diff --git a/examples/http-parser/AUTHORS b/examples/http-parser/AUTHORS new file mode 100644 index 0000000..abe99de --- /dev/null +++ b/examples/http-parser/AUTHORS @@ -0,0 +1,32 @@ +# Authors ordered by first contribution. +Ryan Dahl +Jeremy Hinegardner +Sergey Shepelev +Joe Damato +tomika +Phoenix Sol +Cliff Frey +Ewen Cheslack-Postava +Santiago Gala +Tim Becker +Jeff Terrace +Ben Noordhuis +Nathan Rajlich +Mark Nottingham +Aman Gupta +Tim Becker +Sean Cunningham +Peter Griess +Salman Haq +Cliff Frey +Jon Kolb +Fouad Mardini +Paul Querna +Felix Geisendörfer +koichik +Andre Caron +Ivo Raisr +James McLaughlin +David Gwynne +LE ROUX Thomas +Randy Rizun diff --git a/examples/http-parser/CONTRIBUTIONS b/examples/http-parser/CONTRIBUTIONS new file mode 100644 index 0000000..11ba31e --- /dev/null +++ b/examples/http-parser/CONTRIBUTIONS @@ -0,0 +1,4 @@ +Contributors must agree to the Contributor License Agreement before patches +can be accepted. + +http://spreadsheets2.google.com/viewform?hl=en&formkey=dDJXOGUwbzlYaWM4cHN1MERwQS1CSnc6MQ diff --git a/examples/http-parser/LICENSE-MIT b/examples/http-parser/LICENSE-MIT new file mode 100644 index 0000000..58010b3 --- /dev/null +++ b/examples/http-parser/LICENSE-MIT @@ -0,0 +1,23 @@ +http_parser.c is based on src/http/ngx_http_parse.c from NGINX copyright +Igor Sysoev. + +Additional changes are licensed under the same terms as NGINX and +copyright Joyent, Inc. and other Node contributors. All rights reserved. + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to +deal in the Software without restriction, including without limitation the +rights to use, copy, modify, merge, publish, distribute, sublicense, and/or +sell copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in +all copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING +FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS +IN THE SOFTWARE. diff --git a/examples/http-parser/README.md b/examples/http-parser/README.md new file mode 100644 index 0000000..3815e36 --- /dev/null +++ b/examples/http-parser/README.md @@ -0,0 +1,178 @@ +HTTP Parser +=========== + +This is a parser for HTTP messages written in C. It parses both requests and +responses. The parser is designed to be used in performance HTTP +applications. It does not make any syscalls nor allocations, it does not +buffer data, it can be interrupted at anytime. Depending on your +architecture, it only requires about 40 bytes of data per message +stream (in a web server that is per connection). + +Features: + + * No dependencies + * Handles persistent streams (keep-alive). + * Decodes chunked encoding. + * Upgrade support + * Defends against buffer overflow attacks. + +The parser extracts the following information from HTTP messages: + + * Header fields and values + * Content-Length + * Request method + * Response status code + * Transfer-Encoding + * HTTP version + * Request URL + * Message body + + +Usage +----- + +One `http_parser` object is used per TCP connection. Initialize the struct +using `http_parser_init()` and set the callbacks. That might look something +like this for a request parser: + + http_parser_settings settings; + settings.on_path = my_path_callback; + settings.on_header_field = my_header_field_callback; + /* ... */ + + http_parser *parser = malloc(sizeof(http_parser)); + http_parser_init(parser, HTTP_REQUEST); + parser->data = my_socket; + +When data is received on the socket execute the parser and check for errors. + + size_t len = 80*1024, nparsed; + char buf[len]; + ssize_t recved; + + recved = recv(fd, buf, len, 0); + + if (recved < 0) { + /* Handle error. */ + } + + /* Start up / continue the parser. + * Note we pass recved==0 to signal that EOF has been recieved. + */ + nparsed = http_parser_execute(parser, &settings, buf, recved); + + if (parser->upgrade) { + /* handle new protocol */ + } else if (nparsed != recved) { + /* Handle error. Usually just close the connection. */ + } + +HTTP needs to know where the end of the stream is. For example, sometimes +servers send responses without Content-Length and expect the client to +consume input (for the body) until EOF. To tell http_parser about EOF, give +`0` as the forth parameter to `http_parser_execute()`. Callbacks and errors +can still be encountered during an EOF, so one must still be prepared +to receive them. + +Scalar valued message information such as `status_code`, `method`, and the +HTTP version are stored in the parser structure. This data is only +temporally stored in `http_parser` and gets reset on each new message. If +this information is needed later, copy it out of the structure during the +`headers_complete` callback. + +The parser decodes the transfer-encoding for both requests and responses +transparently. That is, a chunked encoding is decoded before being sent to +the on_body callback. + + +The Special Problem of Upgrade +------------------------------ + +HTTP supports upgrading the connection to a different protocol. An +increasingly common example of this is the Web Socket protocol which sends +a request like + + GET /demo HTTP/1.1 + Upgrade: WebSocket + Connection: Upgrade + Host: example.com + Origin: http://example.com + WebSocket-Protocol: sample + +followed by non-HTTP data. + +(See http://tools.ietf.org/html/draft-hixie-thewebsocketprotocol-75 for more +information the Web Socket protocol.) + +To support this, the parser will treat this as a normal HTTP message without a +body. Issuing both on_headers_complete and on_message_complete callbacks. However +http_parser_execute() will stop parsing at the end of the headers and return. + +The user is expected to check if `parser->upgrade` has been set to 1 after +`http_parser_execute()` returns. Non-HTTP data begins at the buffer supplied +offset by the return value of `http_parser_execute()`. + + +Callbacks +--------- + +During the `http_parser_execute()` call, the callbacks set in +`http_parser_settings` will be executed. The parser maintains state and +never looks behind, so buffering the data is not necessary. If you need to +save certain data for later usage, you can do that from the callbacks. + +There are two types of callbacks: + +* notification `typedef int (*http_cb) (http_parser*);` + Callbacks: on_message_begin, on_headers_complete, on_message_complete. +* data `typedef int (*http_data_cb) (http_parser*, const char *at, size_t length);` + Callbacks: (requests only) on_uri, + (common) on_header_field, on_header_value, on_body; + +Callbacks must return 0 on success. Returning a non-zero value indicates +error to the parser, making it exit immediately. + +In case you parse HTTP message in chunks (i.e. `read()` request line +from socket, parse, read half headers, parse, etc) your data callbacks +may be called more than once. Http-parser guarantees that data pointer is only +valid for the lifetime of callback. You can also `read()` into a heap allocated +buffer to avoid copying memory around if this fits your application. + +Reading headers may be a tricky task if you read/parse headers partially. +Basically, you need to remember whether last header callback was field or value +and apply following logic: + + (on_header_field and on_header_value shortened to on_h_*) + ------------------------ ------------ -------------------------------------------- + | State (prev. callback) | Callback | Description/action | + ------------------------ ------------ -------------------------------------------- + | nothing (first call) | on_h_field | Allocate new buffer and copy callback data | + | | | into it | + ------------------------ ------------ -------------------------------------------- + | value | on_h_field | New header started. | + | | | Copy current name,value buffers to headers | + | | | list and allocate new buffer for new name | + ------------------------ ------------ -------------------------------------------- + | field | on_h_field | Previous name continues. Reallocate name | + | | | buffer and append callback data to it | + ------------------------ ------------ -------------------------------------------- + | field | on_h_value | Value for current header started. Allocate | + | | | new buffer and copy callback data to it | + ------------------------ ------------ -------------------------------------------- + | value | on_h_value | Value continues. Reallocate value buffer | + | | | and append callback data to it | + ------------------------ ------------ -------------------------------------------- + + +Parsing URLs +------------ + +A simplistic zero-copy URL parser is provided as `http_parser_parse_url()`. +Users of this library may wish to use it to parse URLs constructed from +consecutive `on_url` callbacks. + +See examples of reading in headers: + +* [partial example](http://gist.github.com/155877) in C +* [from http-parser tests](http://github.com/joyent/http-parser/blob/37a0ff8/test.c#L403) in C +* [from Node library](http://github.com/joyent/node/blob/842eaf4/src/http.js#L284) in Javascript diff --git a/examples/http-parser/http_parser.c b/examples/http-parser/http_parser.c new file mode 100644 index 0000000..0c11eb8 --- /dev/null +++ b/examples/http-parser/http_parser.c @@ -0,0 +1,2059 @@ +/* Based on src/http/ngx_http_parse.c from NGINX copyright Igor Sysoev + * + * Additional changes are licensed under the same terms as NGINX and + * copyright Joyent, Inc. and other Node contributors. All rights reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + */ +#include "http_parser.h" +#include +#include +#include +#include +#include +#include + +#ifndef ULLONG_MAX +# define ULLONG_MAX ((uint64_t) -1) /* 2^64-1 */ +#endif + +#ifndef MIN +# define MIN(a,b) ((a) < (b) ? (a) : (b)) +#endif + + +#if HTTP_PARSER_DEBUG +#define SET_ERRNO(e) \ +do { \ + parser->http_errno = (e); \ + parser->error_lineno = __LINE__; \ +} while (0) +#else +#define SET_ERRNO(e) \ +do { \ + parser->http_errno = (e); \ +} while(0) +#endif + + +/* Run the notify callback FOR, returning ER if it fails */ +#define CALLBACK_NOTIFY_(FOR, ER) \ +do { \ + assert(HTTP_PARSER_ERRNO(parser) == HPE_OK); \ + \ + if (settings->on_##FOR) { \ + if (0 != settings->on_##FOR(parser)) { \ + SET_ERRNO(HPE_CB_##FOR); \ + } \ + \ + /* We either errored above or got paused; get out */ \ + if (HTTP_PARSER_ERRNO(parser) != HPE_OK) { \ + return (ER); \ + } \ + } \ +} while (0) + +/* Run the notify callback FOR and consume the current byte */ +#define CALLBACK_NOTIFY(FOR) CALLBACK_NOTIFY_(FOR, p - data + 1) + +/* Run the notify callback FOR and don't consume the current byte */ +#define CALLBACK_NOTIFY_NOADVANCE(FOR) CALLBACK_NOTIFY_(FOR, p - data) + +/* Run data callback FOR with LEN bytes, returning ER if it fails */ +#define CALLBACK_DATA_(FOR, LEN, ER) \ +do { \ + assert(HTTP_PARSER_ERRNO(parser) == HPE_OK); \ + \ + if (FOR##_mark) { \ + if (settings->on_##FOR) { \ + if (0 != settings->on_##FOR(parser, FOR##_mark, (LEN))) { \ + SET_ERRNO(HPE_CB_##FOR); \ + } \ + \ + /* We either errored above or got paused; get out */ \ + if (HTTP_PARSER_ERRNO(parser) != HPE_OK) { \ + return (ER); \ + } \ + } \ + FOR##_mark = NULL; \ + } \ +} while (0) + +/* Run the data callback FOR and consume the current byte */ +#define CALLBACK_DATA(FOR) \ + CALLBACK_DATA_(FOR, p - FOR##_mark, p - data + 1) + +/* Run the data callback FOR and don't consume the current byte */ +#define CALLBACK_DATA_NOADVANCE(FOR) \ + CALLBACK_DATA_(FOR, p - FOR##_mark, p - data) + +/* Set the mark FOR; non-destructive if mark is already set */ +#define MARK(FOR) \ +do { \ + if (!FOR##_mark) { \ + FOR##_mark = p; \ + } \ +} while (0) + + +#define PROXY_CONNECTION "proxy-connection" +#define CONNECTION "connection" +#define CONTENT_LENGTH "content-length" +#define TRANSFER_ENCODING "transfer-encoding" +#define UPGRADE "upgrade" +#define CHUNKED "chunked" +#define KEEP_ALIVE "keep-alive" +#define CLOSE "close" + + +static const char *method_strings[] = + { +#define XX(num, name, string) #string, + HTTP_METHOD_MAP(XX) +#undef XX + }; + + +/* Tokens as defined by rfc 2616. Also lowercases them. + * token = 1* + * separators = "(" | ")" | "<" | ">" | "@" + * | "," | ";" | ":" | "\" | <"> + * | "/" | "[" | "]" | "?" | "=" + * | "{" | "}" | SP | HT + */ +static const char tokens[256] = { +/* 0 nul 1 soh 2 stx 3 etx 4 eot 5 enq 6 ack 7 bel */ + 0, 0, 0, 0, 0, 0, 0, 0, +/* 8 bs 9 ht 10 nl 11 vt 12 np 13 cr 14 so 15 si */ + 0, 0, 0, 0, 0, 0, 0, 0, +/* 16 dle 17 dc1 18 dc2 19 dc3 20 dc4 21 nak 22 syn 23 etb */ + 0, 0, 0, 0, 0, 0, 0, 0, +/* 24 can 25 em 26 sub 27 esc 28 fs 29 gs 30 rs 31 us */ + 0, 0, 0, 0, 0, 0, 0, 0, +/* 32 sp 33 ! 34 " 35 # 36 $ 37 % 38 & 39 ' */ + 0, '!', 0, '#', '$', '%', '&', '\'', +/* 40 ( 41 ) 42 * 43 + 44 , 45 - 46 . 47 / */ + 0, 0, '*', '+', 0, '-', '.', 0, +/* 48 0 49 1 50 2 51 3 52 4 53 5 54 6 55 7 */ + '0', '1', '2', '3', '4', '5', '6', '7', +/* 56 8 57 9 58 : 59 ; 60 < 61 = 62 > 63 ? */ + '8', '9', 0, 0, 0, 0, 0, 0, +/* 64 @ 65 A 66 B 67 C 68 D 69 E 70 F 71 G */ + 0, 'a', 'b', 'c', 'd', 'e', 'f', 'g', +/* 72 H 73 I 74 J 75 K 76 L 77 M 78 N 79 O */ + 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', +/* 80 P 81 Q 82 R 83 S 84 T 85 U 86 V 87 W */ + 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', +/* 88 X 89 Y 90 Z 91 [ 92 \ 93 ] 94 ^ 95 _ */ + 'x', 'y', 'z', 0, 0, 0, '^', '_', +/* 96 ` 97 a 98 b 99 c 100 d 101 e 102 f 103 g */ + '`', 'a', 'b', 'c', 'd', 'e', 'f', 'g', +/* 104 h 105 i 106 j 107 k 108 l 109 m 110 n 111 o */ + 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', +/* 112 p 113 q 114 r 115 s 116 t 117 u 118 v 119 w */ + 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', +/* 120 x 121 y 122 z 123 { 124 | 125 } 126 ~ 127 del */ + 'x', 'y', 'z', 0, '|', 0, '~', 0 }; + + +static const int8_t unhex[256] = + {-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 + ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 + ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 + , 0, 1, 2, 3, 4, 5, 6, 7, 8, 9,-1,-1,-1,-1,-1,-1 + ,-1,10,11,12,13,14,15,-1,-1,-1,-1,-1,-1,-1,-1,-1 + ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 + ,-1,10,11,12,13,14,15,-1,-1,-1,-1,-1,-1,-1,-1,-1 + ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 + }; + + +#if HTTP_PARSER_STRICT +# define T 0 +#else +# define T 1 +#endif + + +static const uint8_t normal_url_char[256] = { +/* 0 nul 1 soh 2 stx 3 etx 4 eot 5 enq 6 ack 7 bel */ + 0, 0, 0, 0, 0, 0, 0, 0, +/* 8 bs 9 ht 10 nl 11 vt 12 np 13 cr 14 so 15 si */ + 0, T, 0, 0, T, 0, 0, 0, +/* 16 dle 17 dc1 18 dc2 19 dc3 20 dc4 21 nak 22 syn 23 etb */ + 0, 0, 0, 0, 0, 0, 0, 0, +/* 24 can 25 em 26 sub 27 esc 28 fs 29 gs 30 rs 31 us */ + 0, 0, 0, 0, 0, 0, 0, 0, +/* 32 sp 33 ! 34 " 35 # 36 $ 37 % 38 & 39 ' */ + 0, 1, 1, 0, 1, 1, 1, 1, +/* 40 ( 41 ) 42 * 43 + 44 , 45 - 46 . 