xref: /freebsd/contrib/libucl/src/ucl_internal.h (revision 6b129086dcee14496517fae085b448e3edc69bc7)
1 /* Copyright (c) 2013, Vsevolod Stakhov
2  * All rights reserved.
3  *
4  * Redistribution and use in source and binary forms, with or without
5  * modification, are permitted provided that the following conditions are met:
6  *       * Redistributions of source code must retain the above copyright
7  *         notice, this list of conditions and the following disclaimer.
8  *       * Redistributions in binary form must reproduce the above copyright
9  *         notice, this list of conditions and the following disclaimer in the
10  *         documentation and/or other materials provided with the distribution.
11  *
12  * THIS SOFTWARE IS PROVIDED ''AS IS'' AND ANY
13  * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
14  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
15  * DISCLAIMED. IN NO EVENT SHALL AUTHOR BE LIABLE FOR ANY
16  * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
17  * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
18  * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
19  * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
20  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
21  * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
22  */
23 
24 #ifndef UCL_INTERNAL_H_
25 #define UCL_INTERNAL_H_
26 
27 #ifdef HAVE_CONFIG_H
28 #include "config.h"
29 #else
30 /* Help embedded builds */
31 #define HAVE_SYS_TYPES_H
32 #define HAVE_SYS_MMAN_H
33 #define HAVE_SYS_STAT_H
34 #define HAVE_SYS_PARAM_H
35 #define HAVE_LIMITS_H
36 #define HAVE_FCNTL_H
37 #define HAVE_ERRNO_H
38 #define HAVE_UNISTD_H
39 #define HAVE_CTYPE_H
40 #define HAVE_STDIO_H
41 #define HAVE_STRING_H
42 #define HAVE_FLOAT_H
43 #define HAVE_LIBGEN_H
44 #define HAVE_MATH_H
45 #define HAVE_STDBOOL_H
46 #define HAVE_STDINT_H
47 #define HAVE_STDARG_H
48 #ifndef _WIN32
49 # define HAVE_REGEX_H
50 #endif
51 #endif
52 
53 #ifdef HAVE_SYS_TYPES_H
54 #include <sys/types.h>
55 #endif
56 
57 #ifdef HAVE_SYS_MMAN_H
58 # ifndef _WIN32
59 #  include <sys/mman.h>
60 # endif
61 #endif
62 #ifdef HAVE_SYS_STAT_H
63 #include <sys/stat.h>
64 #endif
65 #ifdef HAVE_SYS_PARAM_H
66 #include <sys/param.h>
67 #endif
68 
69 #ifdef HAVE_LIMITS_H
70 #include <limits.h>
71 #endif
72 #ifdef HAVE_FCNTL_H
73 #include <fcntl.h>
74 #endif
75 #ifdef HAVE_ERRNO_H
76 #include <errno.h>
77 #endif
78 #ifdef HAVE_UNISTD_H
79 #include <unistd.h>
80 #endif
81 #ifdef HAVE_CTYPE_H
82 #include <ctype.h>
83 #endif
84 #ifdef HAVE_STDIO_H
85 #include <stdio.h>
86 #endif
87 #ifdef HAVE_STRING_H
88 #include <string.h>
89 #endif
90 
91 #include "utlist.h"
92 #include "utstring.h"
93 #include "uthash.h"
94 #include "ucl.h"
95 #include "ucl_hash.h"
96 #include "xxhash.h"
97 
98 #ifdef HAVE_OPENSSL
99 #include <openssl/evp.h>
100 #endif
101 
102 #ifndef __DECONST
103 #define __DECONST(type, var)    ((type)(uintptr_t)(const void *)(var))
104 #endif
105 
106 /**
107  * @file rcl_internal.h
108  * Internal structures and functions of UCL library
109  */
110 
111 #define UCL_MAX_RECURSION 16
112 #define UCL_TRASH_KEY 0
113 #define UCL_TRASH_VALUE 1
114 
115 enum ucl_parser_state {
116 	UCL_STATE_INIT = 0,
117 	UCL_STATE_OBJECT,
118 	UCL_STATE_ARRAY,
