|
|
| version 1.32, 2009/08/05 08:57:08 | version 1.73, 2016/07/26 15:22:36 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: http support functions. | Parser: http support functions. |
| Copyright(c) 2001-2009 ArtLebedev Group (http://www.artlebedev.com) | Copyright (c) 2001-2015 Art. Lebedev Studio (http://www.artlebedev.com) |
| Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) | Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) |
| */ | */ |
| static const char * const IDENT_HTTP_C="$Date$"; | |
| #include "pa_http.h" | #include "pa_http.h" |
| #include "pa_common.h" | #include "pa_common.h" |
| #include "pa_charsets.h" | #include "pa_charsets.h" |
| Line 15 static const char * const IDENT_HTTP_C=" | Line 13 static const char * const IDENT_HTTP_C=" |
| #include "pa_vfile.h" | #include "pa_vfile.h" |
| #include "pa_random.h" | #include "pa_random.h" |
| volatile const char * IDENT_PA_HTTP_C="$Id$" IDENT_PA_HTTP_H; | |
| #ifdef _MSC_VER | |
| #include <windows.h> | |
| #else | |
| #define closesocket close | |
| #endif | |
| // defines | // defines |
| #define HTTP_METHOD_NAME "method" | #define HTTP_METHOD_NAME "method" |
| Line 22 static const char * const IDENT_HTTP_C=" | Line 28 static const char * const IDENT_HTTP_C=" |
| #define HTTP_BODY_NAME "body" | #define HTTP_BODY_NAME "body" |
| #define HTTP_TIMEOUT_NAME "timeout" | #define HTTP_TIMEOUT_NAME "timeout" |
| #define HTTP_HEADERS_NAME "headers" | #define HTTP_HEADERS_NAME "headers" |
| #define HTTP_COOKIES_NAME "cookies" | |
| #define HTTP_FORM_ENCTYPE_NAME "enctype" | #define HTTP_FORM_ENCTYPE_NAME "enctype" |
| #define HTTP_ANY_STATUS_NAME "any-status" | #define HTTP_ANY_STATUS_NAME "any-status" |
| #define HTTP_OMIT_POST_CHARSET_NAME "omit-post-charset" // ^file::load[...;http://...;$.form[...]$.method[post]] | #define HTTP_OMIT_POST_CHARSET_NAME "omit-post-charset" // ^file::load[...;http://...;$.method[post]] by default adds charset to content-type |
| // by default add charset to content-type | |
| #define HTTP_TABLES_NAME "tables" | |
| #define HTTP_USER "user" | #define HTTP_USER "user" |
| #define HTTP_PASSWORD "password" | #define HTTP_PASSWORD "password" |
| #define HTTP_USER_AGENT "user-agent" | |
| #define DEFAULT_USER_AGENT "parser3" | #define DEFAULT_USER_AGENT "parser3" |
| # ifndef INADDR_NONE | #ifndef INADDR_NONE |
| # define INADDR_NONE ((ulong) -1) | #define INADDR_NONE ((ulong) -1) |
| # endif | #endif |
| #undef CRLF | #undef CRLF |
| #define CRLF "\r\n" | #define CRLF "\r\n" |
| // helpers | |
| class Cookies_table_template_columns: public ArrayString { | |
| public: | |
| Cookies_table_template_columns() { | |
| *this+=new String("name"); | |
| *this+=new String("value"); | |
| *this+=new String("expires"); | |
| *this+=new String("max-age"); | |
| *this+=new String("domain"); | |
| *this+=new String("path"); | |
| *this+=new String("httponly"); | |
| *this+=new String("secure"); | |
| } | |
| }; | |
| static bool set_addr(struct sockaddr_in *addr, const char* host, const short port){ | static bool set_addr(struct sockaddr_in *addr, const char* host, const short port){ |
| memset(addr, 0, sizeof(*addr)); | memset(addr, 0, sizeof(*addr)); |
| addr->sin_family=AF_INET; | addr->sin_family=AF_INET; |
| addr->sin_port=htons(port); | addr->sin_port=htons(port); |
| if(host) { | if(host) { |
| ulong packed_ip=inet_addr(host); | struct hostent *hostIP=gethostbyname(host); |
| if(packed_ip!=INADDR_NONE) | if(hostIP && hostIP->h_addrtype == AF_INET){ |
| memcpy(&addr->sin_addr, &packed_ip, sizeof(packed_ip)); | memcpy(&addr->sin_addr, hostIP->h_addr, hostIP->h_length); |
| else { | return true; |
| struct hostent *hostIP=gethostbyname(host); | } |
| if(hostIP) | } |
| memcpy(&addr->sin_addr, hostIP->h_addr, hostIP->h_length); | return false; |
| else | |