47 / */ + 1, 1, 1, 1, 1, 1, 1, 1, +/* 48 0 49 1 50 2 51 3 52 4 53 5 54 6 55 7 */ + 1, 1, 1, 1, 1, 1, 1, 1, +/* 56 8 57 9 58 : 59 ; 60 < 61 = 62 > 63 ? */ + 1, 1, 1, 1, 1, 1, 1, 0, +/* 64 @ 65 A 66 B 67 C 68 D 69 E 70 F 71 G */ + 1, 1, 1, 1, 1, 1, 1, 1, +/* 72 H 73 I 74 J 75 K 76 L 77 M 78 N 79 O */ + 1, 1, 1, 1, 1, 1, 1, 1, +/* 80 P 81 Q 82 R 83 S 84 T 85 U 86 V 87 W */ + 1, 1, 1, 1, 1, 1, 1, 1, +/* 88 X 89 Y 90 Z 91 [ 92 \ 93 ] 94 ^ 95 _ */ + 1, 1, 1, 1, 1, 1, 1, 1, +/* 96 ` 97 a 98 b 99 c 100 d 101 e 102 f 103 g */ + 1, 1, 1, 1, 1, 1, 1, 1, +/* 104 h 105 i 106 j 107 k 108 l 109 m 110 n 111 o */ + 1, 1, 1, 1, 1, 1, 1, 1, +/* 112 p 113 q 114 r 115 s 116 t 117 u 118 v 119 w */ + 1, 1, 1, 1, 1, 1, 1, 1, +/* 120 x 121 y 122 z 123 { 124 | 125 } 126 ~ 127 del */ + 1, 1, 1, 1, 1, 1, 1, 0, }; + +#undef T + +enum state + { s_dead = 1 /* important that this is > 0 */ + + , s_start_req_or_res + , s_res_or_resp_H + , s_start_res + , s_res_H + , s_res_HT + , s_res_HTT + , s_res_HTTP + , s_res_first_http_major + , s_res_http_major + , s_res_first_http_minor + , s_res_http_minor + , s_res_first_status_code + , s_res_status_code + , s_res_status + , s_res_line_almost_done + + , s_start_req + + , s_req_method + , s_req_spaces_before_url + , s_req_schema + , s_req_schema_slash + , s_req_schema_slash_slash + , s_req_host_start + , s_req_host_v6_start + , s_req_host_v6 + , s_req_host_v6_end + , s_req_host + , s_req_port_start + , s_req_port + , s_req_path + , s_req_query_string_start + , s_req_query_string + , s_req_fragment_start + , s_req_fragment + , s_req_http_start + , s_req_http_H + , s_req_http_HT + , s_req_http_HTT + , s_req_http_HTTP + , s_req_first_http_major + , s_req_http_major + , s_req_first_http_minor + , s_req_http_minor + , s_req_line_almost_done + + , s_header_field_start + , s_header_field + , s_header_value_start + , s_header_value + , s_header_value_lws + + , s_header_almost_done + + , s_chunk_size_start + , s_chunk_size + , s_chunk_parameters + , s_chunk_size_almost_done + + , s_headers_almost_done + , s_headers_done + + /* Important: 's_headers_done' must be the last 'header' state. All + * states beyond this must be 'body' states. It is used for overflow + * checking. See the PARSING_HEADER() macro. + */ + + , s_chunk_data + , s_chunk_data_almost_done + , s_chunk_data_done + + , s_body_identity + , s_body_identity_eof + + , s_message_done + }; + + +#define PARSING_HEADER(state) (state <= s_headers_done) + + +enum header_states + { h_general = 0 + , h_C + , h_CO + , h_CON + + , h_matching_connection + , h_matching_proxy_connection + , h_matching_content_length + , h_matching_transfer_encoding + , h_matching_upgrade + + , h_connection + , h_content_length + , h_transfer_encoding + , h_upgrade + + , h_matching_transfer_encoding_chunked + , h_matching_connection_keep_alive + , h_matching_connection_close + + , h_transfer_encoding_chunked + , h_connection_keep_alive + , h_connection_close + }; + + +/* Macros for character classes; depends on strict-mode */ +#define CR '\r' +#define LF '\n' +#define LOWER(c) (unsigned char)(c | 0x20) +#define IS_ALPHA(c) (LOWER(c) >= 'a' && LOWER(c) <= 'z') +#define IS_NUM(c) ((c) >= '0' && (c) <= '9') +#define IS_ALPHANUM(c) (IS_ALPHA(c) || IS_NUM(c)) +#define IS_HEX(c) (IS_NUM(c) || (LOWER(c) >= 'a' && LOWER(c) <= 'f')) + +#if HTTP_PARSER_STRICT +#define TOKEN(c) (tokens[(unsigned char)c]) +#define IS_URL_CHAR(c) (normal_url_char[(unsigned char) (c)]) +#define IS_HOST_CHAR(c) (IS_ALPHANUM(c) || (c) == '.' || (c) == '-') +#else +#define TOKEN(c) ((c == ' ') ? ' ' : tokens[(unsigned char)c]) +#define IS_URL_CHAR(c) \ + (normal_url_char[(unsigned char) (c)] || ((c) & 0x80)) +#define IS_HOST_CHAR(c) \ + (IS_ALPHANUM(c) || (c) == '.' || (c) == '-' || (c) == '_') +#endif + + +#define start_state (parser->type == HTTP_REQUEST ? s_start_req : s_start_res) + + +#if HTTP_PARSER_STRICT +# define STRICT_CHECK(cond) \ +do { \ + if (cond) { \ + SET_ERRNO(HPE_STRICT); \ + goto error; \ + } \ +} while (0) +# define NEW_MESSAGE() (http_should_keep_alive(parser) ? start_state : s_dead) +#else +# define STRICT_CHECK(cond) +# define NEW_MESSAGE() start_state +#endif + + +/* Map errno values to strings for human-readable output */ +#define HTTP_STRERROR_GEN(n, s) { "HPE_" #n, s }, +static struct { + const char *name; + const char *description; +} http_strerror_tab[] = { + HTTP_ERRNO_MAP(HTTP_STRERROR_GEN) +}; +#undef HTTP_STRERROR_GEN + +int http_message_needs_eof(http_parser *parser); + +/* Our URL parser. + * + * This is designed to be shared by http_parser_execute() for URL validation, + * hence it has a state transition + byte-for-byte interface. In addition, it + * is meant to be embedded in http_parser_parse_url(), which does the dirty + * work of turning state transitions URL components for its API. + * + * This function should only be invoked with non-space characters. It is + * assumed that the caller cares about (and can detect) the transition between + * URL and non-URL states by looking for these. + */ +static enum state +parse_url_char(enum state s, const char ch) +{ + if (ch == ' ' || ch == '\r' || ch == '\n') { + return s_dead; + } + +#if HTTP_PARSER_STRICT + if (ch == '\t' || ch == '\f') { + return s_dead; + } +#endif + + switch (s) { + case s_req_spaces_before_url: + /* Proxied requests are followed by scheme of an absolute URI (alpha). + * All methods except CONNECT are followed by '/' or '*'. + */ + + if (ch == '/' || ch == '*') { + return s_req_path; + } + + if (IS_ALPHA(ch)) { + return s_req_schema; + } + + break; + + case s_req_schema: + if (IS_ALPHA(ch)) { + return s; + } + + if (ch == ':') { + return s_req_schema_slash; + } + + break; + + case s_req_schema_slash: + if (ch == '/') { + return s_req_schema_slash_slash; + } + + break; + + case s_req_schema_slash_slash: + if (ch == '/') { + return s_req_host_start; + } + + break; + + case s_req_host_start: + if (ch == '[') { + return s_req_host_v6_start; + } + + if (IS_HOST_CHAR(ch)) { + return s_req_host; + } + + break; + + case s_req_host: + if (IS_HOST_CHAR(ch)) { + return s_req_host; + } + + /* FALLTHROUGH */ + case s_req_host_v6_end: + switch (ch) { + case ':': + return s_req_port_start; + + case '/': + return s_req_path; + + case '?': + return s_req_query_string_start; + } + + break; + + case s_req_host_v6: + if (ch == ']') { + return s_req_host_v6_end; + } + + /* FALLTHROUGH */ + case s_req_host_v6_start: + if (IS_HEX(ch) || ch == ':') { + return s_req_host_v6; + } + break; + + case s_req_port: + switch (ch) { + case '/': + return s_req_path; + + case '?': + return s_req_query_string_start; + } + + /* FALLTHROUGH */ + case s_req_port_start: + if (IS_NUM(ch)) { + return s_req_port; + } + + break; + + case s_req_path: + if (IS_URL_CHAR(ch)) { + return s; + } + + switch (ch) { + case '?': + return s_req_query_string_start; + + case '#': + return s_req_fragment_start; + } + + break; + + case s_req_query_string_start: + case s_req_query_string: + if (IS_URL_CHAR(ch)) { + return s_req_query_string; + } + + switch (ch) { + case '?': + /* allow extra '?' in query string */ + return s_req_query_string; + + case '#': + return s_req_fragment_start; + } + + break; + + case s_req_fragment_start: + if (IS_URL_CHAR(ch)) { + return s_req_fragment; + } + + switch (ch) { + case '?': + return s_req_fragment; + + case '#': + return s; + } + + break; + + case s_req_fragment: + if (IS_URL_CHAR(ch)) { + return s; + } + + switch (ch) { + case '?': + case '#': + return s; + } + + break; + + default: + break; + } + + /* We should never fall out of the switch above unless there's an error */ + return s_dead; +} + +size_t http_parser_execute (http_parser *parser, + const http_parser_settings *settings, + const char *data, + size_t len) +{ + char c, ch; + int8_t unhex_val; + const char *p = data; + const char *header_field_mark = 0; + const char *header_value_mark = 0; + const char *url_mark = 0; + const char *body_mark = 0; + + /* We're in an error state. Don't bother doing anything. */ + if (HTTP_PARSER_ERRNO(parser) != HPE_OK) { + return 0; + } + + if (len == 0) { + switch (parser->state) { + case s_body_identity_eof: + /* Use of CALLBACK_NOTIFY() here would erroneously return 1 byte read if + * we got paused. + */ + CALLBACK_NOTIFY_NOADVANCE(message_complete); + return 0; + + case s_dead: + case s_start_req_or_res: + case s_start_res: + case s_start_req: + return 0; + + default: + SET_ERRNO(HPE_INVALID_EOF_STATE); + return 1; + } + } + + + if (parser->state == s_header_field) + header_field_mark = data; + if (parser->state == s_header_value) + header_value_mark = data; + switch (parser->state) { + case s_req_path: + case s_req_schema: + case s_req_schema_slash: + case s_req_schema_slash_slash: + case s_req_host_start: + case s_req_host_v6_start: + case s_req_host_v6: + case s_req_host_v6_end: + case s_req_host: + case s_req_port_start: + case s_req_port: + case s_req_query_string_start: + case s_req_query_string: + case s_req_fragment_start: + case s_req_fragment: + url_mark = data; + break; + } + + for (p=data; p != data + len; p++) { + ch = *p; + + if (PARSING_HEADER(parser->state)) { + ++parser->nread; + /* Buffer overflow attack */ + if (parser->nread > HTTP_MAX_HEADER_SIZE) { + SET_ERRNO(HPE_HEADER_OVERFLOW); + goto error; + } + } + + reexecute_byte: + switch (parser->state) { + + case s_dead: + /* this state is used after a 'Connection: close' message + * the parser will error out if it reads another message + */ + if (ch == CR || ch == LF) + break; + + SET_ERRNO(HPE_CLOSED_CONNECTION); + goto error; + + case s_start_req_or_res: + { + if (ch == CR || ch == LF) + break; + parser->flags = 0; + parser->content_length = ULLONG_MAX; + + if (ch == 'H') { + parser->state = s_res_or_resp_H; + + CALLBACK_NOTIFY(message_begin); + } else { + parser->type = HTTP_REQUEST; + parser->state = s_start_req; + goto reexecute_byte; + } + + break; + } + + case s_res_or_resp_H: + if (ch == 'T') { + parser->type = HTTP_RESPONSE; + parser->state = s_res_HT; + } else { + if (ch != 'E') { + SET_ERRNO(HPE_INVALID_CONSTANT); + goto error; + } + + parser->type = HTTP_REQUEST; + parser->method = HTTP_HEAD; + parser->index = 2; + parser->state = s_req_method; + } + break; + + case s_start_res: + { + parser->flags = 0; + parser->content_length = ULLONG_MAX; + + switch (ch) { + case 'H': + parser->state = s_res_H; + break; + + case CR: + case LF: + break; + + default: + SET_ERRNO(HPE_INVALID_CONSTANT); + goto error; + } + + CALLBACK_NOTIFY(message_begin); + break; + } + + case s_res_H: + STRICT_CHECK(ch != 'T'); + parser->state = s_res_HT; + break; + + case s_res_HT: + STRICT_CHECK(ch != 'T'); + parser->state = s_res_HTT; + break; + + case s_res_HTT: + STRICT_CHECK(ch != 'P'); + parser->state = s_res_HTTP; + break; + + case s_res_HTTP: + STRICT_CHECK(ch != '/'); + parser->state = s_res_first_http_major; + break; + + case s_res_first_http_major: + if (ch < '0' || ch > '9') { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_major = ch - '0'; + parser->state = s_res_http_major; + break; + + /* major HTTP version or dot */ + case s_res_http_major: + { + if (ch == '.') { + parser->state = s_res_first_http_minor; + break; + } + + if (!IS_NUM(ch)) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_major *= 10; + parser->http_major += ch - '0'; + + if (parser->http_major > 999) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + break; + } + + /* first digit of minor HTTP version */ + case s_res_first_http_minor: + if (!IS_NUM(ch)) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_minor = ch - '0'; + parser->state = s_res_http_minor; + break; + + /* minor HTTP version or end of request line */ + case s_res_http_minor: + { + if (ch == ' ') { + parser->state = s_res_first_status_code; + break; + } + + if (!IS_NUM(ch)) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_minor *= 10; + parser->http_minor += ch - '0'; + + if (parser->http_minor > 999) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + break; + } + + case s_res_first_status_code: + { + if (!IS_NUM(ch)) { + if (ch == ' ') { + break; + } + + SET_ERRNO(HPE_INVALID_STATUS); + goto error; + } + parser->status_code = ch - '0'; + parser->state = s_res_status_code; + break; + } + + case s_res_status_code: + { + if (!IS_NUM(ch)) { + switch (ch) { + case ' ': + parser->state = s_res_status; + break; + case CR: + parser->state = s_res_line_almost_done; + break; + case LF: + parser->state = s_header_field_start; + break; + default: + SET_ERRNO(HPE_INVALID_STATUS); + goto error; + } + break; + } + + parser->status_code *= 10; + parser->status_code += ch - '0'; + + if (parser->status_code > 999) { + SET_ERRNO(HPE_INVALID_STATUS); + goto error; + } + + break; + } + + case s_res_status: + /* the human readable status. e.g. "NOT FOUND" + * we are not humans so just ignore this */ + if (ch == CR) { + parser->state = s_res_line_almost_done; + break; + } + + if (ch == LF) { + parser->state = s_header_field_start; + break; + } + break; + + case s_res_line_almost_done: + STRICT_CHECK(ch != LF); + parser->state = s_header_field_start; + break; + + case s_start_req: + { + if (ch == CR || ch == LF) + break; + parser->flags = 0; + parser->content_length = ULLONG_MAX; + + if (!IS_ALPHA(ch)) { + SET_ERRNO(HPE_INVALID_METHOD); + goto error; + } + + parser->method = (enum http_method) 0; + parser->index = 1; + switch (ch) { + case 'C': parser->method = HTTP_CONNECT; /* or COPY, CHECKOUT */ break; + case 'D': parser->method = HTTP_DELETE; break; + case 'G': parser->method = HTTP_GET; break; + case 'H': parser->method = HTTP_HEAD; break; + case 'L': parser->method = HTTP_LOCK; break; + case 'M': parser->method = HTTP_MKCOL; /* or MOVE, MKACTIVITY, MERGE, M-SEARCH */ break; + case 'N': parser->method = HTTP_NOTIFY; break; + case 'O': parser->method = HTTP_OPTIONS; break; + case 'P': parser->method = HTTP_POST; + /* or PROPFIND|PROPPATCH|PUT|PATCH|PURGE */ + break; + case 'R': parser->method = HTTP_REPORT; break; + case 'S': parser->method = HTTP_SUBSCRIBE; /* or SEARCH */ break; + case 'T': parser->method = HTTP_TRACE; break; + case 'U': parser->method = HTTP_UNLOCK; /* or UNSUBSCRIBE */ break; + default: + SET_ERRNO(HPE_INVALID_METHOD); + goto error; + } + parser->state = s_req_method; + + CALLBACK_NOTIFY(message_begin); + + break; + } + + case s_req_method: + { + const char *matcher; + if (ch == '\0') { + SET_ERRNO(HPE_INVALID_METHOD); + goto error; + } + + matcher = method_strings[parser->method]; + if (ch == ' ' && matcher[parser->index] == '\0') { + parser->state = s_req_spaces_before_url; + } else if (ch == matcher[parser->index]) { + ; /* nada */ + } else if (parser->method == HTTP_CONNECT) { + if (parser->index == 1 && ch == 'H') { + parser->method = HTTP_CHECKOUT; + } else if (parser->index == 2 && ch == 'P') { + parser->method = HTTP_COPY; + } else { + goto error; + } + } else if (parser->method == HTTP_MKCOL) { + if (parser->index == 1 && ch == 'O') { + parser->method = HTTP_MOVE; + } else if (parser->index == 1 && ch == 'E') { + parser->method = HTTP_MERGE; + } else if (parser->index == 1 && ch == '-') { + parser->method = HTTP_MSEARCH; + } else if (parser->index == 2 && ch == 'A') { + parser->method = HTTP_MKACTIVITY; + } else { + goto error; + } + } else if (parser->method == HTTP_SUBSCRIBE) { + if (parser->index == 1 && ch == 'E') { + parser->method = HTTP_SEARCH; + } else { + goto error; + } + } else if (parser->index == 1 && parser->method == HTTP_POST) { + if (ch == 'R') { + parser->method = HTTP_PROPFIND; /* or HTTP_PROPPATCH */ + } else if (ch == 'U') { + parser->method = HTTP_PUT; /* or HTTP_PURGE */ + } else if (ch == 'A') { + parser->method = HTTP_PATCH; + } else { + goto error; + } + } else if (parser->index == 2) { + if (parser->method == HTTP_PUT) { + if (ch == 'R') parser->method = HTTP_PURGE; + } else if (parser->method == HTTP_UNLOCK) { + if (ch == 'S') parser->method = HTTP_UNSUBSCRIBE; + } + } else if (parser->index == 4 && parser->method == HTTP_PROPFIND && ch == 'P') { + parser->method = HTTP_PROPPATCH; + } else { + SET_ERRNO(HPE_INVALID_METHOD); + goto error; + } + + ++parser->index; + break; + } + + case s_req_spaces_before_url: + { + if (ch == ' ') break; + + MARK(url); + if (parser->method == HTTP_CONNECT) { + parser->state = s_req_host_start; + } + + parser->state = parse_url_char((enum state)parser->state, ch); + if (parser->state == s_dead) { + SET_ERRNO(HPE_INVALID_URL); + goto error; + } + + break; + } + + case s_req_schema: + case s_req_schema_slash: + case s_req_schema_slash_slash: + case s_req_host_start: + case s_req_host_v6_start: + case s_req_host_v6: + case s_req_port_start: + { + switch (ch) { + /* No whitespace allowed here */ + case ' ': + case CR: + case LF: + SET_ERRNO(HPE_INVALID_URL); + goto error; + default: + parser->state = parse_url_char((enum state)parser->state, ch); + if (parser->state == s_dead) { + SET_ERRNO(HPE_INVALID_URL); + goto error; + } + } + + break; + } + + case s_req_host: + case s_req_host_v6_end: + case s_req_port: + case s_req_path: + case s_req_query_string_start: + case s_req_query_string: + case s_req_fragment_start: + case s_req_fragment: + { + switch (ch) { + case ' ': + parser->state = s_req_http_start; + CALLBACK_DATA(url); + break; + case CR: + case LF: + parser->http_major = 0; + parser->http_minor = 9; + parser->state = (ch == CR) ? + s_req_line_almost_done : + s_header_field_start; + CALLBACK_DATA(url); + break; + default: + parser->state = parse_url_char((enum state)parser->state, ch); + if (parser->state == s_dead) { + SET_ERRNO(HPE_INVALID_URL); + goto error; + } + } + break; + } + + case s_req_http_start: + switch (ch) { + case 'H': + parser->state = s_req_http_H; + break; + case ' ': + break; + default: + SET_ERRNO(HPE_INVALID_CONSTANT); + goto error; + } + break; + + case s_req_http_H: + STRICT_CHECK(ch != 'T'); + parser->state = s_req_http_HT; + break; + + case s_req_http_HT: + STRICT_CHECK(ch != 'T'); + parser->state = s_req_http_HTT; + break; + + case s_req_http_HTT: + STRICT_CHECK(ch != 'P'); + parser->state = s_req_http_HTTP; + break; + + case s_req_http_HTTP: + STRICT_CHECK(ch != '/'); + parser->state = s_req_first_http_major; + break; + + /* first digit of major HTTP version */ + case s_req_first_http_major: + if (ch < '1' || ch > '9') { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_major = ch - '0'; + parser->state = s_req_http_major; + break; + + /* major HTTP version or dot */ + case s_req_http_major: + { + if (ch == '.') { + parser->state = s_req_first_http_minor; + break; + } + + if (!IS_NUM(ch)) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_major *= 10; + parser->http_major += ch - '0'; + + if (parser->http_major > 999) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + break; + } + + /* first digit of minor HTTP version */ + case s_req_first_http_minor: + if (!IS_NUM(ch)) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_minor = ch - '0'; + parser->state = s_req_http_minor; + break; + + /* minor HTTP version or end of request line */ + case s_req_http_minor: + { + if (ch == CR) { + parser->state = s_req_line_almost_done; + break; + } + + if (ch == LF) { + parser->state = s_header_field_start; + break; + } + + /* XXX allow spaces after digit? */ + + if (!IS_NUM(ch)) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_minor *= 10; + parser->http_minor += ch - '0'; + + if (parser->http_minor > 999) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + break; + } + + /* end of request line */ + case s_req_line_almost_done: + { + if (ch != LF) { + SET_ERRNO(HPE_LF_EXPECTED); + goto error; + } + + parser->state = s_header_field_start; + break; + } + + case s_header_field_start: + { + if (ch == CR) { + parser->state = s_headers_almost_done; + break; + } + + if (ch == LF) { + /* they might be just sending \n instead of \r\n so this would be + * the second \n to denote the end of headers*/ + parser->state = s_headers_almost_done; + goto reexecute_byte; + } + + c = TOKEN(ch); + + if (!c) { + SET_ERRNO(HPE_INVALID_HEADER_TOKEN); + goto error; + } + + MARK(header_field); + + parser->index = 0; + parser->state = s_header_field; + + switch (c) { + case 'c': + parser->header_state = h_C; + break; + + case 'p': + parser->header_state = h_matching_proxy_connection; + break; + + case 't': + parser->header_state = h_matching_transfer_encoding; + break; + + case 'u': + parser->header_state = h_matching_upgrade; + break; + + default: + parser->header_state = h_general; + break; + } + break; + } + + case s_header_field: + { + c = TOKEN(ch); + + if (c) { + switch (parser->header_state) { + case h_general: + break; + + case h_C: + parser->index++; + parser->header_state = (c == 'o' ? h_CO : h_general); + break; + + case h_CO: + parser->index++; + parser->header_state = (c == 'n' ? h_CON : h_general); + break; + + case h_CON: + parser->index++; + switch (c) { + case 'n': + parser->header_state = h_matching_connection; + break; + case 't': + parser->header_state = h_matching_content_length; + break; + default: + parser->header_state = h_general; + break; + } + break; + + /* connection */ + + case h_matching_connection: + parser->index++; + if (parser->index > sizeof(CONNECTION)-1 + || c != CONNECTION[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(CONNECTION)-2) { + parser->header_state = h_connection; + } + break; + + /* proxy-connection */ + + case h_matching_proxy_connection: + parser->index++; + if (parser->index > sizeof(PROXY_CONNECTION)-1 + || c != PROXY_CONNECTION[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(PROXY_CONNECTION)-2) { + parser->header_state = h_connection; + } + break; + + /* content-length */ + + case h_matching_content_length: + parser->index++; + if (parser->index > sizeof(CONTENT_LENGTH)-1 + || c != CONTENT_LENGTH[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(CONTENT_LENGTH)-2) { + parser->header_state = h_content_length; + } + break; + + /* transfer-encoding */ + + case h_matching_transfer_encoding: + parser->index++; + if (parser->index > sizeof(TRANSFER_ENCODING)-1 + || c != TRANSFER_ENCODING[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(TRANSFER_ENCODING)-2) { + parser->header_state = h_transfer_encoding; + } + break; + + /* upgrade */ + + case h_matching_upgrade: + parser->index++; + if (parser->index > sizeof(UPGRADE)-1 + || c != UPGRADE[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(UPGRADE)-2) { + parser->header_state = h_upgrade; + } + break; + + case h_connection: + case h_content_length: + case h_transfer_encoding: + case h_upgrade: + if (ch != ' ') parser->header_state = h_general; + break; + + default: + assert(0 && "Unknown header_state"); + break; + } + break; + } + + if (ch == ':') { + parser->state = s_header_value_start; + CALLBACK_DATA(header_field); + break; + } + + if (ch == CR) { + parser->state = s_header_almost_done; + CALLBACK_DATA(header_field); + break; + } + + if (ch == LF) { + parser->state = s_header_field_start; + CALLBACK_DATA(header_field); + break; + } + + SET_ERRNO(HPE_INVALID_HEADER_TOKEN); + goto error; + } + + case s_header_value_start: + { + if (ch == ' ' || ch == '\t') break; + + MARK(header_value); + + parser->state = s_header_value; + parser->index = 0; + + if (ch == CR) { + parser->header_state = h_general; + parser->state = s_header_almost_done; + CALLBACK_DATA(header_value); + break; + } + + if (ch == LF) { + parser->state = s_header_field_start; + CALLBACK_DATA(header_value); + break; + } + + c = LOWER(ch); + + switch (parser->header_state) { + case h_upgrade: + parser->flags |= F_UPGRADE; + parser->header_state = h_general; + break; + + case h_transfer_encoding: + /* looking for 'Transfer-Encoding: chunked' */ + if ('c' == c) { + parser->header_state = h_matching_transfer_encoding_chunked; + } else { + parser->header_state = h_general; + } + break; + + case h_content_length: + if (!IS_NUM(ch)) { + SET_ERRNO(HPE_INVALID_CONTENT_LENGTH); + goto error; + } + + parser->content_length = ch - '0'; + break; + + case h_connection: + /* looking for 'Connection: keep-alive' */ + if (c == 'k') { + parser->header_state = h_matching_connection_keep_alive; + /* looking for 'Connection: close' */ + } else if (c == 'c') { + parser->header_state = h_matching_connection_close; + } else { + parser->header_state = h_general; + } + break; + + default: + parser->header_state = h_general; + break; + } + break; + } + + case s_header_value: + { + + if (ch == CR) { + parser->state = s_header_almost_done; + CALLBACK_DATA(header_value); + break; + } + + if (ch == LF) { + parser->state = s_header_almost_done; + CALLBACK_DATA_NOADVANCE(header_value); + goto reexecute_byte; + } + + c = LOWER(ch); + + switch (parser->header_state) { + case h_general: + break; + + case h_connection: + case h_transfer_encoding: + assert(0 && "Shouldn't get here."); + break; + + case h_content_length: + { + uint64_t t; + + if (ch == ' ') break; + + if (!IS_NUM(ch)) { + SET_ERRNO(HPE_INVALID_CONTENT_LENGTH); + goto error; + } + + t = parser->content_length; + t *= 10; + t += ch - '0'; + + /* Overflow? */ + if (t < parser->content_length || t == ULLONG_MAX) { + SET_ERRNO(HPE_INVALID_CONTENT_LENGTH); + goto error; + } + + parser->content_length = t; + break; + } + + /* Transfer-Encoding: chunked */ + case h_matching_transfer_encoding_chunked: + parser->index++; + if (parser->index > sizeof(CHUNKED)-1 + || c != CHUNKED[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(CHUNKED)-2) { + parser->header_state = h_transfer_encoding_chunked; + } + break; + + /* looking for 'Connection: keep-alive' */ + case h_matching_connection_keep_alive: + parser->index++; + if (parser->index > sizeof(KEEP_ALIVE)-1 + || c != KEEP_ALIVE[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(KEEP_ALIVE)-2) { + parser->header_state = h_connection_keep_alive; + } + break; + + /* looking for 'Connection: close' */ + case h_matching_connection_close: + parser->index++; + if (parser->index > sizeof(CLOSE)-1 || c != CLOSE[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(CLOSE)-2) { + parser->header_state = h_connection_close; + } + break; + + case h_transfer_encoding_chunked: + case h_connection_keep_alive: + case h_connection_close: + if (ch != ' ') parser->header_state = h_general; + break; + + default: + parser->state = s_header_value; + parser->header_state = h_general; + break; + } + break; + } + + case s_header_almost_done: + { + STRICT_CHECK(ch != LF); + + parser->state = s_header_value_lws; + + switch (parser->header_state) { + case h_connection_keep_alive: + parser->flags |= F_CONNECTION_KEEP_ALIVE; + break; + case h_connection_close: + parser->flags |= F_CONNECTION_CLOSE; + break; + case h_transfer_encoding_chunked: + parser->flags |= F_CHUNKED; + break; + default: + break; + } + + break; + } + + case s_header_value_lws: + { + if (ch == ' ' || ch == '\t') + parser->state = s_header_value_start; + else + { + parser->state = s_header_field_start; + goto reexecute_byte; + } + break; + } + + case s_headers_almost_done: + { + STRICT_CHECK(ch != LF); + + if (parser->flags & F_TRAILING) { + /* End of a chunked request */ + parser->state = NEW_MESSAGE(); + CALLBACK_NOTIFY(message_complete); + break; + } + + parser->state = s_headers_done; + + /* Set this here so that on_headers_complete() callbacks can see it */ + parser->upgrade = + (parser->flags & F_UPGRADE || parser->method == HTTP_CONNECT); + + /* Here we call the headers_complete callback. This is somewhat + * different than other callbacks because if the user returns 1, we + * will interpret that as saying that this message has no body. This + * is needed for the annoying case of recieving a response to a HEAD + * request. + * + * We'd like to use CALLBACK_NOTIFY_NOADVANCE() here but we cannot, so + * we have to simulate it by handling a change in errno below. + */ + if (settings->on_headers_complete) { + switch (settings->on_headers_complete(parser)) { + case 0: + break; + + case 1: + parser->flags |= F_SKIPBODY; + break; + + default: + SET_ERRNO(HPE_CB_headers_complete); + return p - data; /* Error */ + } + } + + if (HTTP_PARSER_ERRNO(parser) != HPE_OK) { + return p - data; + } + + goto reexecute_byte; + } + + case s_headers_done: + { + STRICT_CHECK(ch != LF); + + parser->nread = 0; + + /* Exit, the rest of the connect is in a different protocol. */ + if (parser->upgrade) { + parser->state = NEW_MESSAGE(); + CALLBACK_NOTIFY(message_complete); + return (p - data) + 1; + } + + if (parser->flags & F_SKIPBODY) { + parser->state = NEW_MESSAGE(); + CALLBACK_NOTIFY(message_complete); + } else if (parser->flags & F_CHUNKED) { + /* chunked encoding - ignore Content-Length header */ + parser->state = s_chunk_size_start; + } else { + if (parser->content_length == 0) { + /* Content-Length header given but zero: Content-Length: 0\r\n */ + parser->state = NEW_MESSAGE(); + CALLBACK_NOTIFY(message_complete); + } else if (parser->content_length != ULLONG_MAX) { + /* Content-Length header given and non-zero */ + parser->state = s_body_identity; + } else { + if (parser->type == HTTP_REQUEST || + !http_message_needs_eof(parser)) { + /* Assume content-length 0 - read the next */ + parser->state = NEW_MESSAGE(); + CALLBACK_NOTIFY(message_complete); + } else { + /* Read body until EOF */ + parser->state = s_body_identity_eof; + } + } + } + + break; + } + + case s_body_identity: + { + uint64_t to_read = MIN(parser->content_length, + (uint64_t) ((data + len) - p)); + + assert(parser->content_length != 0 + && parser->content_length != ULLONG_MAX); + + /* The difference between advancing content_length and p is because + * the latter will automaticaly advance on the next loop iteration. + * Further, if content_length ends up at 0, we want to see the last + * byte again for our message complete callback. + */ + MARK(body); + parser->content_length -= to_read; + p += to_read - 1; + + if (parser->content_length == 0) { + parser->state = s_message_done; + + /* Mimic CALLBACK_DATA_NOADVANCE() but with one extra byte. + * + * The alternative to doing this is to wait for the next byte to + * trigger the data callback, just as in every other case. The + * problem with this is that this makes it difficult for the test + * harness to distinguish between complete-on-EOF and + * complete-on-length. It's not clear that this distinction is + * important for applications, but let's keep it for now. + */ + CALLBACK_DATA_(body, p - body_mark + 1, p - data); + goto reexecute_byte; + } + + break; + } + + /* read until EOF */ + case s_body_identity_eof: + MARK(body); + p = data + len - 1; + + break; + + case s_message_done: + parser->state = NEW_MESSAGE(); + CALLBACK_NOTIFY(message_complete); + break; + + case s_chunk_size_start: + { + assert(parser->nread == 1); + assert(parser->flags & F_CHUNKED); + + unhex_val = unhex[(unsigned char)ch]; + if (unhex_val == -1) { + SET_ERRNO(HPE_INVALID_CHUNK_SIZE); + goto error; + } + + parser->content_length = unhex_val; + parser->state = s_chunk_size; + break; + } + + case s_chunk_size: + { + uint64_t t; + + assert(parser->flags & F_CHUNKED); + + if (ch == CR) { + parser->state = s_chunk_size_almost_done; + break; + } + + unhex_val = unhex[(unsigned char)ch]; + + if (unhex_val == -1) { + if (ch == ';' || ch == ' ') { + parser->state = s_chunk_parameters; + break; + } + + SET_ERRNO(HPE_INVALID_CHUNK_SIZE); + goto error; + } + + t = parser->content_length; + t *= 16; + t += unhex_val; + + /* Overflow? */ + if (t < parser->content_length || t == ULLONG_MAX) { + SET_ERRNO(HPE_INVALID_CONTENT_LENGTH); + goto error; + } + + parser->content_length = t; + break; + } + + case s_chunk_parameters: + { + assert(parser->flags & F_CHUNKED); + /* just ignore this shit. TODO check for overflow */ + if (ch == CR) { + parser->state = s_chunk_size_almost_done; + break; + } + break; + } + + case s_chunk_size_almost_done: + { + assert(parser->flags & F_CHUNKED); + STRICT_CHECK(ch != LF); + + parser->nread = 0; + + if (parser->content_length == 0) { + parser->flags |= F_TRAILING; + parser->state = s_header_field_start; + } else { + parser->state = s_chunk_data; + } + break; + } + + case s_chunk_data: + { + uint64_t to_read = MIN(parser->content_length, + (uint64_t) ((data + len) - p)); + + assert(parser->flags & F_CHUNKED); + assert(parser->content_length != 0 + && parser->content_length != ULLONG_MAX); + + /* See the explanation in s_body_identity for why the content + * length and data pointers are managed this way. + */ + MARK(body); + parser->content_length -= to_read; + p += to_read - 1; + + if (parser->content_length == 0) { + parser->state = s_chunk_data_almost_done; + } + + break; + } + + case s_chunk_data_almost_done: + assert(parser->flags & F_CHUNKED); + assert(parser->content_length == 0); + STRICT_CHECK(ch != CR); + parser->state = s_chunk_data_done; + CALLBACK_DATA(body); + break; + + case s_chunk_data_done: + assert(parser->flags & F_CHUNKED); + STRICT_CHECK(ch != LF); + parser->nread = 0; + parser->state = s_chunk_size_start; + break; + + default: + assert(0 && "unhandled state"); + SET_ERRNO(HPE_INVALID_INTERNAL_STATE); + goto error; + } + } + + /* Run callbacks for any marks that we have leftover after we ran our of + * bytes. There should be at most one of these set, so it's OK to invoke + * them in series (unset marks will not result in callbacks). + * + * We use the NOADVANCE() variety of callbacks here because 'p' has already + * overflowed 'data' and this allows us to correct for the off-by-one that + * we'd otherwise have (since CALLBACK_DATA() is meant to be run with a 'p' + * value that's in-bounds). + */ + + assert(((header_field_mark ? 1 : 0) + + (header_value_mark ? 1 : 0) + + (url_mark ? 1 : 0) + + (body_mark ? 1 : 0)) <= 1); + + CALLBACK_DATA_NOADVANCE(header_field); + CALLBACK_DATA_NOADVANCE(header_value); + CALLBACK_DATA_NOADVANCE(url); + CALLBACK_DATA_NOADVANCE(body); + + return len; + +error: + if (HTTP_PARSER_ERRNO(parser) == HPE_OK) { + SET_ERRNO(HPE_UNKNOWN); + } + + return (p - data); +} + + +/* Does the parser need to see an EOF to find the end of the message? */ +int +http_message_needs_eof (http_parser *parser) +{ + if (parser->type == HTTP_REQUEST) { + return 0; + } + + /* See RFC 2616 section 4.4 */ + if (parser->status_code / 100 == 1 || /* 1xx e.g. Continue */ + parser->status_code == 204 || /* No Content */ + parser->status_code == 304 || /* Not Modified */ + parser->flags & F_SKIPBODY) { /* response to a HEAD request */ + return 0; + } + + if ((parser->flags & F_CHUNKED) || parser->content_length != ULLONG_MAX) { + return 0; + } + + return 1; +} + + +int +http_should_keep_alive (http_parser *parser) +{ + if (parser->http_major > 0 && parser->http_minor > 0) { + /* HTTP/1.1 */ + if (parser->flags & F_CONNECTION_CLOSE) { + return 0; + } + } else { + /* HTTP/1.0 or earlier */ + if (!