119 	UCL_STATE_KEY,
120 	UCL_STATE_VALUE,
121 	UCL_STATE_AFTER_VALUE,
122 	UCL_STATE_ARRAY_VALUE,
123 	UCL_STATE_SCOMMENT,
124 	UCL_STATE_MCOMMENT,
125 	UCL_STATE_MACRO_NAME,
126 	UCL_STATE_MACRO,
127 	UCL_STATE_ERROR
128 };
129 
130 enum ucl_character_type {
131 	UCL_CHARACTER_DENIED = 0,
132 	UCL_CHARACTER_KEY = 1,
133 	UCL_CHARACTER_KEY_START = 1 << 1,
134 	UCL_CHARACTER_WHITESPACE = 1 << 2,
135 	UCL_CHARACTER_WHITESPACE_UNSAFE = 1 << 3,
136 	UCL_CHARACTER_VALUE_END = 1 << 4,
137 	UCL_CHARACTER_VALUE_STR = 1 << 5,
138 	UCL_CHARACTER_VALUE_DIGIT = 1 << 6,
139 	UCL_CHARACTER_VALUE_DIGIT_START = 1 << 7,
140 	UCL_CHARACTER_ESCAPE = 1 << 8,
141 	UCL_CHARACTER_KEY_SEP = 1 << 9,
142 	UCL_CHARACTER_JSON_UNSAFE = 1 << 10,
143 	UCL_CHARACTER_UCL_UNSAFE = 1 << 11
144 };
145 
146 struct ucl_macro {
147 	char *name;
148 	ucl_macro_handler handler;
149 	void* ud;
150 	UT_hash_handle hh;
151 };
152 
153 struct ucl_stack {
154 	ucl_object_t *obj;
155 	struct ucl_stack *next;
156 	int level;
157 };
158 
159 struct ucl_chunk {
160 	const unsigned char *begin;
161 	const unsigned char *end;
162 	const unsigned char *pos;
163 	size_t remain;
164 	unsigned int line;
165 	unsigned int column;
166 	struct ucl_chunk *next;
167 };
168 
169 #ifdef HAVE_OPENSSL
170 struct ucl_pubkey {
171 	EVP_PKEY *key;
172 	struct ucl_pubkey *next;
173 };
174 #else
175 struct ucl_pubkey {
176 	struct ucl_pubkey *next;
177 };
178 #endif
179 
180 struct ucl_variable {
181 	char *var;
182 	char *value;
183 	size_t var_len;
184 	size_t value_len;
185 	struct ucl_variable *next;
186 };
187 
188 struct ucl_parser {
189 	enum ucl_parser_state state;
190 	enum ucl_parser_state prev_state;
191 	unsigned int recursion;
192 	int flags;
193 	ucl_object_t *top_obj;
194 	ucl_object_t *cur_obj;
195 	struct ucl_macro *macroes;
196 	struct ucl_stack *stack;
197 	struct ucl_chunk *chunks;
198 	struct ucl_pubkey *keys;
199 	struct ucl_variable *variables;
200 	ucl_variable_handler var_handler;
201 	void *var_data;
202 	UT_string *err;
203 };
204 
205 /**
206  * Unescape json string inplace
207  * @param str
208  */
209 size_t ucl_unescape_json_string (char *str, size_t len);
210 
211 /**
212  * Handle include macro
213  * @param data include data
214  * @param len length of data
215  * @param ud user data
216  * @param err error ptr
217  * @return
218  */
219 bool ucl_include_handler (const unsigned char *data, size_t len, void* ud);
220 
221 bool ucl_try_include_handler (const unsigned char *data, size_t len, void* ud);
222 
223 /**
224  * Handle includes macro
225  * @param data include data
226  * @param len length of data
227  * @param ud user data
228  * @param err error ptr
229  * @return
230  */
231 bool ucl_includes_handler (const unsigned char *data, size_t len, void* ud);
232 
233 size_t ucl_strlcpy (char *dst, const char *src, size_t siz);
234 size_t ucl_strlcpy_unsafe (char *dst, const char *src, size_t siz);
235 size_t ucl_strlcpy_tolower (char *dst, const char *src, size_t siz);
236 
237 
238 #ifdef __GNUC__
239 static inline void
240 ucl_create_err (UT_string **err, const char *fmt, ...)
241 __attribute__ (( format( printf, 2, 3) ));
242 #endif
243 
244 static inline void
245 ucl_create_err (UT_string **err, const char *fmt, ...)
