http_parser.rb 0.6.0 → 0.8.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
@@ -12,6 +12,7 @@
12
12
  # RuntimeLibrary MUST MATCH across the entire project
13
13
  'Debug': {
14
14
  'defines': [ 'DEBUG', '_DEBUG' ],
15
+ 'cflags': [ '-Wall', '-Wextra', '-O0', '-g', '-ftrapv' ],
15
16
  'msvs_settings': {
16
17
  'VCCLCompilerTool': {
17
18
  'RuntimeLibrary': 1, # static debug
@@ -20,6 +21,7 @@
20
21
  },
21
22
  'Release': {
22
23
  'defines': [ 'NDEBUG' ],
24
+ 'cflags': [ '-Wall', '-Wextra', '-O3' ],
23
25
  'msvs_settings': {
24
26
  'VCCLCompilerTool': {
25
27
  'RuntimeLibrary': 0, # static release
@@ -51,6 +53,7 @@
51
53
  'type': 'static_library',
52
54
  'include_dirs': [ '.' ],
53
55
  'direct_dependent_settings': {
56
+ 'defines': [ 'HTTP_PARSER_STRICT=0' ],
54
57
  'include_dirs': [ '.' ],
55
58
  },
56
59
  'defines': [ 'HTTP_PARSER_STRICT=0' ],
@@ -69,11 +72,40 @@
69
72
  },
70
73
 
71
74
  {
72
- 'target_name': 'test',
75
+ 'target_name': 'http_parser_strict',
76
+ 'type': 'static_library',
77
+ 'include_dirs': [ '.' ],
78
+ 'direct_dependent_settings': {
79
+ 'defines': [ 'HTTP_PARSER_STRICT=1' ],
80
+ 'include_dirs': [ '.' ],
81
+ },
82
+ 'defines': [ 'HTTP_PARSER_STRICT=1' ],
83
+ 'sources': [ './http_parser.c', ],
84
+ 'conditions': [
85
+ ['OS=="win"', {
86
+ 'msvs_settings': {
87
+ 'VCCLCompilerTool': {
88
+ # Compile as C++. http_parser.c is actually C99, but C++ is
89
+ # close enough in this case.
90
+ 'CompileAs': 2,
91
+ },
92
+ },
93
+ }]
94
+ ],
95
+ },
96
+
97
+ {
98
+ 'target_name': 'test-nonstrict',
73
99
  'type': 'executable',
74
100
  'dependencies': [ 'http_parser' ],
75
101
  'sources': [ 'test.c' ]
102
+ },
103
+
104
+ {
105
+ 'target_name': 'test-strict',
106
+ 'type': 'executable',
107
+ 'dependencies': [ 'http_parser_strict' ],
108
+ 'sources': [ 'test.c' ]
76
109
  }
77
110
  ]
78
111
  }
79
-
@@ -24,11 +24,15 @@
24
24
  extern "C" {
25
25
  #endif
26
26
 
27
- #define HTTP_PARSER_VERSION_MAJOR 1
28
- #define HTTP_PARSER_VERSION_MINOR 0
29
-
30
- #include <sys/types.h>
31
- #if defined(_WIN32) && !defined(__MINGW32__) && (!defined(_MSC_VER) || _MSC_VER<1600)
27
+ /* Also update SONAME in the Makefile whenever you change these. */
28
+ #define HTTP_PARSER_VERSION_MAJOR 2
29
+ #define HTTP_PARSER_VERSION_MINOR 8
30
+ #define HTTP_PARSER_VERSION_PATCH 1
31
+
32
+ #include <stddef.h>
33
+ #if defined(_WIN32) && !defined(__MINGW32__) && \
34
+ (!defined(_MSC_VER) || _MSC_VER<1600) && !defined(__WINE__)
35
+ #include <BaseTsd.h>
32
36
  typedef __int8 int8_t;
33
37
  typedef unsigned __int8 uint8_t;
34
38
  typedef __int16 int16_t;
@@ -37,9 +41,6 @@ typedef __int32 int32_t;
37
41
  typedef unsigned __int32 uint32_t;
38
42
  typedef __int64 int64_t;
39
43
  typedef unsigned __int64 uint64_t;
40
-
41
- typedef unsigned int size_t;
42
- typedef int ssize_t;
43
44
  #else
44
45
  #include <stdint.h>
45
46
  #endif
@@ -51,21 +52,19 @@ typedef int ssize_t;
51
52
  # define HTTP_PARSER_STRICT 1
52
53
  #endif
53
54
 