(parser->flags & F_CONNECTION_KEEP_ALIVE)) { + return 0; + } + } + + return !http_message_needs_eof(parser); +} + + +const char * http_method_str (enum http_method m) +{ + return method_strings[m]; +} + + +void +http_parser_init (http_parser *parser, enum http_parser_type t) +{ + void *data = parser->data; /* preserve application data */ + memset(parser, 0, sizeof(*parser)); + parser->data = data; + parser->type = t; + parser->state = (t == HTTP_REQUEST ? s_start_req : (t == HTTP_RESPONSE ? s_start_res : s_start_req_or_res)); + parser->http_errno = HPE_OK; +} + +const char * +http_errno_name(enum http_errno err) { + assert(err < (sizeof(http_strerror_tab)/sizeof(http_strerror_tab[0]))); + return http_strerror_tab[err].name; +} + +const char * +http_errno_description(enum http_errno err) { + assert(err < (sizeof(http_strerror_tab)/sizeof(http_strerror_tab[0]))); + return http_strerror_tab[err].description; +} + +int +http_parser_parse_url(const char *buf, size_t buflen, int is_connect, + struct http_parser_url *u) +{ + enum state s; + const char *p; + enum http_parser_url_fields uf, old_uf; + + u->port = u->field_set = 0; + s = is_connect ? s_req_host_start : s_req_spaces_before_url; + uf = old_uf = UF_MAX; + + for (p = buf; p < buf + buflen; p++) { + s = parse_url_char(s, *p); + + /* Figure out the next field that we're operating on */ + switch (s) { + case s_dead: + return 1; + + /* Skip delimeters */ + case s_req_schema_slash: + case s_req_schema_slash_slash: + case s_req_host_start: + case s_req_host_v6_start: + case s_req_host_v6_end: + case s_req_port_start: + case s_req_query_string_start: + case s_req_fragment_start: + continue; + + case s_req_schema: + uf = UF_SCHEMA; + break; + + case s_req_host: + case s_req_host_v6: + uf = UF_HOST; + break; + + case s_req_port: + uf = UF_PORT; + break; + + case s_req_path: + uf = UF_PATH; + break; + + case s_req_query_string: + uf = UF_QUERY; + break; + + case s_req_fragment: + uf = UF_FRAGMENT; + break; + + default: + assert(!"Unexpected state"); + return 1; + } + + /* Nothing's changed; soldier on */ + if (uf == old_uf) { + u->field_data[uf].len++; + continue; + } + + u->field_data[uf].off = p - buf; + u->field_data[uf].len = 1; + + u->field_set |= (1 << uf); + old_uf = uf; + } + + /* CONNECT requests can only contain "hostname:port" */ + if (is_connect && u->field_set != ((1 << UF_HOST)|(1 << UF_PORT))) { + return 1; + } + + /* Make sure we don't end somewhere unexpected */ + switch (s) { + case s_req_host_v6_start: + case s_req_host_v6: + case s_req_host_v6_end: + case s_req_host: + case s_req_port_start: + return 1; + default: + break; + } + + if (u->field_set & (1 << UF_PORT)) { + /* Don't bother with endp; we've already validated the string */ + unsigned long v = strtoul(buf + u->field_data[UF_PORT].off, NULL, 10); + + /* Ports have a max value of 2^16 */ + if (v > 0xffff) { + return 1; + } + + u->port = (uint16_t) v; + } + + return 0; +} + +void +http_parser_pause(http_parser *parser, int paused) { + /* Users should only be pausing/unpausing a parser that is not in an error + * state. In non-debug builds, there's not much that we can do about this + * other than ignore it. + */ + if (HTTP_PARSER_ERRNO(parser) == HPE_OK || + HTTP_PARSER_ERRNO(parser) == HPE_PAUSED) { + SET_ERRNO((paused) ? HPE_PAUSED : HPE_OK); + } else { + assert(0 && "Attempting to pause parser in error state"); + } +} diff --git a/examples/http-parser/http_parser.gyp b/examples/http-parser/http_parser.gyp new file mode 100644 index 0000000..c6eada7 --- /dev/null +++ b/examples/http-parser/http_parser.gyp @@ -0,0 +1,79 @@ +# This file is used with the GYP meta build system. +# http://code.google.com/p/gyp/ +# To build try this: +# svn co http://gyp.googlecode.com/svn/trunk gyp +# ./gyp/gyp -f make --depth=`pwd` http_parser.gyp +# ./out/Debug/test +{ + 'target_defaults': { + 'default_configuration': 'Debug', + 'configurations': { + # TODO: hoist these out and put them somewhere common, because + # RuntimeLibrary MUST MATCH across the entire project + 'Debug': { + 'defines': [ 'DEBUG', '_DEBUG' ], + 'msvs_settings': { + 'VCCLCompilerTool': { + 'RuntimeLibrary': 1, # static debug + }, + }, + }, + 'Release': { + 'defines': [ 'NDEBUG' ], + 'msvs_settings': { + 'VCCLCompilerTool': { + 'RuntimeLibrary': 0, # static release + }, + }, + } + }, + 'msvs_settings': { + 'VCCLCompilerTool': { + }, + 'VCLibrarianTool': { + }, + 'VCLinkerTool': { + 'GenerateDebugInformation': 'true', + }, + }, + 'conditions': [ + ['OS == "win"', { + 'defines': [ + 'WIN32' + ], + }] + ], + }, + + 'targets': [ + { + 'target_name': 'http_parser', + 'type': 'static_library', + 'include_dirs': [ '.' ], + 'direct_dependent_settings': { + 'include_dirs': [ '.' ], + }, + 'defines': [ 'HTTP_PARSER_STRICT=0' ], + 'sources': [ './http_parser.c', ], + 'conditions': [ + ['OS=="win"', { + 'msvs_settings': { + 'VCCLCompilerTool': { + # Compile as C++. http_parser.c is actually C99, but C++ is + # close enough in this case. + 'CompileAs': 2, + }, + }, + }] + ], + }, + + { + 'target_name': 'test', + 'type': 'executable', + 'dependencies': [ 'http_parser' ], + 'sources': [ 'test.c' ] + } + ] +} + diff --git a/examples/http-parser/http_parser.h b/examples/http-parser/http_parser.h new file mode 100644 index 0000000..8ed4180 --- /dev/null +++ b/examples/http-parser/http_parser.h @@ -0,0 +1,318 @@ +/* Copyright Joyent, Inc. and other Node contributors. All rights reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + */ +#ifndef http_parser_h +#define http_parser_h +#ifdef __cplusplus +extern "C" { +#endif + +#define HTTP_PARSER_VERSION_MAJOR 1 +#define HTTP_PARSER_VERSION_MINOR 0 + +#include +#if defined(_WIN32) && !defined(__MINGW32__) && (!defined(_MSC_VER) || _MSC_VER<1600) +typedef __int8 int8_t; +typedef unsigned __int8 uint8_t; +typedef __int16 int16_t; +typedef unsigned __int16 uint16_t; +typedef __int32 int32_t; +typedef unsigned __int32 uint32_t; +typedef __int64 int64_t; +typedef unsigned __int64 uint64_t; + +typedef unsigned int size_t; +typedef int ssize_t; +#else +#include +#endif + +/* Compile with -DHTTP_PARSER_STRICT=0 to make less checks, but run + * faster + */ +#ifndef HTTP_PARSER_STRICT +# define HTTP_PARSER_STRICT 1 +#endif + +/* Compile with -DHTTP_PARSER_DEBUG=1 to add extra debugging information to + * the error reporting facility. + */ +#ifndef HTTP_PARSER_DEBUG +# define HTTP_PARSER_DEBUG 0 +#endif + + +/* Maximium header size allowed */ +#define HTTP_MAX_HEADER_SIZE (80*1024) + + +typedef struct http_parser http_parser; +typedef struct http_parser_settings http_parser_settings; + + +/* Callbacks should return non-zero to indicate an error. The parser will + * then halt execution. + * + * The one exception is on_headers_complete. In a HTTP_RESPONSE parser + * returning '1' from on_headers_complete will tell the parser that it + * should not expect a body. This is used when receiving a response to a + * HEAD request which may contain 'Content-Length' or 'Transfer-Encoding: + * chunked' headers that indicate the presence of a body. + * + * http_data_cb does not return data chunks. It will be call arbitrarally + * many times for each string. E.G. you might get 10 callbacks for "on_path" + * each providing just a few characters more data. + */ +typedef int (*http_data_cb) (http_parser*, const char *at, size_t length); +typedef int (*http_cb) (http_parser*); + + +/* Request Methods */ +#define HTTP_METHOD_MAP(XX) \ + XX(0, DELETE, DELETE) \ + XX(1, GET, GET) \ + XX(2, HEAD, HEAD) \ + XX(3, POST, POST) \ + XX(4, PUT, PUT) \ + /* pathological */ \ + XX(5, CONNECT, CONNECT) \ + XX(6, OPTIONS, OPTIONS) \ + XX(7, TRACE, TRACE) \ + /* webdav */ \ + XX(8, COPY, COPY) \ + XX(9, LOCK, LOCK) \ + XX(10, MKCOL, MKCOL) \ + XX(11, MOVE, MOVE) \ + XX(12, PROPFIND, PROPFIND) \ + XX(13, PROPPATCH, PROPPATCH) \ + XX(14, SEARCH, SEARCH) \ + XX(15, UNLOCK, UNLOCK) \ + /* subversion */ \ + XX(16, REPORT, REPORT) \ + XX(17, MKACTIVITY, MKACTIVITY) \ + XX(18, CHECKOUT, CHECKOUT) \ + XX(19, MERGE, MERGE) \ + /* upnp */ \ + XX(20, MSEARCH, M-SEARCH) \ + XX(21, NOTIFY, NOTIFY) \ + XX(22, SUBSCRIBE, SUBSCRIBE) \ + XX(23, UNSUBSCRIBE, UNSUBSCRIBE) \ + /* RFC-5789 */ \ + XX(24, PATCH, PATCH) \ + XX(25, PURGE, PURGE) \ + +enum http_method + { +#define XX(num, name, string) HTTP_##name = num, + HTTP_METHOD_MAP(XX) +#undef XX + }; + + +enum http_parser_type { HTTP_REQUEST, HTTP_RESPONSE, HTTP_BOTH }; + + +/* Flag values for http_parser.flags field */ +enum flags + { F_CHUNKED = 1 << 0 + , F_CONNECTION_KEEP_ALIVE = 1 << 1 + , F_CONNECTION_CLOSE = 1 << 2 + , F_TRAILING = 1 << 3 + , F_UPGRADE = 1 << 4 + , F_SKIPBODY = 1 << 5 + }; + + +/* Map for errno-related constants + * + * The provided argument should be a macro that takes 2 arguments. + */ +#define HTTP_ERRNO_MAP(XX) \ + /* No error */ \ + XX(OK, "success") \ + \ + /* Callback-related errors */ \ + XX(CB_message_begin, "the on_message_begin callback failed") \ + XX(CB_url, "the on_url callback failed") \ + XX(CB_header_field, "the on_header_field callback failed") \ + XX(CB_header_value, "the on_header_value callback failed") \ + XX(CB_headers_complete, "the on_headers_complete callback failed") \ + XX(CB_body, "the on_body callback failed") \ + XX(CB_message_complete, "the on_message_complete callback failed") \ + \ + /* Parsing-related errors */ \ + XX(INVALID_EOF_STATE, "stream ended at an unexpected time") \ + XX(HEADER_OVERFLOW, \ + "too many header bytes seen; overflow detected") \ + XX(CLOSED_CONNECTION, \ + "data received after completed connection: close message") \ + XX(INVALID_VERSION, "invalid HTTP version") \ + XX(INVALID_STATUS, "invalid HTTP status code") \ + XX(INVALID_METHOD, "invalid HTTP method") \ + XX(INVALID_URL, "invalid URL") \ + XX(INVALID_HOST, "invalid host") \ + XX(INVALID_PORT, "invalid port") \ + XX(INVALID_PATH, "invalid path") \ + XX(INVALID_QUERY_STRING, "invalid query string") \ + XX(INVALID_FRAGMENT, "invalid fragment") \ + XX(LF_EXPECTED, "LF character expected") \ + XX(INVALID_HEADER_TOKEN, "invalid character in header") \ + XX(INVALID_CONTENT_LENGTH, \ + "invalid character in content-length header") \ + XX(INVALID_CHUNK_SIZE, \ + "invalid character in chunk size header") \ + XX(INVALID_CONSTANT, "invalid constant string") \ + XX(INVALID_INTERNAL_STATE, "encountered unexpected internal state")\ + XX(STRICT, "strict mode assertion failed") \ + XX(PAUSED, "parser is paused") \ + XX(UNKNOWN, "an unknown error occurred") + + +/* Define HPE_* values for each errno value above */ +#define HTTP_ERRNO_GEN(n, s) HPE_##n, +enum http_errno { + HTTP_ERRNO_MAP(HTTP_ERRNO_GEN) +}; +#undef HTTP_ERRNO_GEN + + +/* Get an http_errno value from an http_parser */ +#define HTTP_PARSER_ERRNO(p) ((enum http_errno) (p)->http_errno) + +/* Get the line number that generated the current error */ +#if HTTP_PARSER_DEBUG +#define HTTP_PARSER_ERRNO_LINE(p) ((p)->error_lineno) +#else +#define HTTP_PARSER_ERRNO_LINE(p) 0 +#endif + + +struct http_parser { + /** PRIVATE **/ + unsigned char type : 2; /* enum http_parser_type */ + unsigned char flags : 6; /* F_* values from 'flags' enum; semi-public */ + unsigned char state; /* enum state from http_parser.c */ + unsigned char header_state; /* enum header_state from http_parser.c */ + unsigned char index; /* index into current matcher */ + + uint32_t nread; /* # bytes read in various scenarios */ + uint64_t content_length; /* # bytes in body (0 if no Content-Length header) */ + + /** READ-ONLY **/ + unsigned short http_major; + unsigned short http_minor; + unsigned short status_code; /* responses only */ + unsigned char method; /* requests only */ + unsigned char http_errno : 7; + + /* 1 = Upgrade header was present and the parser has exited because of that. + * 0 = No upgrade header present. + * Should be checked when http_parser_execute() returns in addition to + * error checking. + */ + unsigned char upgrade : 1; + +#if HTTP_PARSER_DEBUG + uint32_t error_lineno; +#endif + + /** PUBLIC **/ + void *data; /* A pointer to get hook to the "connection" or "socket" object */ +}; + + +struct http_parser_settings { + http_cb on_message_begin; + http_data_cb on_url; + http_data_cb on_header_field; + http_data_cb on_header_value; + http_cb on_headers_complete; + http_data_cb on_body; + http_cb on_message_complete; +}; + + +enum http_parser_url_fields + { UF_SCHEMA = 0 + , UF_HOST = 1 + , UF_PORT = 2 + , UF_PATH = 3 + , UF_QUERY = 4 + , UF_FRAGMENT = 5 + , UF_MAX = 6 + }; + + +/* Result structure for http_parser_parse_url(). + * + * Callers should index into field_data[] with UF_* values iff field_set + * has the relevant (1 << UF_*) bit set. As a courtesy to clients (and + * because we probably have padding left over), we convert any port to + * a uint16_t. + */ +struct http_parser_url { + uint16_t field_set; /* Bitmask of (1 << UF_*) values */ + uint16_t port; /* Converted UF_PORT string */ + + struct { + uint16_t off; /* Offset into buffer in which field starts */ + uint16_t len; /* Length of run in buffer */ + } field_data[UF_MAX]; +}; + + +void http_parser_init(http_parser *parser, enum http_parser_type type); + + +size_t http_parser_execute(http_parser *parser, + const http_parser_settings *settings, + const char *data, + size_t len); + + +/* If http_should_keep_alive() in the on_headers_complete or + * on_message_complete callback returns true, then this will be should be + * the last message on the connection. + * If you are the server, respond with the "Connection: close" header. + * If you are the client, close the connection. + */ +int http_should_keep_alive(http_parser *parser); + +/* Returns a string version of the HTTP method. */ +const char *http_method_str(enum http_method m); + +/* Return a string name of the given error */ +const char *http_errno_name(enum http_errno err); + +/* Return a string description of the given error */ +const char *http_errno_description(enum http_errno err); + +/* Parse a URL; return nonzero on failure */ +int http_parser_parse_url(const char *buf, size_t buflen, + int is_connect, + struct http_parser_url *u); + +/* Pause or un-pause the parser; a nonzero value pauses */ +void http_parser_pause(http_parser *parser, int paused); + +#ifdef __cplusplus +} +#endif +#endif diff --git a/examples/http-parser/test.c b/examples/http-parser/test.c new file mode 100644 index 0000000..6d8c004 --- /dev/null +++ b/examples/http-parser/test.c @@ -0,0 +1,2958 @@ +/* Copyright Joyent, Inc. and other Node contributors. All rights reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + */ +#include "http_parser.h" +#include +#include +#include +#include /* rand */ +#include +#include + +#undef TRUE +#define TRUE 1 +#undef FALSE +#define FALSE 0 + +#define MAX_HEADERS 13 +#define MAX_ELEMENT_SIZE 500 + +#define MIN(a,b) ((a) < (b) ? (a) : (b)) + +static http_parser *parser; + +struct message { + const char *name; // for debugging purposes + const char *raw; + enum http_parser_type type; + enum http_method method; + int status_code; + char request_path[MAX_ELEMENT_SIZE]; + char request_url[MAX_ELEMENT_SIZE]; + char fragment[MAX_ELEMENT_SIZE]; + char query_string[MAX_ELEMENT_SIZE]; + char body[MAX_ELEMENT_SIZE]; + size_t body_size; + uint16_t port; + int num_headers; + enum { NONE=0, FIELD, VALUE } last_header_element; + char headers [MAX_HEADERS][2][MAX_ELEMENT_SIZE]; + int should_keep_alive; + + const char *upgrade; // upgraded body + + unsigned short http_major; + unsigned short http_minor; + + int message_begin_cb_called; + int headers_complete_cb_called; + int message_complete_cb_called; + int message_complete_on_eof; +}; + +static int currently_parsing_eof; + +static struct message messages[5]; +static int num_messages; +static http_parser_settings *current_pause_parser; + +/* * R E Q U E S T S * */ +const struct message requests[] = +#define CURL_GET 0 +{ {.name= "curl get" + ,.type= HTTP_REQUEST + ,.raw= "GET /test HTTP/1.1\r\n" + "User-Agent: curl/7.18.0 (i486-pc-linux-gnu) libcurl/7.18.0 OpenSSL/0.9.8g zlib/1.2.3.3 libidn/1.1\r\n" + "Host: 0.0.0.0=5000\r\n" + "Accept: */*\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/test" + ,.request_url= "/test" + ,.num_headers= 3 + ,.headers= + { { "User-Agent", "curl/7.18.0 (i486-pc-linux-gnu) libcurl/7.18.0 