1 | /* |
2 | * libwebsockets - small server side websockets and web server implementation |
3 | * |
4 | * Copyright (C) 2010 - 2019 Andy Green <andy@warmcat.com> |
5 | * |
6 | * Permission is hereby granted, free of charge, to any person obtaining a copy |
7 | * of this software and associated documentation files (the "Software"), to |
8 | * deal in the Software without restriction, including without limitation the |
9 | * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or |
10 | * sell copies of the Software, and to permit persons to whom the Software is |
11 | * furnished to do so, subject to the following conditions: |
12 | * |
13 | * The above copyright notice and this permission notice shall be included in |
14 | * all copies or substantial portions of the Software. |
15 | * |
16 | * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR |
17 | * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
18 | * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE |
19 | * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER |
20 | * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING |
21 | * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS |
22 | * IN THE SOFTWARE. |
23 | */ |
24 | |
25 | /* Do not treat - as a terminal character, so "my-token" is one token */ |
26 | #define LWS_TOKENIZE_F_MINUS_NONTERM (1 << 0) |
27 | /* Separately report aggregate colon-delimited tokens */ |
28 | #define LWS_TOKENIZE_F_AGG_COLON (1 << 1) |
29 | /* Enforce sequencing for a simple token , token , token ... list */ |
30 | #define LWS_TOKENIZE_F_COMMA_SEP_LIST (1 << 2) |
31 | /* Allow more characters in the tokens and less delimiters... default is |
32 | * only alphanumeric + underscore in tokens */ |
33 | #define LWS_TOKENIZE_F_RFC7230_DELIMS (1 << 3) |
34 | /* Do not treat . as a terminal character, so "warmcat.com" is one token */ |
35 | #define LWS_TOKENIZE_F_DOT_NONTERM (1 << 4) |
36 | /* If something starts looking like a float, like 1.2, force to be string token. |
37 | * This lets you receive dotted-quads like 192.168.0.1 as string tokens, and |
38 | * avoids illegal float format detection like 1.myserver.com */ |
39 | #define LWS_TOKENIZE_F_NO_FLOATS (1 << 5) |
40 | /* Instead of LWS_TOKZE_INTEGER, report integers as any other string token */ |
41 | #define LWS_TOKENIZE_F_NO_INTEGERS (1 << 6) |
42 | /* # makes the rest of the line a comment */ |
43 | #define (1 << 7) |
44 | /* Do not treat / as a terminal character, so "multipart/related" is one token */ |
45 | #define LWS_TOKENIZE_F_SLASH_NONTERM (1 << 8) |
46 | /* Do not treat * as a terminal character, so "myfile*" is one token */ |
47 | #define LWS_TOKENIZE_F_ASTERISK_NONTERM (1 << 9) |
48 | /* Do not treat = as a terminal character, so "x=y" is one token */ |
49 | #define LWS_TOKENIZE_F_EQUALS_NONTERM (1 << 10) |
50 | |
51 | typedef enum { |
52 | |
53 | LWS_TOKZE_ERRS = 5, /* the number of errors defined */ |
54 | |
55 | LWS_TOKZE_ERR_BROKEN_UTF8 = -5, /* malformed or partial utf8 */ |
56 | LWS_TOKZE_ERR_UNTERM_STRING = -4, /* ended while we were in "" */ |
57 | LWS_TOKZE_ERR_MALFORMED_FLOAT = -3, /* like 0..1 or 0.1.1 */ |
58 | LWS_TOKZE_ERR_NUM_ON_LHS = -2, /* like 123= or 0.1= */ |
59 | LWS_TOKZE_ERR_COMMA_LIST = -1, /* like ",tok", or, "tok,," */ |
60 | |
61 | LWS_TOKZE_ENDED = 0, /* no more content */ |
62 | |
63 | /* Note: results have ordinal 1+, EOT is 0 and errors are < 0 */ |
64 | |
65 | LWS_TOKZE_DELIMITER, /* a delimiter appeared */ |
66 | LWS_TOKZE_TOKEN, /* a token appeared */ |
67 | LWS_TOKZE_INTEGER, /* an integer appeared */ |
68 | LWS_TOKZE_FLOAT, /* a float appeared */ |
69 | LWS_TOKZE_TOKEN_NAME_EQUALS, /* token [whitespace] = */ |
70 | LWS_TOKZE_TOKEN_NAME_COLON, /* token [whitespace] : (only with |
71 | LWS_TOKENIZE_F_AGG_COLON flag) */ |
72 | LWS_TOKZE_QUOTED_STRING, /* "*", where * may have any char */ |
73 | |
74 | } lws_tokenize_elem; |
75 | |
76 | /* |
77 | * helper enums to allow caller to enforce legal delimiter sequencing, eg |
78 | * disallow "token,,token", "token,", and ",token" |
79 | */ |
80 | |
81 | enum lws_tokenize_delimiter_tracking { |
82 | LWSTZ_DT_NEED_FIRST_CONTENT, |
