Annotation of mandoc/chars.c, Revision 1.21
1.21 ! kristaps 1: /* $Id: chars.c,v 1.20 2010/06/19 20:46:27 kristaps Exp $ */
1.1 kristaps 2: /*
1.20 kristaps 3: * Copyright (c) 2009 Kristaps Dzonsons <kristaps@bsd.lv>
1.1 kristaps 4: *
5: * Permission to use, copy, modify, and distribute this software for any
6: * purpose with or without fee is hereby granted, provided that the above
7: * copyright notice and this permission notice appear in all copies.
8: *
9: * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
10: * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
11: * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
12: * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
13: * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
14: * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
15: * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
16: */
1.14 kristaps 17: #ifdef HAVE_CONFIG_H
18: #include "config.h"
19: #endif
20:
1.1 kristaps 21: #include <assert.h>
1.10 kristaps 22: #include <stdio.h>
1.1 kristaps 23: #include <stdlib.h>
24: #include <string.h>
25:
1.18 kristaps 26: #include "mandoc.h"
1.1 kristaps 27: #include "chars.h"
28:
1.2 kristaps 29: #define PRINT_HI 126
30: #define PRINT_LO 32
1.1 kristaps 31:
32: struct ln {
33: struct ln *next;
34: const char *code;
1.21 ! kristaps 35: size_t codesz;
1.2 kristaps 36: const char *ascii;
37: size_t asciisz;
1.21 ! kristaps 38: int unicode;
1.1 kristaps 39: int type;
40: #define CHARS_CHAR (1 << 0)
41: #define CHARS_STRING (1 << 1)
1.12 kristaps 42: #define CHARS_BOTH (CHARS_CHAR | CHARS_STRING)
1.1 kristaps 43: };
44:
1.19 kristaps 45: #define LINES_MAX 370
1.1 kristaps 46:
1.21 ! kristaps 47: #define CHAR(in, insz, ch, chsz, code) \
! 48: { NULL, (in), (insz), (ch), (chsz), (code), CHARS_CHAR },
! 49: #define STRING(in, insz, ch, chsz, code) \
! 50: { NULL, (in), (insz), (ch), (chsz), (code), CHARS_STRING },
! 51: #define BOTH(in, insz, ch, chsz, code) \
! 52: { NULL, (in), (insz), (ch), (chsz), (code), CHARS_BOTH },
1.1 kristaps 53:
1.13 kristaps 54: #define CHAR_TBL_START static struct ln lines[LINES_MAX] = {
55: #define CHAR_TBL_END };
56:
1.1 kristaps 57: #include "chars.in"
58:
59: struct tbl {
1.2 kristaps 60: enum chars type;
1.1 kristaps 61: struct ln **htab;
62: };
63:
64: static inline int match(const struct ln *,
65: const char *, size_t, int);
1.21 ! kristaps 66: static const struct ln *find(struct tbl *, const char *, size_t, int);
1.1 kristaps 67:
68:
69: void
70: chars_free(void *arg)
71: {
72: struct tbl *tab;
73:
74: tab = (struct tbl *)arg;
75:
76: free(tab->htab);
77: free(tab);
78: }
79:
80:
81: void *
82: chars_init(enum chars type)
83: {
84: struct tbl *tab;
85: struct ln **htab;
86: struct ln *pp;
87: int i, hash;
88:
89: /*
90: * Constructs a very basic chaining hashtable. The hash routine
91: * is simply the integral value of the first character.
92: * Subsequent entries are chained in the order they're processed
93: * (they're in-line re-ordered during lookup).
94: */
95:
1.10 kristaps 96: tab = malloc(sizeof(struct tbl));
97: if (NULL == tab) {
1.11 kristaps 98: perror(NULL);
1.10 kristaps 99: exit(EXIT_FAILURE);
100: }
1.1 kristaps 101:
1.2 kristaps 102: htab = calloc(PRINT_HI - PRINT_LO + 1, sizeof(struct ln **));
1.10 kristaps 103: if (NULL == htab) {
1.11 kristaps 104: perror(NULL);
1.10 kristaps 105: exit(EXIT_FAILURE);
106: }
1.1 kristaps 107:
108: for (i = 0; i < LINES_MAX; i++) {
1.2 kristaps 109: hash = (int)lines[i].code[0] - PRINT_LO;
1.1 kristaps 110:
111: if (NULL == (pp = htab[hash])) {
112: htab[hash] = &lines[i];
113: continue;
114: }
115:
116: for ( ; pp->next; pp = pp->next)
117: /* Scan ahead. */ ;
118: pp->next = &lines[i];
119: }
120:
121: tab->htab = htab;
1.10 kristaps 122: tab->type = type;
1.1 kristaps 123: return(tab);
124: }
125:
126:
1.21 ! kristaps 127: /*
! 128: * Special character to Unicode codepoint.