OpenSSL/0.9.8g zlib/1.2.3.3 libidn/1.1" } + , { "Host", "0.0.0.0=5000" } + , { "Accept", "*/*" } + } + ,.body= "" + } + +#define FIREFOX_GET 1 +, {.name= "firefox get" + ,.type= HTTP_REQUEST + ,.raw= "GET /favicon.ico HTTP/1.1\r\n" + "Host: 0.0.0.0=5000\r\n" + "User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9) Gecko/2008061015 Firefox/3.0\r\n" + "Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8\r\n" + "Accept-Language: en-us,en;q=0.5\r\n" + "Accept-Encoding: gzip,deflate\r\n" + "Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7\r\n" + "Keep-Alive: 300\r\n" + "Connection: keep-alive\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/favicon.ico" + ,.request_url= "/favicon.ico" + ,.num_headers= 8 + ,.headers= + { { "Host", "0.0.0.0=5000" } + , { "User-Agent", "Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9) Gecko/2008061015 Firefox/3.0" } + , { "Accept", "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8" } + , { "Accept-Language", "en-us,en;q=0.5" } + , { "Accept-Encoding", "gzip,deflate" } + , { "Accept-Charset", "ISO-8859-1,utf-8;q=0.7,*;q=0.7" } + , { "Keep-Alive", "300" } + , { "Connection", "keep-alive" } + } + ,.body= "" + } + +#define DUMBFUCK 2 +, {.name= "dumbfuck" + ,.type= HTTP_REQUEST + ,.raw= "GET /dumbfuck HTTP/1.1\r\n" + "aaaaaaaaaaaaa:++++++++++\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/dumbfuck" + ,.request_url= "/dumbfuck" + ,.num_headers= 1 + ,.headers= + { { "aaaaaaaaaaaaa", "++++++++++" } + } + ,.body= "" + } + +#define FRAGMENT_IN_URI 3 +, {.name= "fragment in url" + ,.type= HTTP_REQUEST + ,.raw= "GET /forums/1/topics/2375?page=1#posts-17408 HTTP/1.1\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "page=1" + ,.fragment= "posts-17408" + ,.request_path= "/forums/1/topics/2375" + /* XXX request url does include fragment? */ + ,.request_url= "/forums/1/topics/2375?page=1#posts-17408" + ,.num_headers= 0 + ,.body= "" + } + +#define GET_NO_HEADERS_NO_BODY 4 +, {.name= "get no headers no body" + ,.type= HTTP_REQUEST + ,.raw= "GET /get_no_headers_no_body/world HTTP/1.1\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE /* would need Connection: close */ + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/get_no_headers_no_body/world" + ,.request_url= "/get_no_headers_no_body/world" + ,.num_headers= 0 + ,.body= "" + } + +#define GET_ONE_HEADER_NO_BODY 5 +, {.name= "get one header no body" + ,.type= HTTP_REQUEST + ,.raw= "GET /get_one_header_no_body HTTP/1.1\r\n" + "Accept: */*\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE /* would need Connection: close */ + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/get_one_header_no_body" + ,.request_url= "/get_one_header_no_body" + ,.num_headers= 1 + ,.headers= + { { "Accept" , "*/*" } + } + ,.body= "" + } + +#define GET_FUNKY_CONTENT_LENGTH 6 +, {.name= "get funky content length body hello" + ,.type= HTTP_REQUEST + ,.raw= "GET /get_funky_content_length_body_hello HTTP/1.0\r\n" + "conTENT-Length: 5\r\n" + "\r\n" + "HELLO" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 0 + ,.method= HTTP_GET + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/get_funky_content_length_body_hello" + ,.request_url= "/get_funky_content_length_body_hello" + ,.num_headers= 1 + ,.headers= + { { "conTENT-Length" , "5" } + } + ,.body= "HELLO" + } + +#define POST_IDENTITY_BODY_WORLD 7 +, {.name= "post identity body world" + ,.type= HTTP_REQUEST + ,.raw= "POST /post_identity_body_world?q=search#hey HTTP/1.1\r\n" + "Accept: */*\r\n" + "Transfer-Encoding: identity\r\n" + "Content-Length: 5\r\n" + "\r\n" + "World" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_POST + ,.query_string= "q=search" + ,.fragment= "hey" + ,.request_path= "/post_identity_body_world" + ,.request_url= "/post_identity_body_world?q=search#hey" + ,.num_headers= 3 + ,.headers= + { { "Accept", "*/*" } + , { "Transfer-Encoding", "identity" } + , { "Content-Length", "5" } + } + ,.body= "World" + } + +#define POST_CHUNKED_ALL_YOUR_BASE 8 +, {.name= "post - chunked body: all your base are belong to us" + ,.type= HTTP_REQUEST + ,.raw= "POST /post_chunked_all_your_base HTTP/1.1\r\n" + "Transfer-Encoding: chunked\r\n" + "\r\n" + "1e\r\nall your base are belong to us\r\n" + "0\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_POST + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/post_chunked_all_your_base" + ,.request_url= "/post_chunked_all_your_base" + ,.num_headers= 1 + ,.headers= + { { "Transfer-Encoding" , "chunked" } + } + ,.body= "all your base are belong to us" + } + +#define TWO_CHUNKS_MULT_ZERO_END 9 +, {.name= "two chunks ; triple zero ending" + ,.type= HTTP_REQUEST + ,.raw= "POST /two_chunks_mult_zero_end HTTP/1.1\r\n" + "Transfer-Encoding: chunked\r\n" + "\r\n" + "5\r\nhello\r\n" + "6\r\n world\r\n" + "000\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_POST + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/two_chunks_mult_zero_end" + ,.request_url= "/two_chunks_mult_zero_end" + ,.num_headers= 1 + ,.headers= + { { "Transfer-Encoding", "chunked" } + } + ,.body= "hello world" + } + +#define CHUNKED_W_TRAILING_HEADERS 10 +, {.name= "chunked with trailing headers. blech." + ,.type= HTTP_REQUEST + ,.raw= "POST /chunked_w_trailing_headers HTTP/1.1\r\n" + "Transfer-Encoding: chunked\r\n" + "\r\n" + "5\r\nhello\r\n" + "6\r\n world\r\n" + "0\r\n" + "Vary: *\r\n" + "Content-Type: text/plain\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_POST + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/chunked_w_trailing_headers" + ,.request_url= "/chunked_w_trailing_headers" + ,.num_headers= 3 + ,.headers= + { { "Transfer-Encoding", "chunked" } + , { "Vary", "*" } + , { "Content-Type", "text/plain" } + } + ,.body= "hello world" + } + +#define CHUNKED_W_BULLSHIT_AFTER_LENGTH 11 +, {.name= "with bullshit after the length" + ,.type= HTTP_REQUEST + ,.raw= "POST /chunked_w_bullshit_after_length HTTP/1.1\r\n" + "Transfer-Encoding: chunked\r\n" + "\r\n" + "5; ihatew3;whatthefuck=aretheseparametersfor\r\nhello\r\n" + "6; blahblah; blah\r\n world\r\n" + "0\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_POST + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/chunked_w_bullshit_after_length" + ,.request_url= "/chunked_w_bullshit_after_length" + ,.num_headers= 1 + ,.headers= + { { "Transfer-Encoding", "chunked" } + } + ,.body= "hello world" + } + +#define WITH_QUOTES 12 +, {.name= "with quotes" + ,.type= HTTP_REQUEST + ,.raw= "GET /with_\"stupid\"_quotes?foo=\"bar\" HTTP/1.1\r\n\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "foo=\"bar\"" + ,.fragment= "" + ,.request_path= "/with_\"stupid\"_quotes" + ,.request_url= "/with_\"stupid\"_quotes?foo=\"bar\"" + ,.num_headers= 0 + ,.headers= { } + ,.body= "" + } + +#define APACHEBENCH_GET 13 +/* The server receiving this request SHOULD NOT wait for EOF + * to know that content-length == 0. + * How to represent this in a unit test? message_complete_on_eof + * Compare with NO_CONTENT_LENGTH_RESPONSE. + */ +, {.name = "apachebench get" + ,.type= HTTP_REQUEST + ,.raw= "GET /test HTTP/1.0\r\n" + "Host: 0.0.0.0:5000\r\n" + "User-Agent: ApacheBench/2.3\r\n" + "Accept: */*\r\n\r\n" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 0 + ,.method= HTTP_GET + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/test" + ,.request_url= "/test" + ,.num_headers= 3 + ,.headers= { { "Host", "0.0.0.0:5000" } + , { "User-Agent", "ApacheBench/2.3" } + , { "Accept", "*/*" } + } + ,.body= "" + } + +#define QUERY_URL_WITH_QUESTION_MARK_GET 14 +/* Some clients include '?' characters in query strings. + */ +, {.name = "query url with question mark" + ,.type= HTTP_REQUEST + ,.raw= "GET /test.cgi?foo=bar?baz HTTP/1.1\r\n\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "foo=bar?baz" + ,.fragment= "" + ,.request_path= "/test.cgi" + ,.request_url= "/test.cgi?foo=bar?baz" + ,.num_headers= 0 + ,.headers= {} + ,.body= "" + } + +#define PREFIX_NEWLINE_GET 15 +/* Some clients, especially after a POST in a keep-alive connection, + * will send an extra CRLF before the next request + */ +, {.name = "newline prefix get" + ,.type= HTTP_REQUEST + ,.raw= "\r\nGET /test HTTP/1.1\r\n\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/test" + ,.request_url= "/test" + ,.num_headers= 0 + ,.headers= { } + ,.body= "" + } + +#define UPGRADE_REQUEST 16 +, {.name = "upgrade request" + ,.type= HTTP_REQUEST + ,.raw= "GET /demo HTTP/1.1\r\n" + "Host: example.com\r\n" + "Connection: Upgrade\r\n" + "Sec-WebSocket-Key2: 12998 5 Y3 1 .P00\r\n" + "Sec-WebSocket-Protocol: sample\r\n" + "Upgrade: WebSocket\r\n" + "Sec-WebSocket-Key1: 4 @1 46546xW%0l 1 5\r\n" + "Origin: http://example.com\r\n" + "\r\n" + "Hot diggity dogg" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/demo" + ,.request_url= "/demo" + ,.num_headers= 7 + ,.upgrade="Hot diggity dogg" + ,.headers= { { "Host", "example.com" } + , { "Connection", "Upgrade" } + , { "Sec-WebSocket-Key2", "12998 5 Y3 1 .P00" } + , { "Sec-WebSocket-Protocol", "sample" } + , { "Upgrade", "WebSocket" } + , { "Sec-WebSocket-Key1", "4 @1 46546xW%0l 1 5" } + , { "Origin", "http://example.com" } + } + ,.body= "" + } + +#define CONNECT_REQUEST 17 +, {.name = "connect request" + ,.type= HTTP_REQUEST + ,.raw= "CONNECT 0-home0.netscape.com:443 HTTP/1.0\r\n" + "User-agent: Mozilla/1.1N\r\n" + "Proxy-authorization: basic aGVsbG86d29ybGQ=\r\n" + "\r\n" + "some data\r\n" + "and yet even more data" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 0 + ,.method= HTTP_CONNECT + ,.query_string= "" + ,.fragment= "" + ,.request_path= "" + ,.request_url= "0-home0.netscape.com:443" + ,.num_headers= 2 + ,.upgrade="some data\r\nand yet even more data" + ,.headers= { { "User-agent", "Mozilla/1.1N" } + , { "Proxy-authorization", "basic aGVsbG86d29ybGQ=" } + } + ,.body= "" + } + +#define REPORT_REQ 18 +, {.name= "report request" + ,.type= HTTP_REQUEST + ,.raw= "REPORT /test HTTP/1.1\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_REPORT + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/test" + ,.request_url= "/test" + ,.num_headers= 0 + ,.headers= {} + ,.body= "" + } + +#define NO_HTTP_VERSION 19 +, {.name= "request with no http version" + ,.type= HTTP_REQUEST + ,.raw= "GET /\r\n" + "\r\n" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= FALSE + ,.http_major= 0 + ,.http_minor= 9 + ,.method= HTTP_GET + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/" + ,.request_url= "/" + ,.num_headers= 0 + ,.headers= {} + ,.body= "" + } + +#define MSEARCH_REQ 20 +, {.name= "m-search request" + ,.type= HTTP_REQUEST + ,.raw= "M-SEARCH * HTTP/1.1\r\n" + "HOST: 239.255.255.250:1900\r\n" + "MAN: \"ssdp:discover\"\r\n" + "ST: \"ssdp:all\"\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_MSEARCH + ,.query_string= "" + ,.fragment= "" + ,.request_path= "*" + ,.request_url= "*" + ,.num_headers= 3 + ,.headers= { { "HOST", "239.255.255.250:1900" } + , { "MAN", "\"ssdp:discover\"" } + , { "ST", "\"ssdp:all\"" } + } + ,.body= "" + } + +#define LINE_FOLDING_IN_HEADER 21 +, {.name= "line folding in header value" + ,.type= HTTP_REQUEST + ,.raw= "GET / HTTP/1.1\r\n" + "Line1: abc\r\n" + "\tdef\r\n" + " ghi\r\n" + "\t\tjkl\r\n" + " mno \r\n" + "\t \tqrs\r\n" + "Line2: \t line2\t\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/" + ,.request_url= "/" + ,.num_headers= 2 + ,.headers= { { "Line1", "abcdefghijklmno qrs" } + , { "Line2", "line2\t" } + } + ,.body= "" + } + + +#define QUERY_TERMINATED_HOST 22 +, {.name= "host terminated by a query string" + ,.type= HTTP_REQUEST + ,.raw= "GET http://hypnotoad.org?hail=all HTTP/1.1\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "hail=all" + ,.fragment= "" + ,.request_path= "" + ,.request_url= "http://hypnotoad.org?hail=all" + ,.num_headers= 0 + ,.headers= { } + ,.body= "" + } + +#define QUERY_TERMINATED_HOSTPORT 23 +, {.name= "host:port terminated by a query string" + ,.type= HTTP_REQUEST + ,.raw= "GET http://hypnotoad.org:1234?hail=all HTTP/1.1\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "hail=all" + ,.fragment= "" + ,.request_path= "" + ,.request_url= "http://hypnotoad.org:1234?hail=all" + ,.port= 1234 + ,.num_headers= 0 + ,.headers= { } + ,.body= "" + } + +#define SPACE_TERMINATED_HOSTPORT 24 +, {.name= "host:port terminated by a space" + ,.type= HTTP_REQUEST + ,.raw= "GET http://hypnotoad.org:1234 HTTP/1.1\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "" + ,.fragment= "" + ,.request_path= "" + ,.request_url= "http://hypnotoad.org:1234" + ,.port= 1234 + ,.num_headers= 0 + ,.headers= { } + ,.body= "" + } + +#define PATCH_REQ 25 +, {.name = "PATCH request" + ,.type= HTTP_REQUEST + ,.raw= "PATCH /file.txt HTTP/1.1\r\n" + "Host: www.example.com\r\n" + "Content-Type: application/example\r\n" + "If-Match: \"e0023aa4e\"\r\n" + "Content-Length: 10\r\n" + "\r\n" + "cccccccccc" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_PATCH + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/file.txt" + ,.request_url= "/file.txt" + ,.num_headers= 4 + ,.headers= { { "Host", "www.example.com" } + , { "Content-Type", "application/example" } + , { "If-Match", "\"e0023aa4e\"" } + , { "Content-Length", "10" } + } + ,.body= "cccccccccc" + } + +#define CONNECT_CAPS_REQUEST 26 +, {.name = "connect caps request" + ,.type= HTTP_REQUEST + ,.raw= "CONNECT HOME0.NETSCAPE.COM:443 HTTP/1.0\r\n" + "User-agent: Mozilla/1.1N\r\n" + "Proxy-authorization: basic aGVsbG86d29ybGQ=\r\n" + "\r\n" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 0 + ,.method= HTTP_CONNECT + ,.query_string= "" + ,.fragment= "" + ,.request_path= "" + ,.request_url= "HOME0.NETSCAPE.COM:443" + ,.num_headers= 2 + ,.upgrade="" + ,.headers= { { "User-agent", "Mozilla/1.1N" } + , { "Proxy-authorization", "basic aGVsbG86d29ybGQ=" } + } + ,.body= "" + } + +#if !HTTP_PARSER_STRICT +#define UTF8_PATH_REQ 27 +, {.name= "utf-8 path request" + ,.type= HTTP_REQUEST + ,.raw= "GET /δ¶/δt/pope?q=1#narf HTTP/1.1\r\n" + "Host: github.com\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_GET + ,.query_string= "q=1" + ,.fragment= "narf" + ,.request_path= "/δ¶/δt/pope" + ,.request_url= "/δ¶/δt/pope?q=1#narf" + ,.num_headers= 1 + ,.headers= { {"Host", "github.com" } + } + ,.body= "" + } + +#define HOSTNAME_UNDERSCORE 28 +, {.name = "hostname underscore" + ,.type= HTTP_REQUEST + ,.raw= "CONNECT home_0.netscape.com:443 HTTP/1.0\r\n" + "User-agent: Mozilla/1.1N\r\n" + "Proxy-authorization: basic aGVsbG86d29ybGQ=\r\n" + "\r\n" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 0 + ,.method= HTTP_CONNECT + ,.query_string= "" + ,.fragment= "" + ,.request_path= "" + ,.request_url= "home_0.netscape.com:443" + ,.num_headers= 2 + ,.upgrade="" + ,.headers= { { "User-agent", "Mozilla/1.1N" } + , { "Proxy-authorization", "basic aGVsbG86d29ybGQ=" } + } + ,.body= "" + } +#endif /* !HTTP_PARSER_STRICT */ + +/* see https://github.com/ry/http-parser/issues/47 */ +#define EAT_TRAILING_CRLF_NO_CONNECTION_CLOSE 29 +, {.name = "eat CRLF between requests, no \"Connection: close\" header" + ,.raw= "POST / HTTP/1.1\r\n" + "Host: www.example.com\r\n" + "Content-Type: application/x-www-form-urlencoded\r\n" + "Content-Length: 4\r\n" + "\r\n" + "q=42\r\n" /* note the trailing CRLF */ + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_POST + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/" + ,.request_url= "/" + ,.num_headers= 3 + ,.upgrade= 0 + ,.headers= { { "Host", "www.example.com" } + , { "Content-Type", "application/x-www-form-urlencoded" } + , { "Content-Length", "4" } + } + ,.body= "q=42" + } + +/* see https://github.com/ry/http-parser/issues/47 */ +#define EAT_TRAILING_CRLF_WITH_CONNECTION_CLOSE 30 +, {.name = "eat CRLF between requests even if \"Connection: close\" is set" + ,.raw= "POST / HTTP/1.1\r\n" + "Host: www.example.com\r\n" + "Content-Type: application/x-www-form-urlencoded\r\n" + "Content-Length: 4\r\n" + "Connection: close\r\n" + "\r\n" + "q=42\r\n" /* note the trailing CRLF */ + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= FALSE /* input buffer isn't empty when on_message_complete is called */ + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_POST + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/" + ,.request_url= "/" + ,.num_headers= 4 + ,.upgrade= 0 + ,.headers= { { "Host", "www.example.com" } + , { "Content-Type", "application/x-www-form-urlencoded" } + , { "Content-Length", "4" } + , { "Connection", "close" } + } + ,.body= "q=42" + } + +#define PURGE_REQ 31 +, {.name = "PURGE request" + ,.type= HTTP_REQUEST + ,.raw= "PURGE /file.txt HTTP/1.1\r\n" + "Host: www.example.com\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_PURGE + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/file.txt" + ,.request_url= "/file.txt" + ,.num_headers= 1 + ,.headers= { { "Host", "www.example.com" } } + ,.body= "" + } + +#define SEARCH_REQ 32 +, {.name = "SEARCH request" + ,.type= HTTP_REQUEST + ,.raw= "SEARCH / HTTP/1.1\r\n" + "Host: www.example.com\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.method= HTTP_SEARCH + ,.query_string= "" + ,.fragment= "" + ,.request_path= "/" + ,.request_url= "/" + ,.num_headers= 1 + ,.headers= { { "Host", "www.example.com" } } + ,.body= "" + } + +, {.name= NULL } /* sentinel */ +}; + +/* * R E S P O N S E S * */ +const struct message responses[] = +#define GOOGLE_301 0 +{ {.name= "google 301" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 301 Moved Permanently\r\n" + "Location: http://www.google.com/\r\n" + "Content-Type: text/html; charset=UTF-8\r\n" + "Date: Sun, 26 Apr 2009 11:11:49 GMT\r\n" + "Expires: Tue, 26 May 2009 11:11:49 GMT\r\n" + "X-$PrototypeBI-Version: 1.6.0.3\r\n" /* $ char in header field */ + "Cache-Control: public, max-age=2592000\r\n" + "Server: gws\r\n" + "Content-Length: 219 \r\n" + "\r\n" + "\n" + "301 Moved\n" + "

