+#include <stdio.h>
+#include <ctype.h>
+
+#ifndef unix
+#define SHIFT 5
+#define TABSIZE (int)(400000/(1<<SHIFT))
+int *tab; /*honeywell loader deficiency*/
+#else
+#define Tolower(c) (isupper(c)?tolower(c):c) /* ugh!!! */
+#define SHIFT 4
+#define TABSIZE 25000 /*(int)(400000/(1<<shift))--pdp11 compiler deficiency*/
+short tab[TABSIZE];
+#endif
+long p[] = {
+ 399871,
+ 399887,
+ 399899,
+ 399911,
+ 399913,
+ 399937,
+ 399941,
+ 399953,
+ 399979,
+ 399983,
+ 399989,
+};
+#define NP (sizeof(p)/sizeof(p[0]))
+#define NW 30
+
+/*
+* Hash table for spelling checker has n bits.
+* Each word w is hashed by k different (modular) hash functions, hi.
+* The bits hi(w), i=1..k, are set for words in the dictionary.
+* Assuming independence, the probability that no word of a d-word
+* dictionary sets a particular bit is given by the Poisson formula
+* P = exp(-y)*y**0/0!, where y=d*k/n.
+* The probability that a random string is recognized as a word is then
+* (1-P)**k. For given n and d this is minimum when y=log(2), P=1/2,
+* whence one finds, for example, that a 25000-word dictionary in a
+* 400000-bit table works best with k=11.
+*/
+
+long pow2[NP][NW];
+
+prime(argc, argv) register char **argv;
+{
+ int i, j;
+ long h;
+ register long *lp;
+
+#ifndef unix
+ if ((tab = (int *)calloc(sizeof(*tab), TABSIZE)) == NULL)
+ return(0);
+#endif
+ if (argc > 1) {
+ FILE *f;
+ if ((f = fopen(argv[1], "ri")) == NULL)
+ return(0);
+ if (fread((char *)tab, sizeof(*tab), TABSIZE, f) != TABSIZE)
+ return(0);
+ fclose(f);
+ }
+ for (i=0; i<NP; i++) {
+ h = *(lp = pow2[i]) = 1<<14;
+ for (j=1; j<NW; j++)
+ h = *++lp = (h<<7) % p[i];
+ }
+ return(1);
+}
+
+#define get(h) (tab[h>>SHIFT]&(1<<((int)h&((1<<SHIFT)-1))))
+#define set(h) tab[h>>SHIFT] |= 1<<((int)h&((1<<SHIFT)-1))
projects / unix-history / commitdiff