54
- /* Compile with -DHTTP_PARSER_DEBUG=1 to add extra debugging information to
55
- * the error reporting facility.
55
+ /* Maximium header size allowed. If the macro is not defined
56
+ * before including this header then the default is used. To
57
+ * change the maximum header size, define the macro in the build
58
+ * environment (e.g. -DHTTP_MAX_HEADER_SIZE=<value>). To remove
59
+ * the effective limit on the size of the header, define the macro
60
+ * to a very large number (e.g. -DHTTP_MAX_HEADER_SIZE=0x7fffffff)
56
61
  */
57
- #ifndef HTTP_PARSER_DEBUG
58
- # define HTTP_PARSER_DEBUG 0
62
+ #ifndef HTTP_MAX_HEADER_SIZE
63
+ # define HTTP_MAX_HEADER_SIZE (80*1024)
59
64
  #endif
60
65
 
61
-
62
- /* Maximium header size allowed */
63
- #define HTTP_MAX_HEADER_SIZE (80*1024)
64
-
65
-
66
66
  typedef struct http_parser http_parser;
67
67
  typedef struct http_parser_settings http_parser_settings;
68
- typedef struct http_parser_result http_parser_result;
69
68
 
70
69
 
71
70
  /* Callbacks should return non-zero to indicate an error. The parser will
@@ -77,46 +76,139 @@ typedef struct http_parser_result http_parser_result;
77
76
  * HEAD request which may contain 'Content-Length' or 'Transfer-Encoding:
78
77
  * chunked' headers that indicate the presence of a body.
79
78
  *
80
- * http_data_cb does not return data chunks. It will be call arbitrarally
81
- * many times for each string. E.G. you might get 10 callbacks for "on_path"
79
+ * Returning `2` from on_headers_complete will tell parser that it should not
80
+ * expect neither a body nor any futher responses on this connection. This is
81
+ * useful for handling responses to a CONNECT request which may not contain
82
+ * `Upgrade` or `Connection: upgrade` headers.
83
+ *
84
+ * http_data_cb does not return data chunks. It will be called arbitrarily
85
+ * many times for each string. E.G. you might get 10 callbacks for "on_url"
82
86
  * each providing just a few characters more data.
83
87
  */
84
88
  typedef int (*http_data_cb) (http_parser*, const char *at, size_t length);
85
89
  typedef int (*http_cb) (http_parser*);
86
90
 