301 Moved

\n" + "The document has moved\n" + "here.\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 301 + ,.num_headers= 8 + ,.headers= + { { "Location", "http://www.google.com/" } + , { "Content-Type", "text/html; charset=UTF-8" } + , { "Date", "Sun, 26 Apr 2009 11:11:49 GMT" } + , { "Expires", "Tue, 26 May 2009 11:11:49 GMT" } + , { "X-$PrototypeBI-Version", "1.6.0.3" } + , { "Cache-Control", "public, max-age=2592000" } + , { "Server", "gws" } + , { "Content-Length", "219 " } + } + ,.body= "\n" + "301 Moved\n" + "

301 Moved

\n" + "The document has moved\n" + "here.\r\n" + "\r\n" + } + +#define NO_CONTENT_LENGTH_RESPONSE 1 +/* The client should wait for the server's EOF. That is, when content-length + * is not specified, and "Connection: close", the end of body is specified + * by the EOF. + * Compare with APACHEBENCH_GET + */ +, {.name= "no content-length response" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 200 OK\r\n" + "Date: Tue, 04 Aug 2009 07:59:32 GMT\r\n" + "Server: Apache\r\n" + "X-Powered-By: Servlet/2.5 JSP/2.1\r\n" + "Content-Type: text/xml; charset=utf-8\r\n" + "Connection: close\r\n" + "\r\n" + "\n" + "\n" + " \n" + " \n" + " SOAP-ENV:Client\n" + " Client Error\n" + " \n" + " \n" + "" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= TRUE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 200 + ,.num_headers= 5 + ,.headers= + { { "Date", "Tue, 04 Aug 2009 07:59:32 GMT" } + , { "Server", "Apache" } + , { "X-Powered-By", "Servlet/2.5 JSP/2.1" } + , { "Content-Type", "text/xml; charset=utf-8" } + , { "Connection", "close" } + } + ,.body= "\n" + "\n" + " \n" + " \n" + " SOAP-ENV:Client\n" + " Client Error\n" + " \n" + " \n" + "" + } + +#define NO_HEADERS_NO_BODY_404 2 +, {.name= "404 no headers no body" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 404 Not Found\r\n\r\n" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= TRUE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 404 + ,.num_headers= 0 + ,.headers= {} + ,.body_size= 0 + ,.body= "" + } + +#define NO_REASON_PHRASE 3 +, {.name= "301 no response phrase" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 301\r\n\r\n" + ,.should_keep_alive = FALSE + ,.message_complete_on_eof= TRUE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 301 + ,.num_headers= 0 + ,.headers= {} + ,.body= "" + } + +#define TRAILING_SPACE_ON_CHUNKED_BODY 4 +, {.name="200 trailing space on chunked body" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 200 OK\r\n" + "Content-Type: text/plain\r\n" + "Transfer-Encoding: chunked\r\n" + "\r\n" + "25 \r\n" + "This is the data in the first chunk\r\n" + "\r\n" + "1C\r\n" + "and this is the second one\r\n" + "\r\n" + "0 \r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 200 + ,.num_headers= 2 + ,.headers= + { {"Content-Type", "text/plain" } + , {"Transfer-Encoding", "chunked" } + } + ,.body_size = 37+28 + ,.body = + "This is the data in the first chunk\r\n" + "and this is the second one\r\n" + + } + +#define NO_CARRIAGE_RET 5 +, {.name="no carriage ret" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 200 OK\n" + "Content-Type: text/html; charset=utf-8\n" + "Connection: close\n" + "\n" + "these headers are from http://news.ycombinator.com/" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= TRUE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 200 + ,.num_headers= 2 + ,.headers= + { {"Content-Type", "text/html; charset=utf-8" } + , {"Connection", "close" } + } + ,.body= "these headers are from http://news.ycombinator.com/" + } + +#define PROXY_CONNECTION 6 +, {.name="proxy connection" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 200 OK\r\n" + "Content-Type: text/html; charset=UTF-8\r\n" + "Content-Length: 11\r\n" + "Proxy-Connection: close\r\n" + "Date: Thu, 31 Dec 2009 20:55:48 +0000\r\n" + "\r\n" + "hello world" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 200 + ,.num_headers= 4 + ,.headers= + { {"Content-Type", "text/html; charset=UTF-8" } + , {"Content-Length", "11" } + , {"Proxy-Connection", "close" } + , {"Date", "Thu, 31 Dec 2009 20:55:48 +0000"} + } + ,.body= "hello world" + } + +#define UNDERSTORE_HEADER_KEY 7 + // shown by + // curl -o /dev/null -v "http://ad.doubleclick.net/pfadx/DARTSHELLCONFIGXML;dcmt=text/xml;" +, {.name="underscore header key" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 200 OK\r\n" + "Server: DCLK-AdSvr\r\n" + "Content-Type: text/xml\r\n" + "Content-Length: 0\r\n" + "DCLK_imp: v7;x;114750856;0-0;0;17820020;0/0;21603567/21621457/1;;~okv=;dcmt=text/xml;;~cs=o\r\n\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 200 + ,.num_headers= 4 + ,.headers= + { {"Server", "DCLK-AdSvr" } + , {"Content-Type", "text/xml" } + , {"Content-Length", "0" } + , {"DCLK_imp", "v7;x;114750856;0-0;0;17820020;0/0;21603567/21621457/1;;~okv=;dcmt=text/xml;;~cs=o" } + } + ,.body= "" + } + +#define BONJOUR_MADAME_FR 8 +/* The client should not merge two headers fields when the first one doesn't + * have a value. + */ +, {.name= "bonjourmadame.fr" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.0 301 Moved Permanently\r\n" + "Date: Thu, 03 Jun 2010 09:56:32 GMT\r\n" + "Server: Apache/2.2.3 (Red Hat)\r\n" + "Cache-Control: public\r\n" + "Pragma: \r\n" + "Location: http://www.bonjourmadame.fr/\r\n" + "Vary: Accept-Encoding\r\n" + "Content-Length: 0\r\n" + "Content-Type: text/html; charset=UTF-8\r\n" + "Connection: keep-alive\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 0 + ,.status_code= 301 + ,.num_headers= 9 + ,.headers= + { { "Date", "Thu, 03 Jun 2010 09:56:32 GMT" } + , { "Server", "Apache/2.2.3 (Red Hat)" } + , { "Cache-Control", "public" } + , { "Pragma", "" } + , { "Location", "http://www.bonjourmadame.fr/" } + , { "Vary", "Accept-Encoding" } + , { "Content-Length", "0" } + , { "Content-Type", "text/html; charset=UTF-8" } + , { "Connection", "keep-alive" } + } + ,.body= "" + } + +#define RES_FIELD_UNDERSCORE 9 +/* Should handle spaces in header fields */ +, {.name= "field underscore" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 200 OK\r\n" + "Date: Tue, 28 Sep 2010 01:14:13 GMT\r\n" + "Server: Apache\r\n" + "Cache-Control: no-cache, must-revalidate\r\n" + "Expires: Mon, 26 Jul 1997 05:00:00 GMT\r\n" + ".et-Cookie: PlaxoCS=1274804622353690521; path=/; domain=.plaxo.com\r\n" + "Vary: Accept-Encoding\r\n" + "_eep-Alive: timeout=45\r\n" /* semantic value ignored */ + "_onnection: Keep-Alive\r\n" /* semantic value ignored */ + "Transfer-Encoding: chunked\r\n" + "Content-Type: text/html\r\n" + "Connection: close\r\n" + "\r\n" + "0\r\n\r\n" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 200 + ,.num_headers= 11 + ,.headers= + { { "Date", "Tue, 28 Sep 2010 01:14:13 GMT" } + , { "Server", "Apache" } + , { "Cache-Control", "no-cache, must-revalidate" } + , { "Expires", "Mon, 26 Jul 1997 05:00:00 GMT" } + , { ".et-Cookie", "PlaxoCS=1274804622353690521; path=/; domain=.plaxo.com" } + , { "Vary", "Accept-Encoding" } + , { "_eep-Alive", "timeout=45" } + , { "_onnection", "Keep-Alive" } + , { "Transfer-Encoding", "chunked" } + , { "Content-Type", "text/html" } + , { "Connection", "close" } + } + ,.body= "" + } + +#define NON_ASCII_IN_STATUS_LINE 10 +/* Should handle non-ASCII in status line */ +, {.name= "non-ASCII in status line" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 500 Oriëntatieprobleem\r\n" + "Date: Fri, 5 Nov 2010 23:07:12 GMT+2\r\n" + "Content-Length: 0\r\n" + "Connection: close\r\n" + "\r\n" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 500 + ,.num_headers= 3 + ,.headers= + { { "Date", "Fri, 5 Nov 2010 23:07:12 GMT+2" } + , { "Content-Length", "0" } + , { "Connection", "close" } + } + ,.body= "" + } + +#define HTTP_VERSION_0_9 11 +/* Should handle HTTP/0.9 */ +, {.name= "http version 0.9" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/0.9 200 OK\r\n" + "\r\n" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= TRUE + ,.http_major= 0 + ,.http_minor= 9 + ,.status_code= 200 + ,.num_headers= 0 + ,.headers= + {} + ,.body= "" + } + +#define NO_CONTENT_LENGTH_NO_TRANSFER_ENCODING_RESPONSE 12 +/* The client should wait for the server's EOF. That is, when neither + * content-length nor transfer-encoding is specified, the end of body + * is specified by the EOF. + */ +, {.name= "neither content-length nor transfer-encoding response" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 200 OK\r\n" + "Content-Type: text/plain\r\n" + "\r\n" + "hello world" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= TRUE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 200 + ,.num_headers= 1 + ,.headers= + { { "Content-Type", "text/plain" } + } + ,.body= "hello world" + } + +#define NO_BODY_HTTP10_KA_200 13 +, {.name= "HTTP/1.0 with keep-alive and EOF-terminated 200 status" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.0 200 OK\r\n" + "Connection: keep-alive\r\n" + "\r\n" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= TRUE + ,.http_major= 1 + ,.http_minor= 0 + ,.status_code= 200 + ,.num_headers= 1 + ,.headers= + { { "Connection", "keep-alive" } + } + ,.body_size= 0 + ,.body= "" + } + +#define NO_BODY_HTTP10_KA_204 14 +, {.name= "HTTP/1.0 with keep-alive and a 204 status" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.0 204 No content\r\n" + "Connection: keep-alive\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 0 + ,.status_code= 204 + ,.num_headers= 1 + ,.headers= + { { "Connection", "keep-alive" } + } + ,.body_size= 0 + ,.body= "" + } + +#define NO_BODY_HTTP11_KA_200 15 +, {.name= "HTTP/1.1 with an EOF-terminated 200 status" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 200 OK\r\n" + "\r\n" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= TRUE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 200 + ,.num_headers= 0 + ,.headers={} + ,.body_size= 0 + ,.body= "" + } + +#define NO_BODY_HTTP11_KA_204 16 +, {.name= "HTTP/1.1 with a 204 status" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 204 No content\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 204 + ,.num_headers= 0 + ,.headers={} + ,.body_size= 0 + ,.body= "" + } + +#define NO_BODY_HTTP11_NOKA_204 17 +, {.name= "HTTP/1.1 with a 204 status and keep-alive disabled" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 204 No content\r\n" + "Connection: close\r\n" + "\r\n" + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 204 + ,.num_headers= 1 + ,.headers= + { { "Connection", "close" } + } + ,.body_size= 0 + ,.body= "" + } + +#define NO_BODY_HTTP11_KA_CHUNKED_200 18 +, {.name= "HTTP/1.1 with chunked endocing and a 200 response" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 200 OK\r\n" + "Transfer-Encoding: chunked\r\n" + "\r\n" + "0\r\n" + "\r\n" + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 200 + ,.num_headers= 1 + ,.headers= + { { "Transfer-Encoding", "chunked" } + } + ,.body_size= 0 + ,.body= "" + } + +#if !HTTP_PARSER_STRICT +#define SPACE_IN_FIELD_RES 19 +/* Should handle spaces in header fields */ +, {.name= "field space" + ,.type= HTTP_RESPONSE + ,.raw= "HTTP/1.1 200 OK\r\n" + "Server: Microsoft-IIS/6.0\r\n" + "X-Powered-By: ASP.NET\r\n" + "en-US Content-Type: text/xml\r\n" /* this is the problem */ + "Content-Type: text/xml\r\n" + "Content-Length: 16\r\n" + "Date: Fri, 23 Jul 2010 18:45:38 GMT\r\n" + "Connection: keep-alive\r\n" + "\r\n" + "hello" /* fake body */ + ,.should_keep_alive= TRUE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 1 + ,.status_code= 200 + ,.num_headers= 7 + ,.headers= + { { "Server", "Microsoft-IIS/6.0" } + , { "X-Powered-By", "ASP.NET" } + , { "en-US Content-Type", "text/xml" } + , { "Content-Type", "text/xml" } + , { "Content-Length", "16" } + , { "Date", "Fri, 23 Jul 2010 18:45:38 GMT" } + , { "Connection", "keep-alive" } + } + ,.body= "hello" + } +#endif /* !HTTP_PARSER_STRICT */ + +, {.name= NULL } /* sentinel */ +}; + +int +request_url_cb (http_parser *p, const char *buf, size_t len) +{ + assert(p == parser); + strncat(messages[num_messages].request_url, buf, len); + return 0; +} + +int +header_field_cb (http_parser *p, const char *buf, size_t len) +{ + assert(p == parser); + struct message *m = &messages[num_messages]; + + if (m->last_header_element != FIELD) + m->num_headers++; + + strncat(m->headers[m->num_headers-1][0], buf, len); + + m->last_header_element = FIELD; + + return 0; +} + +int +header_value_cb (http_parser *p, const char *buf, size_t len) +{ + assert(p == parser); + struct message *m = &messages[num_messages]; + + strncat(m->headers[m->num_headers-1][1], buf, len); + + m->last_header_element = VALUE; + + return 0; +} + +int +body_cb (http_parser *p, const char *buf, size_t len) +{ + assert(p == parser); + strncat(messages[num_messages].body, buf, len); + messages[num_messages].body_size += len; + // printf("body_cb: '%s'\n", requests[num_messages].body); + return 0; +} + +int +count_body_cb (http_parser *p, const char *buf, size_t len) +{ + assert(p == parser); + assert(buf); + messages[num_messages].body_size += len; + return 0; +} + +int +message_begin_cb (http_parser *p) +{ + assert(p == parser); + messages[num_messages].message_begin_cb_called = TRUE; + return 0; +} + +int +headers_complete_cb (http_parser *p) +{ + assert(p == parser); + messages[num_messages].method = parser->method; + messages[num_messages].status_code = parser->status_code; + messages[num_messages].http_major = parser->http_major; + messages[num_messages].http_minor = parser->http_minor; + messages[num_messages].headers_complete_cb_called = TRUE; + messages[num_messages].should_keep_alive = http_should_keep_alive(parser); + return 0; +} + +int +message_complete_cb (http_parser *p) +{ + assert(p == parser); + if (messages[num_messages].should_keep_alive != http_should_keep_alive(parser)) + { + fprintf(stderr, "\n\n *** Error http_should_keep_alive() should have same " + "value in both on_message_complete and on_headers_complete " + "but it doesn't! ***\n\n"); + assert(0); + abort(); + } + messages[num_messages].message_complete_cb_called = TRUE; + + messages[num_messages].message_complete_on_eof = currently_parsing_eof; + + num_messages++; + return 0; +} + +/* These dontcall_* callbacks exist so that we can verify that when we're + * paused, no additional callbacks are invoked */ +int +dontcall_message_begin_cb (http_parser *p) +{ + if (p) { } // gcc + fprintf(stderr, "\n\n*** on_message_begin() called on paused parser ***\n\n"); + abort(); +} + +int +dontcall_header_field_cb (http_parser *p, const char *buf, size_t len) +{ + if (p || buf || len) { } // gcc + fprintf(stderr, "\n\n*** on_header_field() called on paused parser ***\n\n"); + abort(); +} + +int +dontcall_header_value_cb (http_parser *p, const char *buf, size_t len) +{ + if (p || buf || len) { } // gcc + fprintf(stderr, "\n\n*** on_header_value() called on paused parser ***\n\n"); + abort(); +} + +int +dontcall_request_url_cb (http_parser *p, const char *buf, size_t len) +{ + if (p || buf || len) { } // gcc + fprintf(stderr, "\n\n*** on_request_url() called on paused parser ***\n\n"); + abort(); +} + +int +dontcall_body_cb (http_parser *p, const char *buf, size_t len) +{ + if (p || buf || len) { } // gcc + fprintf(stderr, "\n\n*** on_body_cb() called on paused parser ***\n\n"); + abort(); +} + +int +dontcall_headers_complete_cb (http_parser *p) +{ + if (p) { } // gcc + fprintf(stderr, "\n\n*** on_headers_complete() called on paused " + "parser ***\n\n"); + abort(); +} + +int +dontcall_message_complete_cb (http_parser *p) +{ + if (p) { } // gcc + fprintf(stderr, "\n\n*** on_message_complete() called on paused " + "parser ***\n\n"); + abort(); +} + +static http_parser_settings settings_dontcall = + {.on_message_begin = dontcall_message_begin_cb + ,.on_header_field = dontcall_header_field_cb + ,.on_header_value = dontcall_header_value_cb + ,.on_url = dontcall_request_url_cb + ,.on_body = dontcall_body_cb + ,.on_headers_complete = dontcall_headers_complete_cb + ,.on_message_complete = dontcall_message_complete_cb + }; + +/* These pause_* callbacks always pause the parser and just invoke the regular + * callback that tracks content. Before returning, we overwrite the parser + * settings to point to the _dontcall variety so that we can verify that + * the pause actually did, you know, pause. */ +int +pause_message_begin_cb (http_parser *p) +{ + http_parser_pause(p, 1); + *current_pause_parser = settings_dontcall; + return message_begin_cb(p); +} + +int +pause_header_field_cb (http_parser *p, const char *buf, size_t len) +{ + http_parser_pause(p, 1); + *current_pause_parser = settings_dontcall; + return header_field_cb(p, buf, len); +} + +int +pause_header_value_cb (http_parser *p, const char *buf, size_t len) +{ + http_parser_pause(p, 1); + *current_pause_parser = settings_dontcall; + return header_value_cb(p, buf, len); +} + +int +pause_request_url_cb (http_parser *p, const char *buf, size_t len) +{ + http_parser_pause(p, 1); + *current_pause_parser = settings_dontcall; + return request_url_cb(p, buf, len); +} + +int +pause_body_cb (http_parser *p, const char *buf, size_t len) +{ + http_parser_pause(p, 1); + *current_pause_parser = settings_dontcall; + return body_cb(p, buf, len); +} + +int +pause_headers_complete_cb (http_parser *p) +{ + http_parser_pause(p, 1); + *current_pause_parser = settings_dontcall; + return headers_complete_cb(p); +} + +int +pause_message_complete_cb (http_parser *p) +{ + http_parser_pause(p, 1); + *current_pause_parser = settings_dontcall; + return message_complete_cb(p); +} + +static http_parser_settings settings_pause = + {.on_message_begin = pause_message_begin_cb + ,.on_header_field = pause_header_field_cb + ,.on_header_value = pause_header_value_cb + ,.on_url = pause_request_url_cb + ,.on_body = pause_body_cb + ,.on_headers_complete = pause_headers_complete_cb + ,.on_message_complete = pause_message_complete_cb + }; + +static http_parser_settings settings = + {.on_message_begin = message_begin_cb + ,.on_header_field = header_field_cb + ,.on_header_value = header_value_cb + ,.on_url = request_url_cb + ,.on_body = body_cb + ,.on_headers_complete = headers_complete_cb + ,.on_message_complete = message_complete_cb + }; + +static http_parser_settings settings_count_body = + {.on_message_begin = message_begin_cb + ,.on_header_field = header_field_cb + ,.on_header_value = header_value_cb + ,.on_url = request_url_cb + ,.on_body = count_body_cb + ,.on_headers_complete = headers_complete_cb + ,.on_message_complete = message_complete_cb + }; + +static http_parser_settings settings_null = + {.on_message_begin = 0 + ,.on_header_field = 0 + ,.on_header_value = 0 + ,.on_url = 0 + ,.on_body = 0 + ,.on_headers_complete = 0 + ,.on_message_complete = 0 + }; + +void +parser_init (enum http_parser_type type) +{ + num_messages = 0; + + assert(parser == NULL); + + parser = malloc(sizeof(http_parser)); + + http_parser_init(parser, type); + + memset(&messages, 0, sizeof messages); + +} + +void +parser_free () +{ + assert(parser); + free(parser); + parser = NULL; +} + +size_t parse (const char *buf, size_t len) +{ + size_t nparsed; + currently_parsing_eof = (len == 0); + nparsed = http_parser_execute(parser, &settings, buf, len); + return nparsed; +} + +size_t parse_count_body (const char *buf, size_t len) +{ + size_t nparsed; + currently_parsing_eof = (len == 0); + nparsed = http_parser_execute(parser, &settings_count_body, buf, len); + return nparsed; +} + +size_t parse_pause (const char *buf, size_t len) +{ + size_t nparsed; + http_parser_settings s = settings_pause; + + currently_parsing_eof = (len == 0); + current_pause_parser = &s; + nparsed = http_parser_execute(parser, current_pause_parser, buf, len); + return nparsed; +} + +static inline int +check_str_eq (const struct message *m, + const char *prop, + const char *expected, + const char *found) { + if ((expected == NULL) != (found == NULL)) { + printf("\n*** Error: %s in '%s' ***\n\n", prop, m->name); + printf("expected %s\n", (expected == NULL) ? "NULL" : expected); + printf(" found %s\n", (found == NULL) ? "NULL" : found); + return 0; + } + if (expected != NULL && 0 != strcmp(expected, found)) { + printf("\n*** Error: %s in '%s' ***\n\n", prop, m->name); + printf("expected '%s'\n", expected); + printf(" found '%s'\n", found); + return 0; + } + return 1; +} + +static inline int +check_num_eq (const struct message *m, + const char *prop, + int expected, + int found) { + if (expected != found) { + printf("\n*** Error: %s in '%s' ***\n\n", prop, m->name); + printf("expected %d\n", expected); + printf(" found %d\n", found); + return 0; + } + return 1; +} + +#define MESSAGE_CHECK_STR_EQ(expected, found, prop) \ + if (!check_str_eq(expected, #prop, expected->prop, found->prop)) return 0 + +#define MESSAGE_CHECK_NUM_EQ(expected, found, prop) \ + if (!check_num_eq(expected, #prop, expected->prop, found->prop)) return 0 + +#define MESSAGE_CHECK_URL_EQ(u, expected, found, prop, fn) \ +do { \ + char ubuf[256]; \ + \ + if ((u)->field_set & (1 << (fn))) { \ + memcpy(ubuf, (found)->request_url + (u)->field_data[(fn)].off, \ + (u)->field_data[(fn)].len); \ + ubuf[(u)->field_data[(fn)].len] = '\0'; \ + } else { \ + ubuf[0] = '\0'; \ + } \ + \ + check_str_eq(expected, #prop, expected->prop, ubuf); \ +} while(0) + +int +message_eq (int index, const struct message *expected) +{ + int i; + struct message *m = &messages[index]; + + MESSAGE_CHECK_NUM_EQ(expected, m, http_major); + MESSAGE_CHECK_NUM_EQ(expected, m, http_minor); + + if (expected->type == HTTP_REQUEST) { + MESSAGE_CHECK_NUM_EQ(expected, m, method); + } else { + MESSAGE_CHECK_NUM_EQ(expected, m, status_code); + } + + MESSAGE_CHECK_NUM_EQ(expected, m, should_keep_alive); + MESSAGE_CHECK_NUM_EQ(expected, m, message_complete_on_eof); + + assert(m->message_begin_cb_called); + assert(m->headers_complete_cb_called); + assert(m->message_complete_cb_called); + + + MESSAGE_CHECK_STR_EQ(expected, m, request_url); + + /* Check URL components; we can't do this w/ CONNECT since it doesn't + * send us a well-formed URL. + */ + if (*m->request_url && m->method != HTTP_CONNECT) { + struct http_parser_url u; + + if (http_parser_parse_url(m->request_url, strlen(m->request_url), 0, &u)) { + fprintf(stderr, "\n\n*** failed to parse URL %s ***\n\n", + m->request_url); + abort(); + } + + m->port = (u.field_set & (1 << UF_PORT)) ? + u.port : 0; + + MESSAGE_CHECK_URL_EQ(&u, expected, m, query_string, UF_QUERY); + MESSAGE_CHECK_URL_EQ(&u, expected, m, fragment, UF_FRAGMENT); + MESSAGE_CHECK_URL_EQ(&u, expected, m, request_path, UF_PATH); + MESSAGE_CHECK_NUM_EQ(expected, m, port); + } + + if (expected->body_size) { + MESSAGE_CHECK_NUM_EQ(expected, m, body_size); + } else { + MESSAGE_CHECK_STR_EQ(expected, m, body); + } + + MESSAGE_CHECK_NUM_EQ(expected, m, num_headers); + + int r; + for (i = 0; i < m->num_headers; i++) { + r = check_str_eq(expected, "header field", expected->headers[i][0], m->headers[i][0]); + if (!r) return 0; + r = check_str_eq(expected, "header value", expected->headers[i][1], m->headers[i][1]); + if (!r) return 0; + } + + MESSAGE_CHECK_STR_EQ(expected, m, upgrade); + + return 1; +} + +/* Given a sequence of varargs messages, return the number of them that the + * parser should successfully parse, taking into account that upgraded + * messages prevent all subsequent messages from being parsed. + */ +size_t +count_parsed_messages(const size_t nmsgs, ...) { + size_t i; + va_list ap; + + va_start(ap, nmsgs); + + for (i = 0; i < nmsgs; i++) { + struct message *m = va_arg(ap, struct message *); + + if (m->upgrade) { + va_end(ap); + return i + 1; + } + } + + va_end(ap); + return nmsgs; +} + +/* Given a sequence of bytes and the number of these that we were able to + * parse, verify that upgrade bodies are correct. + */ +void +upgrade_message_fix(char *body, const size_t nread, const size_t nmsgs, ...) { + va_list ap; + size_t i; + size_t off = 0; + + va_start(ap, nmsgs); + + for (i = 0; i < nmsgs; i++) { + struct message *m = va_arg(ap, struct message *); + + off += strlen(m->raw); + + if (m->upgrade) { + off -= strlen(m->upgrade); + + /* Check the portion of the response after its specified upgrade */ + if (!check_str_eq(m, "upgrade", body + off, body + nread)) { + abort(); + } + + /* Fix up the response so that message_eq() will verify the beginning + * of the upgrade */ + *(body + nread + strlen(m->upgrade)) = '\0'; + messages[num_messages -1 ].upgrade = body + nread; + + va_end(ap); + return; + } + } + + va_end(ap); + printf("\n\n*** Error: expected a message with upgrade ***\n"); + + abort(); +} + +static void +print_error (const char *raw, size_t error_location) +{ + fprintf(stderr, "\n*** %s:%d -- %s ***\n\n", + "http_parser.c", HTTP_PARSER_ERRNO_LINE(parser), + http_errno_description(HTTP_PARSER_ERRNO(parser))); + + int this_line = 0, char_len = 0; + size_t i, j, len = strlen(raw), error_location_line = 0; + for (i = 0; i < len; i++) { + if (i == error_location) this_line = 1; + switch (raw[i]) { + case '\r': + char_len = 2; + fprintf(stderr, "\\r"); + break; + + case '\n': + char_len = 2; + fprintf(stderr, "\\n\n"); + + if (this_line) goto print; + + error_location_line = 0; + continue; + + default: + char_len = 1; + fputc(raw[i], stderr); + break; + } + if (!this_line) error_location_line += char_len; + } + + fprintf(stderr, "[eof]\n"); + + print: + for (j = 0; j < error_location_line; j++) { + fputc(' ', stderr); + } + fprintf(stderr, "^\n\nerror location: %u\n", (unsigned int)error_location); +} + +void +test_preserve_data (void) +{ + char my_data[] = "application-specific data"; + http_parser parser; + parser.data = my_data; + http_parser_init(&parser, HTTP_REQUEST); + if (parser.data != my_data) { + printf("\n*** parser.data not preserved accross http_parser_init ***\n\n"); + abort(); + } +} + +struct url_test { + const char *name; + const char *url; + int is_connect; + struct http_parser_url u; + int rv; +}; + +const struct url_test url_tests[] = +{ {.name="proxy request" + ,.url="http://hostname/" + ,.is_connect=0 + ,.u= + {.field_set=(1 << UF_SCHEMA) | (1 << UF_HOST) | (1 << UF_PATH) + ,.port=0 + ,.field_data= + {{ 0, 4 } /* UF_SCHEMA */ + ,{ 7, 8 } /* UF_HOST */ + ,{ 0, 0 } /* UF_PORT */ + ,{ 15, 1 } /* UF_PATH */ + ,{ 0, 0 } /* UF_QUERY */ + ,{ 0, 0 } /* UF_FRAGMENT */ + } + } + ,.rv=0 + } + +, {.name="CONNECT request" + ,.url="hostname:443" + ,.is_connect=1 + ,.u= + {.field_set=(1 << UF_HOST) | (1 << UF_PORT) + ,.port=443 + ,.field_data= + {{ 0, 0 } /* UF_SCHEMA */ + ,{ 0, 8 } /* UF_HOST */ + ,{ 9, 3 } /* UF_PORT */ + ,{ 0, 0 } /* UF_PATH */ + ,{ 0, 0 } /* UF_QUERY */ + ,{ 0, 0 } /* UF_FRAGMENT */ + } + } + ,.rv=0 + } + +, {.name="proxy ipv6 request" + ,.url="http://[1:2::3:4]/" + ,.is_connect=0 + ,.u= + {.field_set=(1 << UF_SCHEMA) | (1 << UF_HOST) | (1 << UF_PATH) + ,.port=0 + ,.field_data= + {{ 0, 4 } /* UF_SCHEMA */ + ,{ 8, 8 } /* UF_HOST */ + ,{ 0, 0 } /* UF_PORT */ + ,{ 17, 1 } /* UF_PATH */ + ,{ 0, 0 } /* UF_QUERY */ + ,{ 0, 0 } /* UF_FRAGMENT */ + } + } + ,.rv=0 + } + +, {.name="CONNECT ipv6 address" + ,.url="[1:2::3:4]:443" + ,.is_connect=1 + ,.u= + {.field_set=(1 << UF_HOST) | (1 << UF_PORT) + ,.port=443 + ,.field_data= + {{ 0, 0 } /* UF_SCHEMA */ + ,{ 1, 8 } /* UF_HOST */ + ,{ 11, 3 } /* UF_PORT */ + ,{ 0, 0 } /* UF_PATH */ + ,{ 0, 0 } /* UF_QUERY */ + ,{ 0, 0 } /* UF_FRAGMENT */ + } + } + ,.rv=0 + } + +, {.name="extra ? in query string" + ,.url="http://a.tbcdn.cn/p/fp/2010c/??fp-header-min.css,fp-base-min.css,fp-channel-min.css,fp-product-min.css,fp-mall-min.css,fp-category-min.css,fp-sub-min.css,fp-gdp4p-min.css,fp-css3-min.css,fp-misc-min.css?t=20101022.css" + ,.is_connect=0 + ,.u= + {.field_set=(1<field_set, u->port); + for (i = 0; i < UF_MAX; i++) { + if ((u->field_set & (1 << i)) == 0) { + printf("\tfield_data[%u]: unset\n", i); + continue; + } + + memcpy(part, url + u->field_data[i].off, u->field_data[i].len); + part[u->field_data[i].len] = '\0'; + + printf("\tfield_data[%u]: off: %u len: %u part: \"%s\"\n", + i, + u->field_data[i].off, + u->field_data[i].len, + part); + } +} + +void +test_parse_url (void) +{ + struct http_parser_url u; + const struct url_test *test; + unsigned int i; + int rv; + + for (i = 0; i < (sizeof(url_tests) / sizeof(url_tests[0])); i++) { + test = &url_tests[i]; + memset(&u, 0, sizeof(u)); + + rv = http_parser_parse_url(test->url, + strlen(test->url), + test->is_connect, + &u); + + if (test->rv == 0) { + if (rv != 0) { + printf("\n*** http_parser_parse_url(\"%s\") \"%s\" test failed, " + "unexpected rv %d ***\n\n", test->url, test->name, rv); + abort(); + } + + if (memcmp(&u, &test->u, sizeof(u)) != 0) { + printf("\n*** http_parser_parse_url(\"%s\") \"%s\" failed ***\n", + test->url, test->name); + + printf("target http_parser_url:\n"); + dump_url(test->url, &test->u); + printf("result http_parser_url:\n"); + dump_url(test->url, &u); + + abort(); + } + } else { + /* test->rv != 0 */ + if (rv == 0) { + printf("\n*** http_parser_parse_url(\"%s\") \"%s\" test failed, " + "unexpected rv %d ***\n\n", test->url, test->name, rv); + abort(); + } + } + } +} + +void +test_message (const struct message *message) +{ + size_t raw_len = strlen(message->raw); + size_t msg1len; + for (msg1len = 0; msg1len < raw_len; msg1len++) { + parser_init(message->type); + + size_t read; + const char *msg1 = message->raw; + const char *msg2 = msg1 + msg1len; + size_t msg2len = raw_len - msg1len; + + if (msg1len) { + read = parse(msg1, msg1len); + + if (message->upgrade && parser->upgrade) { + messages[num_messages - 1].upgrade = msg1 + read; + goto test; + } + + if (read != msg1len) { + print_error(msg1, read); + abort(); + } + } + + + read = parse(msg2, msg2len); + + if (message->upgrade && parser->upgrade) { + messages[num_messages - 1].upgrade = msg2 + read; + goto test; + } + + if (read != msg2len) { + print_error(msg2, read); + abort(); + } + + read = parse(NULL, 0); + + if (read != 0) { + print_error(message->raw, read); + abort(); + } + + test: + + if (num_messages != 1) { + printf("\n*** num_messages != 1 after testing '%s' ***\n\n", message->name); + abort(); + } + + if(!message_eq(0, message)) abort(); + + parser_free(); + } +} + +void +test_message_count_body (const struct message *message) +{ + parser_init(message->type); + + size_t read; + size_t l = strlen(message->raw); + size_t i, toread; + size_t chunk = 4024; + + for (i = 0; i < l; i+= chunk) { + toread = MIN(l-i, chunk); + read = parse_count_body(message->raw + i, toread); + if (read != toread) { + print_error(message->raw, read); + abort(); + } + } + + + read = parse_count_body(NULL, 0); + if (read != 0) { + print_error(message->raw, read); + abort(); + } + + if (num_messages != 1) { + printf("\n*** num_messages != 1 after testing '%s' ***\n\n", message->name); + abort(); + } + + if(!message_eq(0, message)) abort(); + + parser_free(); +} + +void +test_simple (const char *buf, enum http_errno err_expected) +{ + parser_init(HTTP_REQUEST); + + size_t parsed; + int pass; + enum http_errno err; + + parsed = parse(buf, strlen(buf)); + pass = (parsed == strlen(buf)); + err = HTTP_PARSER_ERRNO(parser); + parsed = parse(NULL, 0); + pass &= (parsed == 0); + + parser_free(); + + /* In strict mode, allow us to pass with an unexpected HPE_STRICT as + * long as the caller isn't expecting success. + */ +#if HTTP_PARSER_STRICT + if (err_expected != err && err_expected != HPE_OK && err != HPE_STRICT) { +#else + if (err_expected != err) { +#endif + fprintf(stderr, "\n*** test_simple expected %s, but saw %s ***\n\n%s\n", + http_errno_name(err_expected), http_errno_name(err), buf); + abort(); + } +} + +void +test_header_overflow_error (int req) +{ + http_parser