83 | LWSTZ_DT_NEED_DELIM, |
84 | LWSTZ_DT_NEED_NEXT_CONTENT, |
85 | }; |
86 | |
87 | typedef struct lws_tokenize { |
88 | const char *start; /**< set to the start of the string to tokenize */ |
89 | const char *token; /**< the start of an identified token or delimiter */ |
90 | size_t len; /**< set to the length of the string to tokenize */ |
91 | size_t token_len; /**< the length of the identied token or delimiter */ |
92 | |
93 | uint16_t flags; /**< optional LWS_TOKENIZE_F_ flags, or 0 */ |
94 | uint8_t delim; |
95 | |
96 | int8_t e; /**< convenient for storing lws_tokenize return */ |
97 | } lws_tokenize_t; |
98 | |
99 | /** |
100 | * lws_tokenize() - breaks down a string into tokens and delimiters in-place |
101 | * |
102 | * \param ts: the lws_tokenize struct to init |
103 | * \param start: the string to tokenize |
104 | * \param flags: LWS_TOKENIZE_F_ option flags |
105 | * |
106 | * This initializes the tokenize struct to point to the given string, and |
107 | * sets the length to 2GiB - 1 (so there must be a terminating NUL)... you can |
108 | * override this requirement by setting ts.len yourself before using it. |
109 | * |
110 | * .delim is also initialized to LWSTZ_DT_NEED_FIRST_CONTENT. |
111 | */ |
112 | |
113 | LWS_VISIBLE LWS_EXTERN void |
114 | lws_tokenize_init(struct lws_tokenize *ts, const char *start, int flags); |
115 | |
116 | /** |
117 | * lws_tokenize() - breaks down a string into tokens and delimiters in-place |
118 | * |
119 | * \param ts: the lws_tokenize struct with information and state on what to do |
120 | * |
121 | * The \p ts struct should have its start, len and flags members initialized to |
122 | * reflect the string to be tokenized and any options. |
123 | * |
124 | * Then `lws_tokenize()` may be called repeatedly on the struct, returning one |
125 | * of `lws_tokenize_elem` each time, and with the struct's `token` and |
126 | * `token_len` members set to describe the content of the delimiter or token |
127 | * payload each time. |
128 | * |
129 | * There are no allocations during the process. |
130 | * |
131 | * returns lws_tokenize_elem that was identified (LWS_TOKZE_ENDED means reached |
132 | * the end of the string). |
133 | */ |
134 | |
135 | LWS_VISIBLE LWS_EXTERN lws_tokenize_elem |
136 | lws_tokenize(struct lws_tokenize *ts); |
137 | |
138 | /** |
139 | * lws_tokenize_cstr() - copy token string to NUL-terminated buffer |
140 | * |
141 | * \param ts: pointer to lws_tokenize struct to operate on |
142 | * \param str: destination buffer |
143 | * \pparam max: bytes in destination buffer |
144 | * |
145 | * returns 0 if OK or nonzero if the string + NUL won't fit. |
146 | */ |
147 | |
148 | LWS_VISIBLE LWS_EXTERN int |
149 | lws_tokenize_cstr(struct lws_tokenize *ts, char *str, size_t max); |
150 | |
151 | |
152 | /* |
153 | * lws_strexp: flexible string expansion helper api |
154 | * |
155 | * This stateful helper can handle multiple separate input chunks and multiple |
156 | * output buffer loads with arbitrary boundaries between literals and expanded |
157 | * symbols. This allows it to handle fragmented input as well as arbitrarily |
158 | * long symbol expansions that are bigger than the output buffer itself. |
159 | * |
160 | * A user callback is used to convert symbol names to the symbol value. |
161 | * |
162 | * A single byte buffer for input and another for output can process any |
163 | * length substitution then. The state object is around 64 bytes on a 64-bit |
164 | * system and it only uses 8 bytes stack. |
165 | */ |
166 | |
167 | |
168 | typedef int (*lws_strexp_expand_cb)(void *priv, const char *name, char *out, |
169 | size_t *pos, size_t olen, size_t *exp_ofs); |
170 | |
171 | typedef struct lws_strexp { |
172 | char name[32]; |
173 | lws_strexp_expand_cb cb; |
174 | void *priv; |
175 | char *out; |
176 | size_t olen; |
177 | size_t pos; |
178 | |
179 | size_t exp_ofs; |
180 | |
181 | uint8_t name_pos; |
182 | char state; |
183 | } lws_strexp_t; |
184 | |
185 | enum { |
186 | LSTRX_DONE, /* it completed OK */ |
187 | LSTRX_FILLED_OUT, /* out buf filled and needs resetting */ |
