1 #include <assert.h>
2 #include <stdio.h>
3 #include <string.h>
4
5 #include "buffer.h"
6 #include "houdini.h"
7 #include "utf8.h"
8 #include "entities.inc"
9
10 /* Binary tree lookup code for entities added by JGM */
11
S_lookup(int i,int low,int hi,const unsigned char * s,int len)12 static const unsigned char *S_lookup(int i, int low, int hi,
13 const unsigned char *s, int len) {
14 int j;
15 int cmp =
16 strncmp((const char *)s, (const char *)cmark_entities[i].entity, len);
17 if (cmp == 0 && cmark_entities[i].entity[len] == 0) {
18 return (const unsigned char *)cmark_entities[i].bytes;
19 } else if (cmp <= 0 && i > low) {
20 j = i - ((i - low) / 2);
21 if (j == i)
22 j -= 1;
23 return S_lookup(j, low, i - 1, s, len);
24 } else if (cmp > 0 && i < hi) {
25 j = i + ((hi - i) / 2);
26 if (j == i)
27 j += 1;
28 return S_lookup(j, i + 1, hi, s, len);
29 } else {
30 return NULL;
31 }
32 }
33
S_lookup_entity(const unsigned char * s,int len)34 static const unsigned char *S_lookup_entity(const unsigned char *s, int len) {
35 return S_lookup(CMARK_NUM_ENTITIES / 2, 0, CMARK_NUM_ENTITIES - 1, s, len);
36 }
37
houdini_unescape_ent(cmark_strbuf * ob,const uint8_t * src,bufsize_t size)38 bufsize_t houdini_unescape_ent(cmark_strbuf *ob, const uint8_t *src,
39 bufsize_t size) {
40 bufsize_t i = 0;
41
42 if (size >= 3 && src[0] == '#') {
43 int codepoint = 0;
44 int num_digits = 0;
45
46 if (_isdigit(src[1])) {
47 for (i = 1; i < size && _isdigit(src[i]); ++i) {
48 codepoint = (codepoint * 10) + (src[i] - '0');
49
50 if (codepoint >= 0x110000) {
51 // Keep counting digits but
52 // avoid integer overflow.
53 codepoint = 0x110000;
54 }
55 }
56
57 num_digits = i - 1;
58 }
59
60 else if (src[1] == 'x' || src[1] == 'X') {
61 for (i = 2; i < size && _isxdigit(src[i]); ++i) {
62 codepoint = (codepoint * 16) + ((src[i] | 32) % 39 - 9);
63
64 if (codepoint >= 0x110000) {
65 // Keep counting digits but
66 // avoid integer overflow.
67 codepoint = 0x110000;
68 }
69 }
70
71 num_digits = i - 2;
72 }
73
74 if (num_digits >= 1 && num_digits <= 8 && i < size && src[i] == ';') {
75 if (codepoint == 0 || (codepoint >= 0xD800 && codepoint < 0xE000) ||
76 codepoint >= 0x110000) {
77 codepoint = 0xFFFD;
78 }
79 cmark_utf8proc_encode_char(codepoint, ob);
80 return i + 1;
81 }
82 }
83
84 else {
85 if (size > CMARK_ENTITY_MAX_LENGTH)
86 size = CMARK_ENTITY_MAX_LENGTH;
87
88 for (i = CMARK_ENTITY_MIN_LENGTH; i < size; ++i) {
89 if (src[i] == ' ')
90 break;
91
92 if (src[i] == ';') {
93 const unsigned char *entity = S_lookup_entity(src, i);
94
95 if (entity != NULL) {
96 cmark_strbuf_puts(ob, (const char *)entity);
97 return i + 1;
98 }
99
100 break;
101 }
102 }
103 }
104
105 return 0;
106 }
107
houdini_unescape_html(cmark_strbuf * ob,const uint8_t * src,bufsize_t size)108 int houdini_unescape_html(cmark_strbuf *ob, const uint8_t *src,
109 bufsize_t size) {
110 bufsize_t i = 0, org, ent;
111
112 while (i < size) {
113 org = i;
114 while (i < size && src[i] != '&')
115 i++;
116
117 if (likely(i > org)) {
118 if (unlikely(org == 0)) {
119 if (i >= size)
120 return 0;
121
122 cmark_strbuf_grow(ob, HOUDINI_UNESCAPED_SIZE(size));
123 }
124
125 cmark_strbuf_put(ob, src + org, i - org);
126 }
127
128 /* escaping */
129 if (i >= size)
130 break;
131
132 i++;
133
134 ent = houdini_unescape_ent(ob, src + i, size - i);
135 i += ent;
136
137 /* not really an entity */
138 if (ent == 0)
139 cmark_strbuf_putc(ob, '&');
140 }
141
142 return 1;
143 }
144
houdini_unescape_html_f(cmark_strbuf * ob,const uint8_t * src,bufsize_t size)145 void houdini_unescape_html_f(cmark_strbuf *ob, const uint8_t *src,
146 bufsize_t size) {
147 if (!houdini_unescape_html(ob, src, size))
148 cmark_strbuf_put(ob, src, size);
149 }
150