87
91
 
92
+ /* Status Codes */
93
+ #define HTTP_STATUS_MAP(XX) \
94
+ XX(100, CONTINUE, Continue) \
95
+ XX(101, SWITCHING_PROTOCOLS, Switching Protocols) \
96
+ XX(102, PROCESSING, Processing) \
97
+ XX(200, OK, OK) \
98
+ XX(201, CREATED, Created) \
99
+ XX(202, ACCEPTED, Accepted) \
100
+ XX(203, NON_AUTHORITATIVE_INFORMATION, Non-Authoritative Information) \
101
+ XX(204, NO_CONTENT, No Content) \
102
+ XX(205, RESET_CONTENT, Reset Content) \
103
+ XX(206, PARTIAL_CONTENT, Partial Content) \
104
+ XX(207, MULTI_STATUS, Multi-Status) \
105
+ XX(208, ALREADY_REPORTED, Already Reported) \
106
+ XX(226, IM_USED, IM Used) \
107
+ XX(300, MULTIPLE_CHOICES, Multiple Choices) \
108
+ XX(301, MOVED_PERMANENTLY, Moved Permanently) \
109
+ XX(302, FOUND, Found) \
110
+ XX(303, SEE_OTHER, See Other) \
111
+ XX(304, NOT_MODIFIED, Not Modified) \
112
+ XX(305, USE_PROXY, Use Proxy) \
113
+ XX(307, TEMPORARY_REDIRECT, Temporary Redirect) \
114
+ XX(308, PERMANENT_REDIRECT, Permanent Redirect) \
115
+ XX(400, BAD_REQUEST, Bad Request) \
116
+ XX(401, UNAUTHORIZED, Unauthorized) \
117
+ XX(402, PAYMENT_REQUIRED, Payment Required) \
118
+ XX(403, FORBIDDEN, Forbidden) \
119
+ XX(404, NOT_FOUND, Not Found) \
120
+ XX(405, METHOD_NOT_ALLOWED, Method Not Allowed) \
121
+ XX(406, NOT_ACCEPTABLE, Not Acceptable) \
122
+ XX(407, PROXY_AUTHENTICATION_REQUIRED, Proxy Authentication Required) \
123
+ XX(408, REQUEST_TIMEOUT, Request Timeout) \
124
+ XX(409, CONFLICT, Conflict) \
125
+ XX(410, GONE, Gone) \
126
+ XX(411, LENGTH_REQUIRED, Length Required) \
127
+ XX(412, PRECONDITION_FAILED, Precondition Failed) \
128
+ XX(413, PAYLOAD_TOO_LARGE, Payload Too Large) \
129
+ XX(414, URI_TOO_LONG, URI Too Long) \
130
+ XX(415, UNSUPPORTED_MEDIA_TYPE, Unsupported Media Type) \
131
+ XX(416, RANGE_NOT_SATISFIABLE, Range Not Satisfiable) \
132
+ XX(417, EXPECTATION_FAILED, Expectation Failed) \
133
+ XX(421, MISDIRECTED_REQUEST, Misdirected Request) \
134
+ XX(422, UNPROCESSABLE_ENTITY, Unprocessable Entity) \
135
+ XX(423, LOCKED, Locked) \
136
+ XX(424, FAILED_DEPENDENCY, Failed Dependency) \
137
+ XX(426, UPGRADE_REQUIRED, Upgrade Required) \
138
+ XX(428, PRECONDITION_REQUIRED, Precondition Required) \
139
+ XX(429, TOO_MANY_REQUESTS, Too Many Requests) \
140
+ XX(431, REQUEST_HEADER_FIELDS_TOO_LARGE, Request Header Fields Too Large) \
141
+ XX(451, UNAVAILABLE_FOR_LEGAL_REASONS, Unavailable For Legal Reasons) \
142
+ XX(500, INTERNAL_SERVER_ERROR, Internal Server Error) \
143
+ XX(501, NOT_IMPLEMENTED, Not Implemented) \
144
+ XX(502, BAD_GATEWAY, Bad Gateway) \
145
+ XX(503, SERVICE_UNAVAILABLE, Service Unavailable) \
146