188 | LSTRX_FATAL_NAME_TOO_LONG = -1, /* fatal */ |
189 | LSTRX_FATAL_NAME_UNKNOWN = -2, |
190 | }; |
191 | |
192 | |
193 | /** |
194 | * lws_strexp_init() - initialize an lws_strexp_t for use |
195 | * |
196 | * \p exp: the exp object to init |
197 | * \p priv: the user's object pointer to pass to callback |
198 | * \p cb: the callback to expand named objects |
199 | * \p out: the start of the output buffer, or NULL just to get the length |
200 | * \p olen: the length of the output buffer in bytes |
201 | * |
202 | * Prepares an lws_strexp_t for use and sets the initial output buffer |
203 | * |
204 | * If \p out is NULL, substitution proceeds normally, but no output is produced, |
205 | * only the length is returned. olen should be set to the largest feasible |
206 | * overall length. To use this mode, the substitution callback must also check |
207 | * for NULL \p out and avoid producing the output. |
208 | */ |
209 | LWS_VISIBLE LWS_EXTERN void |
210 | lws_strexp_init(lws_strexp_t *exp, void *priv, lws_strexp_expand_cb cb, |
211 | char *out, size_t olen); |
212 | |
213 | /** |
214 | * lws_strexp_reset_out() - reset the output buffer on an existing strexp |
215 | * |
216 | * \p exp: the exp object to init |
217 | * \p out: the start of the output buffer, or NULL to just get length |
218 | * \p olen: the length of the output buffer in bytes |
219 | * |
220 | * Provides a new output buffer for lws_strexp_expand() to continue to write |
221 | * into. It can be the same as the old one if it has been copied out or used. |
222 | * The position of the next write will be reset to the start of the given buf. |
223 | * |
224 | * If \p out is NULL, substitution proceeds normally, but no output is produced, |
225 | * only the length is returned. \p olen should be set to the largest feasible |
226 | * overall length. To use this mode, the substitution callback must also check |
227 | * for NULL \p out and avoid producing the output. |
228 | */ |
229 | LWS_VISIBLE LWS_EXTERN void |
230 | lws_strexp_reset_out(lws_strexp_t *exp, char *out, size_t olen); |
231 | |
232 | /** |
233 | * lws_strexp_expand() - copy / expand a string into the output buffer |
234 | * |
235 | * \p exp: the exp object for the copy / expansion |
236 | * \p in: the start of the next input data |
237 | * \p len: the length of the input data |
238 | * \p pused_in: pointer to write the amount of input used |
239 | * \p pused_out: pointer to write the amount of output used |
240 | * |
241 | * Copies in to the output buffer set in exp, expanding any ${name} tokens using |
242 | * the callback. \p *pused_in is set to the number of input chars used and |
243 | * \p *pused_out the number of output characters used |
244 | * |
245 | * May return LSTRX_FILLED_OUT early with *pused < len if the output buffer is |
246 | * filled. Handle the output buffer and reset it with lws_strexp_reset_out() |
247 | * before calling again with adjusted in / len to continue. |
248 | * |
249 | * In the case of large expansions, the expansion itself may fill the output |
250 | * buffer, in which case the expansion callback returns the LSTRX_FILLED_OUT |
251 | * and will be called again to continue with its *exp_ofs parameter set |
252 | * appropriately. |
253 | */ |
254 | LWS_VISIBLE LWS_EXTERN int |
255 | lws_strexp_expand(lws_strexp_t *exp, const char *in, size_t len, |
256 | size_t *pused_in, size_t *pused_out); |
257 | |
258 | /** |
259 | * lws_strcmp_wildcard() - strcmp but the first arg can have wildcards |
260 | * |
261 | * \p wildcard: a string that may contain zero to three *, and may lack a NUL |
262 | * \p wlen: length of the wildcard string |
263 | * \p check: string to test to see if it matches wildcard |
264 | * \p clen: length of check string |
265 | * |
266 | * Like strcmp, but supports patterns like "a*", "a*b", "a*b*" etc |
267 | * where a and b are arbitrary substrings. Both the wc and check strings need |
268 | * not be NUL terminated, but are specified by lengths. |
269 | */ |
270 | LWS_VISIBLE LWS_EXTERN int |
271 | lws_strcmp_wildcard(const char *wildcard, size_t wlen, const char *check, |
272 | size_t clen); |
273 | |