246 
247 {
248 	if (*err == NULL) {
249 		utstring_new (*err);
250 		va_list ap;
251 		va_start (ap, fmt);
252 		utstring_printf_va (*err, fmt, ap);
253 		va_end (ap);
254 	}
255 }
256 
257 /**
258  * Check whether a given string contains a boolean value
259  * @param obj object to set
260  * @param start start of a string
261  * @param len length of a string
262  * @return true if a string is a boolean value
263  */
264 static inline bool
265 ucl_maybe_parse_boolean (ucl_object_t *obj, const unsigned char *start, size_t len)
266 {
267 	const unsigned char *p = start;
268 	bool ret = false, val = false;
269 
270 	if (len == 5) {
271 		if ((p[0] == 'f' || p[0] == 'F') && strncasecmp (p, "false", 5) == 0) {
272 			ret = true;
273 			val = false;
274 		}
275 	}
276 	else if (len == 4) {
277 		if ((p[0] == 't' || p[0] == 'T') && strncasecmp (p, "true", 4) == 0) {
278 			ret = true;
279 			val = true;
280 		}
281 	}
282 	else if (len == 3) {
283 		if ((p[0] == 'y' || p[0] == 'Y') && strncasecmp (p, "yes", 3) == 0) {
284 			ret = true;
285 			val = true;
286 		}
287 		else if ((p[0] == 'o' || p[0] == 'O') && strncasecmp (p, "off", 3) == 0) {
288 			ret = true;
289 			val = false;
290 		}
291 	}
292 	else if (len == 2) {
293 		if ((p[0] == 'n' || p[0] == 'N') && strncasecmp (p, "no", 2) == 0) {
294 			ret = true;
295 			val = false;
296 		}
297 		else if ((p[0] == 'o' || p[0] == 'O') && strncasecmp (p, "on", 2) == 0) {
298 			ret = true;
299 			val = true;
300 		}
301 	}
302 
303 	if (ret) {
304 		obj->type = UCL_BOOLEAN;
305 		obj->value.iv = val;
306 	}
307 
308 	return ret;
309 }
310 
311 /**
312  * Check numeric string
313  * @param obj object to set if a string is numeric
314  * @param start start of string
315  * @param end end of string
316  * @param pos position where parsing has stopped
317  * @param allow_double allow parsing of floating point values
318  * @return 0 if string is numeric and error code (EINVAL or ERANGE) in case of conversion error
319  */
320 int ucl_maybe_parse_number (ucl_object_t *obj,
321 		const char *start, const char *end, const char **pos,
322 		bool allow_double, bool number_bytes, bool allow_time);
323 
324 
325 static inline const ucl_object_t *
326 ucl_hash_search_obj (ucl_hash_t* hashlin, ucl_object_t *obj)
327 {
328 	return (const ucl_object_t *)ucl_hash_search (hashlin, obj->key, obj->keylen);
329 }
330 
331 static inline ucl_hash_t *
332 ucl_hash_insert_object (ucl_hash_t *hashlin, const ucl_object_t *obj) UCL_WARN_UNUSED_RESULT;
333 
334 static inline ucl_hash_t *
335 ucl_hash_insert_object (ucl_hash_t *hashlin, const ucl_object_t *obj)
336 {
337 	if (hashlin == NULL) {
338 		hashlin = ucl_hash_create ();
339 	}
340 	ucl_hash_insert (hashlin, obj, obj->key, obj->keylen);
341 
342 	return hashlin;
343 }
344 
345 /**
346  * Get standard emitter context for a specified emit_type
347  * @param emit_type type of emitter
348  * @return context or NULL if input is invalid
349  */
350 const struct ucl_emitter_context *
351 ucl_emit_get_standard_context (enum ucl_emitter emit_type);
352 
353 /**
354  * Serialise string
355  * @param str string to emit
356  * @param buf target buffer
357  */
358 void ucl_elt_string_write_json (const char *str, size_t size,
359 		struct ucl_emitter_context *ctx);
360 
361 /**
362  * Emit a single object to string
363  * @param obj
364  * @return
365  */
366 unsigned char * ucl_object_emit_single_json (const ucl_object_t *obj);
367 
368 #endif /* UCL_INTERNAL_H_ */
369