+ XX(504, GATEWAY_TIMEOUT, Gateway Timeout) \
147
+ XX(505, HTTP_VERSION_NOT_SUPPORTED, HTTP Version Not Supported) \
148
+ XX(506, VARIANT_ALSO_NEGOTIATES, Variant Also Negotiates) \
149
+ XX(507, INSUFFICIENT_STORAGE, Insufficient Storage) \
150
+ XX(508, LOOP_DETECTED, Loop Detected) \
151
+ XX(510, NOT_EXTENDED, Not Extended) \
152
+ XX(511, NETWORK_AUTHENTICATION_REQUIRED, Network Authentication Required) \
153
+
154
+ enum http_status
155
+ {
156
+ #define XX(num, name, string) HTTP_STATUS_##name = num,
157
+ HTTP_STATUS_MAP(XX)
158
+ #undef XX
159
+ };
160
+
161
+
88
162
  /* Request Methods */
163
+ #define HTTP_METHOD_MAP(XX) \
164
+ XX(0, DELETE, DELETE) \
165
+ XX(1, GET, GET) \
166
+ XX(2, HEAD, HEAD) \
167
+ XX(3, POST, POST) \
168
+ XX(4, PUT, PUT) \
169
+ /* pathological */ \
170
+ XX(5, CONNECT, CONNECT) \
171
+ XX(6, OPTIONS, OPTIONS) \
172
+ XX(7, TRACE, TRACE) \
173
+ /* WebDAV */ \
174
+ XX(8, COPY, COPY) \
175
+ XX(9, LOCK, LOCK) \
176
+ XX(10, MKCOL, MKCOL) \
177
+ XX(11, MOVE, MOVE) \
178
+ XX(12, PROPFIND, PROPFIND) \
179
+ XX(13, PROPPATCH, PROPPATCH) \
180
+ XX(14, SEARCH, SEARCH) \
181
+ XX(15, UNLOCK, UNLOCK) \
182
+ XX(16, BIND, BIND) \
183
+ XX(17, REBIND, REBIND) \
184
+ XX(18, UNBIND, UNBIND) \
185
+ XX(19, ACL, ACL) \
186
+ /* subversion */ \
187
+ XX(20, REPORT, REPORT) \
188
+ XX(21, MKACTIVITY, MKACTIVITY) \
189
+ XX(22, CHECKOUT, CHECKOUT) \
190
+ XX(23, MERGE, MERGE) \
191
+ /* upnp */ \
192
+ XX(24, MSEARCH, M-SEARCH) \
193
+ XX(25, NOTIFY, NOTIFY) \
194
+ XX(26, SUBSCRIBE, SUBSCRIBE) \
195
+ XX(27, UNSUBSCRIBE, UNSUBSCRIBE) \
196
+ /* RFC-5789 */ \
197
+ XX(28, PATCH, PATCH) \
198
+ XX(29, PURGE, PURGE) \
199
+ /* CalDAV */ \
200
+ XX(30, MKCALENDAR, MKCALENDAR) \
201
+ /* RFC-2068, section 19.6.1.2 */ \
202
+ XX(31, LINK, LINK) \
203
+ XX(32, UNLINK, UNLINK) \
204
+ /* icecast */ \
205
+ XX(33, SOURCE, SOURCE) \
206
+
89
207
  enum http_method
90
- { HTTP_DELETE = 0
91
- , HTTP_GET
92
- , HTTP_HEAD
93
- , HTTP_POST
94
- , HTTP_PUT
95
- /* pathological */
96
- , HTTP_CONNECT
97
- , HTTP_OPTIONS
98
- , HTTP_TRACE
99
- /* webdav */
100
- , HTTP_COPY
101
- , HTTP_LOCK
102
- , HTTP_MKCOL
103
- , HTTP_MOVE
104
- , HTTP_PROPFIND
105
- , HTTP_PROPPATCH
106
- , HTTP_UNLOCK
107
- /* subversion */
108
- , HTTP_REPORT
109
- , HTTP_MKACTIVITY
110
- , HTTP_CHECKOUT
111
- , HTTP_MERGE
112
- /* upnp */
113
- , HTTP_MSEARCH
114
- , HTTP_NOTIFY
115
- , HTTP_SUBSCRIBE
116
- , HTTP_UNSUBSCRIBE
117
- /* RFC-5789 */
118
- , HTTP_PATCH
119
- , HTTP_PURGE
208
+ {
209
+ #define XX(num, name, string) HTTP_##name = num,
210
+ HTTP_METHOD_MAP(XX)
211
+ #undef XX
120
212
  };
121
213
 
