Annotation of mandoc/chars.c, Revision 1.22
1.22 ! kristaps 1: /* $Id: chars.c,v 1.21 2010/07/16 22:33:30 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;
1.2 kristaps 35: const char *ascii;
36: size_t asciisz;
1.21 kristaps 37: int unicode;
1.1 kristaps 38: int type;
39: #define CHARS_CHAR (1 << 0)
40: #define CHARS_STRING (1 << 1)
1.12 kristaps 41: #define CHARS_BOTH (CHARS_CHAR | CHARS_STRING)
1.1 kristaps 42: };
43:
1.19 kristaps 44: #define LINES_MAX 370
1.1 kristaps 45:
1.22 ! kristaps 46: #define CHAR(in, ch, chsz, code) \
! 47: { NULL, (in), (ch), (chsz), (code), CHARS_CHAR },
! 48: #define STRING(in, ch, chsz, code) \
! 49: { NULL, (in), (ch), (chsz), (code), CHARS_STRING },
! 50: #define BOTH(in, ch, chsz, code) \
! 51: { NULL, (in), (ch), (chsz), (code), CHARS_BOTH },
1.1 kristaps 52:
1.13 kristaps 53: #define CHAR_TBL_START static struct ln lines[LINES_MAX] = {
54: #define CHAR_TBL_END };
55:
1.1 kristaps 56: #include "chars.in"
57:
58: struct tbl {
1.2 kristaps 59: enum chars type;
1.1 kristaps 60: struct ln **htab;
61: };
62:
63: static inline int match(const struct ln *,
64: const char *, size_t, int);
1.21 kristaps 65: static const struct ln *find(struct tbl *, const char *, size_t, int);
1.1 kristaps 66:
67:
68: void
69: chars_free(void *arg)
70: {
71: struct tbl *tab;
72:
73: tab = (struct tbl *)arg;
74:
75: free(tab->htab);
76: free(tab);
77: }
78:
79:
80: void *
81: chars_init(enum chars type)
82: {
83: struct tbl *tab;
84: struct ln **htab;
85: struct ln *pp;
86: int i, hash;
87:
88: /*
89: * Constructs a very basic chaining hashtable. The hash routine
90: * is simply the integral value of the first character.
91: * Subsequent entries are chained in the order they're processed
92: * (they're in-line re-ordered during lookup).
93: */
94:
1.10 kristaps 95: tab = malloc(sizeof(struct tbl));
96: if (NULL == tab) {
1.11 kristaps 97: perror(NULL);
1.10 kristaps 98: exit(EXIT_FAILURE);
99: }
1.1 kristaps 100:
1.2 kristaps 101: htab = calloc(PRINT_HI - PRINT_LO + 1, sizeof(struct ln **));
1.10 kristaps 102: if (NULL == htab) {
1.11 kristaps 103: perror(NULL);
1.10 kristaps 104: exit(EXIT_FAILURE);
105: }
1.1 kristaps 106:
107: for (i = 0; i < LINES_MAX; i++) {
1.2 kristaps 108: hash = (int)lines[i].code[0] - PRINT_LO;
1.1 kristaps 109:
110: if (NULL == (pp = htab[hash])) {
111: htab[hash] = &lines[i];
112: continue;
113: }
114:
115: for ( ; pp->next; pp = pp->next)
116: /* Scan ahead. */ ;
117: pp->next = &lines[i];
118: }
119:
120: tab->htab = htab;
1.10 kristaps 121: tab->type = type;
1.1 kristaps 122: return(tab);
123: }
124:
125:
1.21 kristaps 126: /*
127: * Special character to Unicode codepoint.
128: */
129: int
130: chars_spec2cp(void *arg, const char *p, size_t sz)
131: {
132: const struct ln *ln;
133:
134: ln = find((struct tbl *)arg, p, sz, CHARS_CHAR);
135: if (NULL == ln)
136: return(-1);
137: return(ln->unicode);
138: }
139:
140:
141: /*
142: * Reserved word to Unicode codepoint.
143: */
144: int
145: chars_res2cp(void *arg, const char *p, size_t sz)
146: {
147: const struct ln *ln;
148:
149: ln = find((struct tbl *)arg, p, sz, CHARS_STRING);
150: if (NULL == ln)
151: return(-1);
152: return(ln->unicode);
153: }
154:
155:
156: /*
157: * Special character to string array.
158: */
1.1 kristaps 159: const char *
1.21 kristaps 160: chars_spec2str(void *arg, const char *p, size_t sz, size_t *rsz)
1.1 kristaps 161: {
1.21 kristaps 162: const struct ln *ln;
163:
164: ln = find((struct tbl *)arg, p, sz, CHARS_CHAR);
165: if (NULL == ln)
166: return(NULL);
1.1 kristaps 167:
1.21 kristaps 168: *rsz = ln->asciisz;
169: return(ln->ascii);
1.1 kristaps 170: }
171:
172:
1.21 kristaps 173: /*
174: * Reserved word to string array.
175: */
1.1 kristaps 176: const char *
1.21 kristaps 177: chars_res2str(void *arg, const char *p, size_t sz, size_t *rsz)
1.1 kristaps 178: {
1.21 kristaps 179: const struct ln *ln;
1.1 kristaps 180:
1.21 kristaps 181: ln = find((struct tbl *)arg, p, sz, CHARS_STRING);
182: if (NULL == ln)
183: return(NULL);
184:
185: *rsz = ln->asciisz;
186: return(ln->ascii);
1.1 kristaps 187: }
188:
189:
1.21 kristaps 190: static const struct ln *
191: find(struct tbl *tab, const char *p, size_t sz, int type)
1.1 kristaps 192: {
193: struct ln *pp, *prev;
194: struct ln **htab;
195: int hash;
196:
197: assert(p);
198: assert(sz > 0);
199:
1.2 kristaps 200: if (p[0] < PRINT_LO || p[0] > PRINT_HI)
1.1 kristaps 201: return(NULL);
202:
203: /*
204: * Lookup the symbol in the symbol hash. See ascii2htab for the
205: * hashtable specs. This dynamically re-orders the hash chain
206: * to optimise for repeat hits.
207: */
208:
1.2 kristaps 209: hash = (int)p[0] - PRINT_LO;
1.1 kristaps 210: htab = tab->htab;
211:
212: if (NULL == (pp = htab[hash]))
213: return(NULL);
214:
215: for (prev = NULL; pp; pp = pp->next) {
216: if ( ! match(pp, p, sz, type)) {
217: prev = pp;
218: continue;
219: }
220:
221: if (prev) {
222: prev->next = pp->next;
223: pp->next = htab[hash];
224: htab[hash] = pp;
225: }
226:
1.21 kristaps 227: return(pp);
1.1 kristaps 228: }
229:
230: return(NULL);
231: }
232:
233:
234: static inline int
235: match(const struct ln *ln, const char *p, size_t sz, int type)
236: {
237:
238: if ( ! (ln->type & type))
239: return(0);
1.22 ! kristaps 240: if (strncmp(ln->code, p, sz))
1.1 kristaps 241: return(0);
1.22 ! kristaps 242: return('\0' == ln->code[(int)sz]);
1.1 kristaps 243: }
CVSweb