1 /*
2 * Copyright (c) 2012 Tim Ruehsen
3 * Copyright (c) 2015-2021 Free Software Foundation, Inc.
4 *
5 * This file is part of libwget.
6 *
7 * Libwget is free software: you can redistribute it and/or modify
8 * it under the terms of the GNU Lesser General Public License as published by
9 * the Free Software Foundation, either version 3 of the License, or
10 * (at your option) any later version.
11 *
12 * Libwget is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 * GNU Lesser General Public License for more details.
16 *
17 * You should have received a copy of the GNU Lesser General Public License
18 * along with libwget. If not, see <https://www.gnu.org/licenses/>.
19 *
20 *
21 * css parsing routines
22 *
23 * Changelog
24 * 03.07.2012 Tim Ruehsen created
25 *
26 * A parser using the flex tokenizer, created with flex tokens from
27 * https://www.w3.org/TR/css3-syntax/
28 *
29 * TODO:
30 * - since we are just interested in @import ... and url(...), we could use
31 * a simplistic hand-written parser which might be much smaller and faster
32 */
33
34 #include <config.h>
35
36 #include <stddef.h>
37 #include <unistd.h>
38 #include <string.h>
39 #include <c-ctype.h>
40 #include <fcntl.h>
41 #include <sys/stat.h>
42 #ifdef HAVE_MMAP
43 #include <sys/mman.h>
44 #endif
45
46 #include <wget.h>
47 #include "private.h"
48
49 #include "css_tokenizer.h"
50
51 // see css_tokenizer.c
52 typedef void* yyscan_t;
53 int yyget_leng(yyscan_t yyscanner);
54 char *yyget_text(yyscan_t yyscanner);
55 typedef struct yy_buffer_state *YY_BUFFER_STATE;
56 int yylex_init(yyscan_t* scanner);
57 YY_BUFFER_STATE yy_scan_string(const char * yystr, yyscan_t yyscanner);
58 YY_BUFFER_STATE yy_scan_bytes(const char * yystr, int len, yyscan_t yyscanner);
59 int yylex(yyscan_t yyscanner);
60 int yylex_destroy(yyscan_t yyscanner);
61 void *yyalloc(size_t size);
62 void *yyrealloc(void *p, size_t size);
63
yyalloc(size_t size)64 void *yyalloc(size_t size) {
65 return wget_malloc(size);
66 }
yyrealloc(void * p,size_t size)67 void *yyrealloc(void *p, size_t size) {
68 return wget_realloc(p, size);
69 }
70
wget_css_parse_buffer(const char * buf,size_t len,wget_css_parse_uri_callback * callback_uri,wget_css_parse_encoding_callback * callback_encoding,void * user_ctx)71 void wget_css_parse_buffer(
72 const char *buf,
73 size_t len,
74 wget_css_parse_uri_callback *callback_uri,
75 wget_css_parse_encoding_callback *callback_encoding,
76 void *user_ctx)
77 {
78 int token;
79 size_t length, pos = 0;
80 char *text;
81 yyscan_t scanner;
82
83 yylex_init(&scanner);
84 yy_scan_bytes(buf, (int) len, scanner);
85
86 while ((token = yylex(scanner)) != CSSEOF) {
87 if (token == IMPORT_SYM) {
88 // e.g. @import "https://example.com/index.html"
89 pos += yyget_leng(scanner);
90
91 // skip whitespace before URI/STRING
92 while ((token = yylex(scanner)) == S)
93 pos += yyget_leng(scanner);
94
95 // now token should be STRING or URI
96 if (token == STRING)
97 token = URI;
98 }
99
100 if (token == URI && callback_uri) {
101 // e.g. url(https://example.com/index.html)
102 text = yyget_text(scanner);
103 length = yyget_leng(scanner);
104
105 if (*text == '\'' || *text == '\"') {
106 // a string - remove the quotes
107 callback_uri(user_ctx, text + 1, length - 2, pos + 1);
108 } else {
109 // extract URI from url(...)
110 if (!wget_strncasecmp_ascii(text, "url(", 4)) {
111 char *otext = text;
112
113 // remove trailing ) and any spaces before
114 for (length--; c_isspace(text[length - 1]); length--);
115
116 // remove leading url( and any spaces after
117 for (length -= 4, text += 4; length && c_isspace(*text); text++, length--);
118
119 // remove quotes
120 if (length && (*text == '\'' || *text == '\"')) {
121 text++;
122 length--;
123 }
124
125 if (length && (text[length - 1] == '\'' || text[length - 1] == '\"'))
126 length--;
127
128 callback_uri(user_ctx, text, length, pos + (text - otext));
129 }
130 }
131 } else if (token == CHARSET_SYM && callback_encoding) {
132 // e.g. @charset "UTF-8"
133 pos += yyget_leng(scanner);
134
135 // skip whitespace before charset name
136 while ((token = yylex(scanner)) == S)
137 pos += yyget_leng(scanner);
138
139 // now token should be STRING
140 if (token == STRING) {
141 text = yyget_text(scanner);
142 length = yyget_leng(scanner);
143
144 if (*text == '\'' || *text == '\"') {
145 // a string - remove the quotes
146 callback_encoding(user_ctx, text + 1, length - 2);
147 } else {
148 // a string without quotes
149 callback_encoding(user_ctx, text, length);
150 }
151 } else {
152 error_printf(_("Unknown token after @charset: %d\n"), token);
153 }
154 }
155 pos += yyget_leng(scanner);
156 }
157
158 yylex_destroy(scanner);
159 }
160
wget_css_parse_file(const char * fname,wget_css_parse_uri_callback * callback_uri,wget_css_parse_encoding_callback * callback_encoding,void * user_ctx)161 void wget_css_parse_file(
162 const char *fname,
163 wget_css_parse_uri_callback *callback_uri,
164 wget_css_parse_encoding_callback *callback_encoding,
165 void *user_ctx)
166 {
167 if (strcmp(fname,"-")) {
168 int fd;
169
170 if ((fd = open(fname, O_RDONLY|O_BINARY)) != -1) {
171 struct stat st;
172 if (fstat(fd, &st) == 0) {
173 #ifdef HAVE_MMAP
174 size_t nread = st.st_size;
175 char *buf = mmap(NULL, nread + 1, PROT_READ|PROT_WRITE, MAP_PRIVATE, fd, 0);
176 #else
177 char *buf=wget_malloc(st.st_size+1);
178 size_t nread=read(fd,buf,st.st_size);
179 #endif
180
181 if (nread > 0) {
182 buf[nread] = 0; // PROT_WRITE allows this write, MAP_PRIVATE prevents changes in underlying file system
183 wget_css_parse_buffer(buf, st.st_size, callback_uri, callback_encoding, user_ctx);
184 }
185
186 #ifdef HAVE_MMAP
187 munmap(buf, nread);
188 #else
189 xfree(buf);
190 #endif
191 }
192 close(fd);
193 } else
194 error_printf(_("Failed to open %s\n"), fname);
195 } else {
196 // read data from STDIN.
197 // maybe should use yy_scan_bytes instead of buffering into memory.
198 char tmp[4096];
199 ssize_t nbytes;
200 wget_buffer buf;
201
202 wget_buffer_init(&buf, NULL, 4096);
203
204 while ((nbytes = read(STDIN_FILENO, tmp, sizeof(tmp))) > 0) {
205 wget_buffer_memcat(&buf, tmp, nbytes);
206 }
207
208 if (buf.length)
209 wget_css_parse_buffer(buf.data, buf.length, callback_uri, callback_encoding, user_ctx);
210
211 wget_buffer_deinit(&buf);
212 }
213 }
214