122
214
 
@@ -128,14 +220,16 @@ enum flags
128
220
  { F_CHUNKED = 1 << 0
129
221
  , F_CONNECTION_KEEP_ALIVE = 1 << 1
130
222
  , F_CONNECTION_CLOSE = 1 << 2
131
- , F_TRAILING = 1 << 3
132
- , F_UPGRADE = 1 << 4
133
- , F_SKIPBODY = 1 << 5
223
+ , F_CONNECTION_UPGRADE = 1 << 3
224
+ , F_TRAILING = 1 << 4
225
+ , F_UPGRADE = 1 << 5
226
+ , F_SKIPBODY = 1 << 6
227
+ , F_CONTENTLENGTH = 1 << 7
134
228
  };
135
229
 
136
230
 
137
231
  /* Map for errno-related constants
138
- *
232
+ *
139
233
  * The provided argument should be a macro that takes 2 arguments.
140
234
  */
141
235
  #define HTTP_ERRNO_MAP(XX) \
@@ -150,6 +244,9 @@ enum flags
150
244
  XX(CB_headers_complete, "the on_headers_complete callback failed") \
151
245
  XX(CB_body, "the on_body callback failed") \
152
246
  XX(CB_message_complete, "the on_message_complete callback failed") \
247
+ XX(CB_status, "the on_status callback failed") \
248
+ XX(CB_chunk_header, "the on_chunk_header callback failed") \
249
+ XX(CB_chunk_complete, "the on_chunk_complete callback failed") \
153
250
  \
154
251
  /* Parsing-related errors */ \
155
252
  XX(INVALID_EOF_STATE, "stream ended at an unexpected time") \
@@ -170,6 +267,8 @@ enum flags
170
267
  XX(INVALID_HEADER_TOKEN, "invalid character in header") \
171
268
  XX(INVALID_CONTENT_LENGTH, \
172
269
  "invalid character in content-length header") \
270
+ XX(UNEXPECTED_CONTENT_LENGTH, \
271
+ "unexpected content-length header") \
173
272
  XX(INVALID_CHUNK_SIZE, \
174
273
  "invalid character in chunk size header") \
175
274
  XX(INVALID_CONSTANT, "invalid constant string") \
@@ -190,21 +289,15 @@ enum http_errno {
190
289
  /* Get an http_errno value from an http_parser */
191
290
  #define HTTP_PARSER_ERRNO(p) ((enum http_errno) (p)->http_errno)
192
291
 
193
- /* Get the line number that generated the current error */
194
- #if HTTP_PARSER_DEBUG
195
- #define HTTP_PARSER_ERRNO_LINE(p) ((p)->error_lineno)
196
- #else
197
- #define HTTP_PARSER_ERRNO_LINE(p) 0
198
- #endif
199
-
200
292
 
201
293
  struct http_parser {
202
294
  /** PRIVATE **/
203
- unsigned char type : 2; /* enum http_parser_type */
204
- unsigned char flags : 6; /* F_* values from 'flags' enum; semi-public */
205
- unsigned char state; /* enum state from http_parser.c */
206
- unsigned char header_state; /* enum header_state from http_parser.c */
207
- unsigned char index; /* index into current matcher */
295
+ unsigned int type : 2; /* enum http_parser_type */
296
+ unsigned int flags : 8; /* F_* values from 'flags' enum; semi-public */
297
+ unsigned int state : 7; /* enum state from http_parser.c */
298
+ unsigned int header_state : 7; /* enum header_state from http_parser.c */
299
+ unsigned int index : 7; /* index into current matcher */
300
+ unsigned int lenient_http_headers : 1;
208
301
 
209
302
  uint32_t nread; /* # bytes read in various scenarios */
210
303
  uint64_t content_length; /* # bytes in body (0 if no Content-Length header) */
@@ -212,20 +305,16 @@ struct http_parser {
212
305
  /** READ-ONLY **/
213
306
  unsigned short http_major;
214
307
  unsigned short http_minor;
215
- unsigned short status_code; /* responses only */
216
- unsigned char method; /* requests only */
217
- unsigned char http_errno : 7;
308
+ unsigned int status_code : 16; /* responses only */
309
+ unsigned int method : 8; /* requests only */
310
+ unsigned int http_errno : 7;
218
311
 
219
312
  /* 1 = Upgrade header was present and the parser has exited because of that.
220
313
  * 0 = No upgrade header present.
221
314
  * Should be checked when http_parser_execute() returns in addition to
222
315
  * error checking.
223
316
  */
224
- unsigned char upgrade : 1;
225
-
226
- #if HTTP_PARSER_DEBUG
227
- uint32_t error_lineno;
228
- #endif
317
+ unsigned int upgrade : 1;
229
318
 
