		for (t=0; t<c; ++t) { 
			u=z[t+s*c];
			v=u/4; 
			x=(v==(2*(v/2)))?'|':' '; 
			v=u/8; 
			y=(v==(2*(v/2)))?'_':' ';
			printf("%c%c",x,y); 
		} 
		printf("|\n"); 
	}
}

Gotta love the extensive commentary...

smbaker · Aug 22, 2022

Chuck(G) said:

I've got this one from an early 80's BSD tape:

Code:

main(argc,argv)
int argc;
char *argv[];
{
    int a,c,e,f,g,i,j,k,q,r,s,t,u,v;
    int d[4],h[4],m[4],n[10000],o[4],o2[4],p[4],z[5000],w[4];
    char x,y;
    if (argc == 1) {
        printf("syntax error\n");
        exit();
    }
    if (argc != 3) {
        printf("length error\n");
        exit();
    }
    r=fii(argv[1])*vi(argv[1]);
    c=fii(argv[2])*vi(argv[2]);
    if (r*c == 0) {
        printf("domain error\n");
        exit();
    }
    i=r*c;
    e=j=v=0;
    for(s=0; s<i; ++s) z[s]=0;
    for(s=0; s<i+i; ++s) n[s]=0;
    o2[3]=z[0]=u=2;
    z[n[1]=i-1]=24;
    for(s=0; s<4; ++s) o[s]=1<<s;
    o2[0]=4;
    o2[1]=8;
    o2[2]=1;
    m[0]=1;
    m[1]=(-c);
    m[2]=(-1);
    m[3]=c;
    goto lb;
la:
    s=1+roll(k);
    t=0;
    for(a=0; a<4; ++a) {
        t+=p[a]&&h[a];
        if(s==t) break;
    }
    q=z[e=d[a]];
    j=(q>j)?q:j;
    z[g]=o[a]+f;
    z[e]=o2[a]+q+(j||f<16 ?0:16);
    if (k!=1) n[u++]=e;
lb:
    for(s=k=0; s<4; ++s) {
        t=d[s]=e+m[s];
        if (t<0 || t>=i) h[s]=0;
        else {
            switch (s) {
            case 0:
            case 2:
                h[s]=(t/c == e/c)?1:0;
                break;
            case 1:
            case 3:
                h[s]=1;
            }
            f=z[e];
            t=z[t];
            k+=p[s]=h[s]&&(t==0||j<(f<16!=t<16));
        }
    }
    g=e;
    if (k!=0) goto la;
    if (u==v) goto lc;
    e=n[v++];
    goto lb;
lc:
    printf("  ");
    for(s=2; s<=c; ++s) printf(" _");
    printf("\n");
    for(s=0; s<r; ++s) {
        for (t=0; t<c; ++t) {
            u=z[t+s*c];
            v=u/4;
            x=(v==(2*(v/2)))?'|':' ';
            v=u/8;
            y=(v==(2*(v/2)))?'_':' ';
            printf("%c%c",x,y);
        }
        printf("|\n");
    }
}

Gotta love the extensive commentary...

Thanks Chuck, that looks very plausible. Perhaps a slightly older version as I know the binary I have prints "maze seed is <x>" at the start of it. The binary I have also assumes a default number of rows and columns. But this will do! Thanks again!

Was there anything handy (header file, another source file, etc) that expanded the functions vi() fii() and roll() ? I can assume roll() is a random number generator, and fii() and vi() have something to do with converting arguments from strings to integers.

The following seemed to suffice to get it to compile under linux:

Code:

int fii(char *s)
{
  return 1;
}

int vi(char *s)
{
  return atoi(s);
}

int roll(int m)
{
  return rand() % m;
}

and produce a maze:

Code:

$ ./maze 10 10
   _ _ _ _ _ _ _ _ _
| |_   _ _  |  _|   |
|  _ _  |_ _|_ _ _| |
|_  |_ _|   |  _ _ _|
|_ _    | | | |_    |
|   |_| | | |_ _ _| |
| |_|    _| |  _  | |
|_ _ _| | |_  | |  _|
|  _ _ _|_ _ _|_ _  |
|_ _ _  |   |   |  _|
|_ _ _ _|_|_ _|_ _  |

Chuck(G) · Aug 22, 2022

No--and this was the "beautified" version of the source. If you want to see the "ugly" version, here it is:

Code:

main(argc,argv)
int argc; char *argv[];
{
      int a,c,e,f,g,i,j,k,q,r,s,t,u,v;
      int d[4],h[4],m[4],n[10000],o[4],o2[4],p[4],z[5000],w[4];
      char x,y;
	  if (argc == 1) {printf("syntax error\n"); exit();}
	  if (argc != 3) {printf("length error\n"); exit();}
	  r=fii(argv[1])*vi(argv[1]); c=fii(argv[2])*vi(argv[2]);
	  if (r*c == 0) {printf("domain error\n"); exit();}
	  i=r*c; e=j=v=0; for(s=0; s<i; ++s) z[s]=0;  for(s=0; s<i+i; ++s) n[s]=0;
      o2[3]=z[0]=u=2; z[n[1]=i-1]=24; for(s=0; s<4; ++s) o[s]=1<<s;
      o2[0]=4; o2[1]=8; o2[2]=1; m[0]=1; m[1]=(-c); m[2]=(-1); m[3]=c; goto lb;
    la:s=1+roll(k); t=0; for(a=0; a<4; ++a) {t+=p[a]&&h[a]; if(s==t) break;}
      q=z[e=d[a]]; j=(q>j)?q:j; z[g]=o[a]+f;
      z[e]=o2[a]+q+(j||f<16 ?0:16); if (k!=1) n[u++]=e;
    lb:for(s=k=0; s<4; ++s) {t=d[s]=e+m[s]; if (t<0 || t>=i) h[s]=0; 
               else {switch (s) {case 0:
      			         case 2: h[s]=(t/c == e/c)?1:0; break;
      			         case 1:
      			         case 3: h[s]=1;  }
      f=z[e]; t=z[t]; k+=p[s]=h[s]&&(t==0||j<(f<16!=t<16));} }
      g=e; if (k!=0) goto la; if (u==v) goto lc; e=n[v++]; goto lb;
    lc:printf("  "); for(s=2; s<=c; ++s) printf(" _"); printf("\n");
      for(s=0; s<r; ++s) { for (t=0; t<c; ++t) { u=z[t+s*c];
     v=u/4; x=(v==(2*(v/2)))?'|':' '; v=u/8; y=(v==(2*(v/2)))?'_':' ';
      printf("%c%c",x,y); } printf("|\n"); }
}

Comes from a BSD tape at UCB from 1981. Likely created on Ernie.

Chuck(G) · Aug 22, 2022

The functions you cited are implemented originally as follows:

Code:

int roll(n) /* returns a "random" integer from 0 to n-1 */
int n;
{
        int a;
        a=rand(); a=a/(1+32767/n);
        return (a);
}

Code:

/*  for string <a>, returns the integer value represented.  Leading blank */
/*  space is ignored.  The result is undefined if <a> contains any funny stuff*/
long fii(a)
char a[];
{
        int i,n,s;

        for(i=0; a[i]==' ' || a[i]=='\n' || a[i]=='\t'; i++);
        s=1;
        if (a[i]=='+' || a[i]=='-') {s=(a[i]=='+')?1:-1; ++i;}
        for(n=0; a[i]>='0' && a[i]<='9'; i++) n=10*n+a[i]-'0';
        n=n*s;
        if (a[i] != '.') return (n);
        ++i;
        if (a[i]>='0' && a[i]<'5') return (n);
        if (a[i]>'4' && a[i]<='9') return (n+s);
        return;
}

Code:

/* for string <a> , returns integer boolean indicating whether or not <a>   */
/* represents a valid number.  Leading and trailing white space is ignored. */
vi(a)
char a[];
{
        int i;
        for(i=0; a[i]==' ' || a[i]=='\n' || a[i]=='\t'; i++) ;
        i+=a[i]=='+' || a[i]=='-';
        for(;a[i]>='0' && a[i]<='9';i++) ;
        if (a[i]=='.') {++i; for(;a[i]>='0' && a[i]<='9';i++) ;}
        for(;a[i]==' ' || a[i]=='\n' || a[i]=='\t'; i++) ;
        return (a[i]=='\0');
}

FWIW, I also have the source for a few other games, such as rogue on the same tape.

smbaker · Aug 22, 2022

Excellent. Now I can keep it original.

I wouldn't mind any source code (for games or utilities) that you can dig up. This old stuff is a match for the ancient compiler I have been using.

Chuck(G) · Aug 22, 2022

Here's rogue, complete with makefile:

acgs · Aug 22, 2022

smbaker said:
Thanks Chuck, that looks very plausible. Perhaps a slightly older version as I know the binary I have prints "maze seed is <x>" at the start of it. The binary I have also assumes a default number of rows and columns. But this will do! Thanks again!

This code prints the maze seed like the binary you're describing:
https://www.tuhs.org/cgi-bin/utree.pl?file=pdp11v/usr/src/games/maze.c

But the maze size is set with a #define.

It's from 1983 and therefore older than the one you're trying to match.

Chuck(G) · Aug 22, 2022

I may have other versions, but I'll have to go digging for them.

Caluser2000 · Aug 22, 2022

Is there much difference in the sizes of the source files compared to later BSD Games in newer BSDs or Linux?

smbaker · Aug 23, 2022

acgs said:
This code prints the maze seed like the binary you're describing:
https://www.tuhs.org/cgi-bin/utree.pl?file=pdp11v/usr/src/games/maze.c

But the maze size is set with a #define.

It's from 1983 and therefore older than the one you're trying to match.

We have a winner!

My binary also has the following strings, which match your source:

Code:

path length = %d
back-outs = %d

Looking for the source code for /usr/games/maze

smbaker

Experienced Member

Chuck(G)

25k Member

smbaker

Experienced Member

Chuck(G)

25k Member

Chuck(G)

25k Member

smbaker

Experienced Member

Chuck(G)

25k Member

Attachments

acgs

Experienced Member

Chuck(G)

25k Member

Caluser2000

Banned

smbaker

Experienced Member