parser; + http_parser_init(&parser, req ? HTTP_REQUEST : HTTP_RESPONSE); + size_t parsed; + const char *buf; + buf = req ? "GET / HTTP/1.1\r\n" : "HTTP/1.0 200 OK\r\n"; + parsed = http_parser_execute(&parser, &settings_null, buf, strlen(buf)); + assert(parsed == strlen(buf)); + + buf = "header-key: header-value\r\n"; + size_t buflen = strlen(buf); + + int i; + for (i = 0; i < 10000; i++) { + parsed = http_parser_execute(&parser, &settings_null, buf, buflen); + if (parsed != buflen) { + //fprintf(stderr, "error found on iter %d\n", i); + assert(HTTP_PARSER_ERRNO(&parser) == HPE_HEADER_OVERFLOW); + return; + } + } + + fprintf(stderr, "\n*** Error expected but none in header overflow test ***\n"); + abort(); +} + +static void +test_content_length_overflow (const char *buf, size_t buflen, int expect_ok) +{ + http_parser parser; + http_parser_init(&parser, HTTP_RESPONSE); + http_parser_execute(&parser, &settings_null, buf, buflen); + + if (expect_ok) + assert(HTTP_PARSER_ERRNO(&parser) == HPE_OK); + else + assert(HTTP_PARSER_ERRNO(&parser) == HPE_INVALID_CONTENT_LENGTH); +} + +void +test_header_content_length_overflow_error (void) +{ +#define X(size) \ + "HTTP/1.1 200 OK\r\n" \ + "Content-Length: " #size "\r\n" \ + "\r\n" + const char a[] = X(18446744073709551614); /* 2^64-2 */ + const char b[] = X(18446744073709551615); /* 2^64-1 */ + const char c[] = X(18446744073709551616); /* 2^64 */ +#undef X + test_content_length_overflow(a, sizeof(a) - 1, 1); /* expect ok */ + test_content_length_overflow(b, sizeof(b) - 1, 0); /* expect failure */ + test_content_length_overflow(c, sizeof(c) - 1, 0); /* expect failure */ +} + +void +test_chunk_content_length_overflow_error (void) +{ +#define X(size) \ + "HTTP/1.1 200 OK\r\n" \ + "Transfer-Encoding: chunked\r\n" \ + "\r\n" \ + #size "\r\n" \ + "..." + const char a[] = X(FFFFFFFFFFFFFFFE); /* 2^64-2 */ + const char b[] = X(FFFFFFFFFFFFFFFF); /* 2^64-1 */ + const char c[] = X(10000000000000000); /* 2^64 */ +#undef X + test_content_length_overflow(a, sizeof(a) - 1, 1); /* expect ok */ + test_content_length_overflow(b, sizeof(b) - 1, 0); /* expect failure */ + test_content_length_overflow(c, sizeof(c) - 1, 0); /* expect failure */ +} + +void +test_no_overflow_long_body (int req, size_t length) +{ + http_parser parser; + http_parser_init(&parser, req ? HTTP_REQUEST : HTTP_RESPONSE); + size_t parsed; + size_t i; + char buf1[3000]; + size_t buf1len = sprintf(buf1, "%s\r\nConnection: Keep-Alive\r\nContent-Length: %lu\r\n\r\n", + req ? "POST / HTTP/1.0" : "HTTP/1.0 200 OK", (unsigned long)length); + parsed = http_parser_execute(&parser, &settings_null, buf1, buf1len); + if (parsed != buf1len) + goto err; + + for (i = 0; i < length; i++) { + char foo = 'a'; + parsed = http_parser_execute(&parser, &settings_null, &foo, 1); + if (parsed != 1) + goto err; + } + + parsed = http_parser_execute(&parser, &settings_null, buf1, buf1len); + if (parsed != buf1len) goto err; + return; + + err: + fprintf(stderr, + "\n*** error in test_no_overflow_long_body %s of length %lu ***\n", + req ? "REQUEST" : "RESPONSE", + (unsigned long)length); + abort(); +} + +void +test_multiple3 (const struct message *r1, const struct message *r2, const struct message *r3) +{ + int message_count = count_parsed_messages(3, r1, r2, r3); + + char total[ strlen(r1->raw) + + strlen(r2->raw) + + strlen(r3->raw) + + 1 + ]; + total[0] = '\0'; + + strcat(total, r1->raw); + strcat(total, r2->raw); + strcat(total, r3->raw); + + parser_init(r1->type); + + size_t read; + + read = parse(total, strlen(total)); + + if (parser->upgrade) { + upgrade_message_fix(total, read, 3, r1, r2, r3); + goto test; + } + + if (read != strlen(total)) { + print_error(total, read); + abort(); + } + + read = parse(NULL, 0); + + if (read != 0) { + print_error(total, read); + abort(); + } + +test: + + if (message_count != num_messages) { + fprintf(stderr, "\n\n*** Parser didn't see 3 messages only %d *** \n", num_messages); + abort(); + } + + if (!message_eq(0, r1)) abort(); + if (message_count > 1 && !message_eq(1, r2)) abort(); + if (message_count > 2 && !message_eq(2, r3)) abort(); + + parser_free(); +} + +/* SCAN through every possible breaking to make sure the + * parser can handle getting the content in any chunks that + * might come from the socket + */ +void +test_scan (const struct message *r1, const struct message *r2, const struct message *r3) +{ + char total[80*1024] = "\0"; + char buf1[80*1024] = "\0"; + char buf2[80*1024] = "\0"; + char buf3[80*1024] = "\0"; + + strcat(total, r1->raw); + strcat(total, r2->raw); + strcat(total, r3->raw); + + size_t read; + + int total_len = strlen(total); + + int total_ops = 2 * (total_len - 1) * (total_len - 2) / 2; + int ops = 0 ; + + size_t buf1_len, buf2_len, buf3_len; + int message_count = count_parsed_messages(3, r1, r2, r3); + + int i,j,type_both; + for (type_both = 0; type_both < 2; type_both ++ ) { + for (j = 2; j < total_len; j ++ ) { + for (i = 1; i < j; i ++ ) { + + if (ops % 1000 == 0) { + printf("\b\b\b\b%3.0f%%", 100 * (float)ops /(float)total_ops); + fflush(stdout); + } + ops += 1; + + parser_init(type_both ? HTTP_BOTH : r1->type); + + buf1_len = i; + strncpy(buf1, total, buf1_len); + buf1[buf1_len] = 0; + + buf2_len = j - i; + strncpy(buf2, total+i, buf2_len); + buf2[buf2_len] = 0; + + buf3_len = total_len - j; + strncpy(buf3, total+j, buf3_len); + buf3[buf3_len] = 0; + + read = parse(buf1, buf1_len); + + if (parser->upgrade) goto test; + + if (read != buf1_len) { + print_error(buf1, read); + goto error; + } + + read += parse(buf2, buf2_len); + + if (parser->upgrade) goto test; + + if (read != buf1_len + buf2_len) { + print_error(buf2, read); + goto error; + } + + read += parse(buf3, buf3_len); + + if (parser->upgrade) goto test; + + if (read != buf1_len + buf2_len + buf3_len) { + print_error(buf3, read); + goto error; + } + + parse(NULL, 0); + +test: + if (parser->upgrade) { + upgrade_message_fix(total, read, 3, r1, r2, r3); + } + + if (message_count != num_messages) { + fprintf(stderr, "\n\nParser didn't see %d messages only %d\n", + message_count, num_messages); + goto error; + } + + if (!message_eq(0, r1)) { + fprintf(stderr, "\n\nError matching messages[0] in test_scan.\n"); + goto error; + } + + if (message_count > 1 && !message_eq(1, r2)) { + fprintf(stderr, "\n\nError matching messages[1] in test_scan.\n"); + goto error; + } + + if (message_count > 2 && !message_eq(2, r3)) { + fprintf(stderr, "\n\nError matching messages[2] in test_scan.\n"); + goto error; + } + + parser_free(); + } + } + } + puts("\b\b\b\b100%"); + return; + + error: + fprintf(stderr, "i=%d j=%d\n", i, j); + fprintf(stderr, "buf1 (%u) %s\n\n", (unsigned int)buf1_len, buf1); + fprintf(stderr, "buf2 (%u) %s\n\n", (unsigned int)buf2_len , buf2); + fprintf(stderr, "buf3 (%u) %s\n", (unsigned int)buf3_len, buf3); + abort(); +} + +// user required to free the result +// string terminated by \0 +char * +create_large_chunked_message (int body_size_in_kb, const char* headers) +{ + int i; + size_t wrote = 0; + size_t headers_len = strlen(headers); + size_t bufsize = headers_len + (5+1024+2)*body_size_in_kb + 6; + char * buf = malloc(bufsize); + + memcpy(buf, headers, headers_len); + wrote += headers_len; + + for (i = 0; i < body_size_in_kb; i++) { + // write 1kb chunk into the body. + memcpy(buf + wrote, "400\r\n", 5); + wrote += 5; + memset(buf + wrote, 'C', 1024); + wrote += 1024; + strcpy(buf + wrote, "\r\n"); + wrote += 2; + } + + memcpy(buf + wrote, "0\r\n\r\n", 6); + wrote += 6; + assert(wrote == bufsize); + + return buf; +} + +/* Verify that we can pause parsing at any of the bytes in the + * message and still get the result that we're expecting. */ +void +test_message_pause (const struct message *msg) +{ + char *buf = (char*) msg->raw; + size_t buflen = strlen(msg->raw); + size_t nread; + + parser_init(msg->type); + + do { + nread = parse_pause(buf, buflen); + + // We can only set the upgrade buffer once we've gotten our message + // completion callback. + if (messages[0].message_complete_cb_called && + msg->upgrade && + parser->upgrade) { + messages[0].upgrade = buf + nread; + goto test; + } + + if (nread < buflen) { + + // Not much do to if we failed a strict-mode check + if (HTTP_PARSER_ERRNO(parser) == HPE_STRICT) { + parser_free(); + return; + } + + assert (HTTP_PARSER_ERRNO(parser) == HPE_PAUSED); + } + + buf += nread; + buflen -= nread; + http_parser_pause(parser, 0); + } while (buflen > 0); + + nread = parse_pause(NULL, 0); + assert (nread == 0); + +test: + if (num_messages != 1) { + printf("\n*** num_messages != 1 after testing '%s' ***\n\n", msg->name); + abort(); + } + + if(!message_eq(0, msg)) abort(); + + parser_free(); +} + +int +main (void) +{ + parser = NULL; + int i, j, k; + int request_count; + int response_count; + + printf("sizeof(http_parser) = %u\n", (unsigned int)sizeof(http_parser)); + + for (request_count = 0; requests[request_count].name; request_count++); + for (response_count = 0; responses[response_count].name; response_count++); + + //// API + test_preserve_data(); + test_parse_url(); + + //// OVERFLOW CONDITIONS + + test_header_overflow_error(HTTP_REQUEST); + test_no_overflow_long_body(HTTP_REQUEST, 1000); + test_no_overflow_long_body(HTTP_REQUEST, 100000); + + test_header_overflow_error(HTTP_RESPONSE); + test_no_overflow_long_body(HTTP_RESPONSE, 1000); + test_no_overflow_long_body(HTTP_RESPONSE, 100000); + + test_header_content_length_overflow_error(); + test_chunk_content_length_overflow_error(); + + //// RESPONSES + + for (i = 0; i < response_count; i++) { + test_message(&responses[i]); + } + + for (i = 0; i < response_count; i++) { + test_message_pause(&responses[i]); + } + + for (i = 0; i < response_count; i++) { + if (!responses[i].should_keep_alive) continue; + for (j = 0; j < response_count; j++) { + if (!responses[j].should_keep_alive) continue; + for (k = 0; k < response_count; k++) { + test_multiple3(&responses[i], &responses[j], &responses[k]); + } + } + } + + test_message_count_body(&responses[NO_HEADERS_NO_BODY_404]); + test_message_count_body(&responses[TRAILING_SPACE_ON_CHUNKED_BODY]); + + // test very large chunked response + { + char * msg = create_large_chunked_message(31337, + "HTTP/1.0 200 OK\r\n" + "Transfer-Encoding: chunked\r\n" + "Content-Type: text/plain\r\n" + "\r\n"); + struct message large_chunked = + {.name= "large chunked" + ,.type= HTTP_RESPONSE + ,.raw= msg + ,.should_keep_alive= FALSE + ,.message_complete_on_eof= FALSE + ,.http_major= 1 + ,.http_minor= 0 + ,.status_code= 200 + ,.num_headers= 2 + ,.headers= + { { "Transfer-Encoding", "chunked" } + , { "Content-Type", "text/plain" } + } + ,.body_size= 31337*1024 + }; + test_message_count_body(&large_chunked); + free(msg); + } + + + + printf("response scan 1/2 "); + test_scan( &responses[TRAILING_SPACE_ON_CHUNKED_BODY] + , &responses[NO_BODY_HTTP10_KA_204] + , &responses[NO_REASON_PHRASE] + ); + + printf("response scan 2/2 "); + test_scan( &responses[BONJOUR_MADAME_FR] + , &responses[UNDERSTORE_HEADER_KEY] + , &responses[NO_CARRIAGE_RET] + ); + + puts("responses okay"); + + + /// REQUESTS + + test_simple("hello world", HPE_INVALID_METHOD); + test_simple("GET / HTP/1.1\r\n\r\n", HPE_INVALID_VERSION); + + + test_simple("ASDF / HTTP/1.1\r\n\r\n", HPE_INVALID_METHOD); + test_simple("PROPPATCHA / HTTP/1.1\r\n\r\n", HPE_INVALID_METHOD); + test_simple("GETA / HTTP/1.1\r\n\r\n", HPE_INVALID_METHOD); + + // Well-formed but incomplete + test_simple("GET / HTTP/1.1\r\n" + "Content-Type: text/plain\r\n" + "Content-Length: 6\r\n" + "\r\n" + "fooba", + HPE_OK); + + static const char *all_methods[] = { + "DELETE", + "GET", + "HEAD", + "POST", + "PUT", + //"CONNECT", //CONNECT can't be tested like other methods, it's a tunnel + "OPTIONS", + "TRACE", + "COPY", + "LOCK", + "MKCOL", + "MOVE", + "PROPFIND", + "PROPPATCH", + "UNLOCK", + "REPORT", + "MKACTIVITY", + "CHECKOUT", + "MERGE", + "M-SEARCH", + "NOTIFY", + "SUBSCRIBE", + "UNSUBSCRIBE", + "PATCH", + 0 }; + const char **this_method; + for (this_method = all_methods; *this_method; this_method++) { + char buf[200]; + sprintf(buf, "%s / HTTP/1.1\r\n\r\n", *this_method); + test_simple(buf, HPE_OK); + } + + static const char *bad_methods[] = { + "C******", + "M****", + 0 }; + for (this_method = bad_methods; *this_method; this_method++) { + char buf[200]; + sprintf(buf, "%s / HTTP/1.1\r\n\r\n", *this_method); + test_simple(buf, HPE_UNKNOWN); + } + + const char *dumbfuck2 = + "GET / HTTP/1.1\r\n" + "X-SSL-Bullshit: -----BEGIN CERTIFICATE-----\r\n" + "\tMIIFbTCCBFWgAwIBAgICH4cwDQYJKoZIhvcNAQEFBQAwcDELMAkGA1UEBhMCVUsx\r\n" + "\tETAPBgNVBAoTCGVTY2llbmNlMRIwEAYDVQQLEwlBdXRob3JpdHkxCzAJBgNVBAMT\r\n" + "\tAkNBMS0wKwYJKoZIhvcNAQkBFh5jYS1vcGVyYXRvckBncmlkLXN1cHBvcnQuYWMu\r\n" + "\tdWswHhcNMDYwNzI3MTQxMzI4WhcNMDcwNzI3MTQxMzI4WjBbMQswCQYDVQQGEwJV\r\n" + "\tSzERMA8GA1UEChMIZVNjaWVuY2UxEzARBgNVBAsTCk1hbmNoZXN0ZXIxCzAJBgNV\r\n" + "\tBAcTmrsogriqMWLAk1DMRcwFQYDVQQDEw5taWNoYWVsIHBhcmQYJKoZIhvcNAQEB\r\n" + "\tBQADggEPADCCAQoCggEBANPEQBgl1IaKdSS1TbhF3hEXSl72G9J+WC/1R64fAcEF\r\n" + "\tW51rEyFYiIeZGx/BVzwXbeBoNUK41OK65sxGuflMo5gLflbwJtHBRIEKAfVVp3YR\r\n" + "\tgW7cMA/s/XKgL1GEC7rQw8lIZT8RApukCGqOVHSi/F1SiFlPDxuDfmdiNzL31+sL\r\n" + "\t0iwHDdNkGjy5pyBSB8Y79dsSJtCW/iaLB0/n8Sj7HgvvZJ7x0fr+RQjYOUUfrePP\r\n" + "\tu2MSpFyf+9BbC/aXgaZuiCvSR+8Snv3xApQY+fULK/xY8h8Ua51iXoQ5jrgu2SqR\r\n" + "\twgA7BUi3G8LFzMBl8FRCDYGUDy7M6QaHXx1ZWIPWNKsCAwEAAaOCAiQwggIgMAwG\r\n" + "\tA1UdEwEB/wQCMAAwEQYJYIZIAYb4QgHTTPAQDAgWgMA4GA1UdDwEB/wQEAwID6DAs\r\n" + "\tBglghkgBhvhCAQ0EHxYdVUsgZS1TY2llbmNlIFVzZXIgQ2VydGlmaWNhdGUwHQYD\r\n" + "\tVR0OBBYEFDTt/sf9PeMaZDHkUIldrDYMNTBZMIGaBgNVHSMEgZIwgY+AFAI4qxGj\r\n" + "\tloCLDdMVKwiljjDastqooXSkcjBwMQswCQYDVQQGEwJVSzERMA8GA1UEChMIZVNj\r\n" + "\taWVuY2UxEjAQBgNVBAsTCUF1dGhvcml0eTELMAkGA1UEAxMCQ0ExLTArBgkqhkiG\r\n" + "\t9w0BCQEWHmNhLW9wZXJhdG9yQGdyaWQtc3VwcG9ydC5hYy51a4IBADApBgNVHRIE\r\n" + "\tIjAggR5jYS1vcGVyYXRvckBncmlkLXN1cHBvcnQuYWMudWswGQYDVR0gBBIwEDAO\r\n" + "\tBgwrBgEEAdkvAQEBAQYwPQYJYIZIAYb4QgEEBDAWLmh0dHA6Ly9jYS5ncmlkLXN1\r\n" + "\tcHBvcnQuYWMudmT4sopwqlBWsvcHViL2NybC9jYWNybC5jcmwwPQYJYIZIAYb4QgEDBDAWLmh0\r\n" + "\tdHA6Ly9jYS5ncmlkLXN1cHBvcnQuYWMudWsvcHViL2NybC9jYWNybC5jcmwwPwYD\r\n" + "\tVR0fBDgwNjA0oDKgMIYuaHR0cDovL2NhLmdyaWQt5hYy51ay9wdWIv\r\n" + "\tY3JsL2NhY3JsLmNybDANBgkqhkiG9w0BAQUFAAOCAQEAS/U4iiooBENGW/Hwmmd3\r\n" + "\tXCy6Zrt08YjKCzGNjorT98g8uGsqYjSxv/hmi0qlnlHs+k/3Iobc3LjS5AMYr5L8\r\n" + "\tUO7OSkgFFlLHQyC9JzPfmLCAugvzEbyv4Olnsr8hbxF1MbKZoQxUZtMVu29wjfXk\r\n" + "\thTeApBv7eaKCWpSp7MCbvgzm74izKhu3vlDk9w6qVrxePfGgpKPqfHiOoGhFnbTK\r\n" + "\twTC6o2xq5y0qZ03JonF7OJspEd3I5zKY3E+ov7/ZhW6DqT8UFvsAdjvQbXyhV8Eu\r\n" + "\tYhixw1aKEPzNjNowuIseVogKOLXxWI5vAi5HgXdS0/ES5gDGsABo4fqovUKlgop3\r\n" + "\tRA==\r\n" + "\t-----END CERTIFICATE-----\r\n" + "\r\n"; + test_simple(dumbfuck2, HPE_OK); + +#if 0 + // NOTE(Wed Nov 18 11:57:27 CET 2009) this seems okay. we just read body + // until EOF. + // + // no content-length + // error if there is a body without content length + const char *bad_get_no_headers_no_body = "GET /bad_get_no_headers_no_body/world HTTP/1.1\r\n" + "Accept: */*\r\n" + "\r\n" + "HELLO"; + test_simple(bad_get_no_headers_no_body, 0); +#endif + /* TODO sending junk and large headers gets rejected */ + + + /* check to make sure our predefined requests are okay */ + for (i = 0; requests[i].name; i++) { + test_message(&requests[i]); + } + + for (i = 0; i < request_count; i++) { + test_message_pause(&requests[i]); + } + + for (i = 0; i < request_count; i++) { + if (!requests[i].should_keep_alive) continue; + for (j = 0; j < request_count; j++) { + if (!requests[j].should_keep_alive) continue; + for (k = 0; k < request_count; k++) { + test_multiple3(&requests[i], &requests[j], &requests[k]); + } + } + } + + printf("request scan 1/4 "); + test_scan( &requests[GET_NO_HEADERS_NO_BODY] + , &requests[GET_ONE_HEADER_NO_BODY] + , &requests[GET_NO_HEADERS_NO_BODY] + ); + + printf("request scan 2/4 "); + test_scan( &requests[POST_CHUNKED_ALL_YOUR_BASE] + , &requests[POST_IDENTITY_BODY_WORLD] + , &requests[GET_FUNKY_CONTENT_LENGTH] + ); + + printf("request scan 3/4 "); + test_scan( &requests[TWO_CHUNKS_MULT_ZERO_END] + , &requests[CHUNKED_W_TRAILING_HEADERS] + , &requests[CHUNKED_W_BULLSHIT_AFTER_LENGTH] + ); + + printf("request scan 4/4 "); + test_scan( &requests[QUERY_URL_WITH_QUESTION_MARK_GET] + , &requests[PREFIX_NEWLINE_GET ] + , &requests[CONNECT_REQUEST] + ); + + puts("requests okay"); + + return 0; +}