230
319
  /** PUBLIC **/
231
320
  void *data; /* A pointer to get hook to the "connection" or "socket" object */
@@ -235,11 +324,17 @@ struct http_parser {
235
324
  struct http_parser_settings {
236
325
  http_cb on_message_begin;
237
326
  http_data_cb on_url;
327
+ http_data_cb on_status;
238
328
  http_data_cb on_header_field;
239
329
  http_data_cb on_header_value;
240
330
  http_cb on_headers_complete;
241
331
  http_data_cb on_body;
242
332
  http_cb on_message_complete;
333
+ /* When on_chunk_header is called, the current chunk length is stored
334
+ * in parser->content_length.
335
+ */
336
+ http_cb on_chunk_header;
337
+ http_cb on_chunk_complete;
243
338
  };
244
339
 
245
340
 
@@ -250,7 +345,8 @@ enum http_parser_url_fields
250
345
  , UF_PATH = 3
251
346
  , UF_QUERY = 4
252
347
  , UF_FRAGMENT = 5
253
- , UF_MAX = 6
348
+ , UF_USERINFO = 6
349
+ , UF_MAX = 7
254
350
  };
255
351
 
256
352
 
@@ -272,9 +368,28 @@ struct http_parser_url {
272
368
  };
273
369
 
274
370
 
371
+ /* Returns the library version. Bits 16-23 contain the major version number,
372
+ * bits 8-15 the minor version number and bits 0-7 the patch level.
373
+ * Usage example:
374
+ *
375
+ * unsigned long version = http_parser_version();
376
+ * unsigned major = (version >> 16) & 255;
377
+ * unsigned minor = (version >> 8) & 255;
378
+ * unsigned patch = version & 255;
379
+ * printf("http_parser v%u.%u.%u\n", major, minor, patch);
380
+ */
381
+ unsigned long http_parser_version(void);
382
+
275
383
  void http_parser_init(http_parser *parser, enum http_parser_type type);
276
384
 
277
385
 
386
+ /* Initialize http_parser_settings members to 0
387
+ */
388
+ void http_parser_settings_init(http_parser_settings *settings);
389
+
390
+
391
+ /* Executes the parser. Returns number of parsed bytes. Sets
392
+ * `parser->http_errno` on error. */
278
393
  size_t http_parser_execute(http_parser *parser,
279
394
  const http_parser_settings *settings,
280
395
  const char *data,
@@ -282,12 +397,12 @@ size_t http_parser_execute(http_parser *parser,
282
397
 
283
398
 
284
399
  /* If http_should_keep_alive() in the on_headers_complete or
285
- * on_message_complete callback returns true, then this will be should be
400
+ * on_message_complete callback returns 0, then this should be
286
401
  * the last message on the connection.
287
402
  * If you are the server, respond with the "Connection: close" header.
288
403
  * If you are the client, close the connection.
289
404
  */
290
- int http_should_keep_alive(http_parser *parser);
405
+ int http_should_keep_alive(const http_parser *parser);
291
406
 
292
407
  /* Returns a string version of the HTTP method. */
293
408
  const char *http_method_str(enum http_method m);
@@ -298,6 +413,9 @@ const char *http_errno_name(enum http_errno err);
298
413
  /* Return a string description of the given error */
299
414
  const char *http_errno_description(enum http_errno err);
300
415
 
416
+ /* Initialize all http_parser_url members to 0 */
417
+ void http_parser_url_init(struct http_parser_url *u);
418
+
301
419
  /* Parse a URL; return nonzero on failure */
302
420
  int http_parser_parse_url(const char *buf, size_t buflen,
303
421
  int is_connect,
@@ -306,6 +424,9 @@ int http_parser_parse_url(const char *buf, size_t buflen,
306
424
  /* Pause or un-pause the parser; a nonzero value pauses */
307
425
  void http_parser_pause(http_parser *parser, int paused);
308
426
 
427
+ /* Checks if this is the final chunk of the body. */
428
+ int http_body_is_final(const http_parser *parser);
429
+
309
430
  #ifdef __cplusplus
310
431
  }
311
432
  #endif