! 129: */
! 130: int
! 131: chars_spec2cp(void *arg, const char *p, size_t sz)
! 132: {
! 133: const struct ln *ln;
! 134:
! 135: ln = find((struct tbl *)arg, p, sz, CHARS_CHAR);
! 136: if (NULL == ln)
! 137: return(-1);
! 138: return(ln->unicode);
! 139: }
! 140:
! 141:
! 142: /*
! 143: * Reserved word to Unicode codepoint.
! 144: */
! 145: int
! 146: chars_res2cp(void *arg, const char *p, size_t sz)
! 147: {
! 148: const struct ln *ln;
! 149:
! 150: ln = find((struct tbl *)arg, p, sz, CHARS_STRING);
! 151: if (NULL == ln)
! 152: return(-1);
! 153: return(ln->unicode);
! 154: }
! 155:
! 156:
! 157: /*
! 158: * Special character to string array.
! 159: */
1.1 kristaps 160: const char *
1.21 ! kristaps 161: chars_spec2str(void *arg, const char *p, size_t sz, size_t *rsz)
1.1 kristaps 162: {
1.21 ! kristaps 163: const struct ln *ln;
! 164:
! 165: ln = find((struct tbl *)arg, p, sz, CHARS_CHAR);
! 166: if (NULL == ln)
! 167: return(NULL);
1.1 kristaps 168:
1.21 ! kristaps 169: *rsz = ln->asciisz;
! 170: return(ln->ascii);
1.1 kristaps 171: }
172:
173:
1.21 ! kristaps 174: /*
! 175: * Reserved word to string array.
! 176: */
1.1 kristaps 177: const char *
1.21 ! kristaps 178: chars_res2str(void *arg, const char *p, size_t sz, size_t *rsz)
1.1 kristaps 179: {
1.21 ! kristaps 180: const struct ln *ln;
1.1 kristaps 181:
1.21 ! kristaps 182: ln = find((struct tbl *)arg, p, sz, CHARS_STRING);
! 183: if (NULL == ln)
! 184: return(NULL);
! 185:
! 186: *rsz = ln->asciisz;
! 187: return(ln->ascii);
1.1 kristaps 188: }
189:
190:
1.21 ! kristaps 191: static const struct ln *
! 192: find(struct tbl *tab, const char *p, size_t sz, int type)
1.1 kristaps 193: {
194: struct ln *pp, *prev;
195: struct ln **htab;
196: int hash;
197:
198: assert(p);
199: assert(sz > 0);
200:
1.2 kristaps 201: if (p[0] < PRINT_LO || p[0] > PRINT_HI)
1.1 kristaps 202: return(NULL);
203:
204: /*
205: * Lookup the symbol in the symbol hash. See ascii2htab for the
206: * hashtable specs. This dynamically re-orders the hash chain
207: * to optimise for repeat hits.
208: */
209:
1.2 kristaps 210: hash = (int)p[0] - PRINT_LO;
1.1 kristaps 211: htab = tab->htab;
212:
213: if (NULL == (pp = htab[hash]))
214: return(NULL);
215:
216: for (prev = NULL; pp; pp = pp->next) {
217: if ( ! match(pp, p, sz, type)) {
218: prev = pp;
219: continue;
220: }
221:
222: if (prev) {
223: prev->next = pp->next;
224: pp->next = htab[hash];
225: htab[hash] = pp;
226: }
227:
1.21 ! kristaps 228: return(pp);
1.1 kristaps 229: }
230:
231: return(NULL);
232: }
233:
234:
235: static inline int
236: match(const struct ln *ln, const char *p, size_t sz, int type)
237: {
238:
239: if ( ! (ln->type & type))
240: return(0);
241: if (ln->codesz != sz)
242: return(0);
243: return(0 == strncmp(ln->code, p, sz));
244: }
CVSweb