| return false; | |
| } | |
| } else | |
| addr->sin_addr.s_addr=INADDR_ANY; | |
| return true; | |
| } | } |
| size_t guess_content_length(char* buf) { | size_t guess_content_length(char* buf) { |
| char* ptr; | char* ptr; |
| if((ptr=strstr(buf, "Content-Length:"))) // Apache | if((ptr=strstr(buf, "Content-Length:"))) // Apache |
| goto found; | goto found; |
| if((ptr=strstr(buf, "content-length:"))) // Parser 3 | if((ptr=strstr(buf, "content-length:"))) // Parser 3 before 3.4.0 |
| goto found; | goto found; |
| if((ptr=strstr(buf, "Content-length:"))) // maybe 1 | if((ptr=strstr(buf, "Content-length:"))) // maybe 1 |
| goto found; | goto found; |
| Line 75 size_t guess_content_length(char* buf) { | Line 89 size_t guess_content_length(char* buf) { |
| return 0; | return 0; |
| found: | found: |
| char *error_pos; | char *error_pos; |
| size_t result=(size_t)strtol(ptr+15/*strlen("CONTENT-LENGTH:")*/, &error_pos, 0); | size_t result=(size_t)strtol(ptr+15/*strlen("Content-Length:")*/, &error_pos, 0); |
| const size_t reasonable_initial_max=0x400*0x400*10 /*10M*/; | const size_t reasonable_initial_max=0x400*0x400*10 /*10M*/; |
| if(result>reasonable_initial_max) // sanity check | if(result>reasonable_initial_max) // sanity check |
| Line 85 found: | Line 99 found: |
| static int http_read_response(char*& response, size_t& response_size, int sock, bool fail_on_status_ne_200) { | static int http_read_response(char*& response, size_t& response_size, int sock, bool fail_on_status_ne_200) { |
| int result=0; | int result=0; |
| // fetching some to local buffer, guessing on possible content-length | // fetching some to local buffer, guessing on possible Content-Length |
| response_size=0x400*20; // initial size if content-length could not be determined | response_size=0x400*20; // initial size if Content-Length could not be determined |
| const size_t preview_size=0x400*20; | const size_t preview_size=0x400*20; |
| char preview_buf[preview_size+1/*terminator*/]; // 20K buffer to preview headers | char preview_buf[preview_size+1/*terminator*/]; // 20K buffer to preview headers |
| ssize_t received_size=recv(sock, preview_buf, preview_size, 0); | ssize_t received_size=recv(sock, preview_buf, preview_size, 0); |
| Line 94 static int http_read_response(char*& res | Line 108 static int http_read_response(char*& res |
| goto done; | goto done; |
| if(received_size<0) { | if(received_size<0) { |
| if(int no=pa_socks_errno()) | if(int no=pa_socks_errno()) |
| throw Exception("http.timeout", | throw Exception("http.timeout", 0, "error receiving response header: %s (%d)", pa_socks_strerr(no), no); |
| 0, | |
| "error receiving response header: %s (%d)", pa_socks_strerr(no), no); | |
| goto done; | goto done; |
| } | } |
| // terminator [helps futher string searches] | // terminator [helps futher string searches] |
| Line 111 static int http_read_response(char*& res | Line 123 static int http_read_response(char*& res |
| result=status_code.as_int(); | result=status_code.as_int(); |
| if(fail_on_status_ne_200 && result!=200) | if(fail_on_status_ne_200 && result!=200) |
| throw Exception("http.status", | throw Exception("http.status", &status_code, "invalid HTTP response status"); |
| &status_code, | |
| "invalid HTTP response status"); | |
| } | } |
| // detecting response_size | // detecting response_size |
| { | { |
| Line 135 static int http_read_response(char*& res | Line 145 static int http_read_response(char*& res |
| // we use terminator byte for two purposes here: | // we use terminator byte for two purposes here: |
| // 1. we return there zero always, not knowing: maybe they would want to create String form $file.body? | // 1. we return there zero always, not knowing: maybe they would want to create String form $file.body? |
| // invariant: all Strings should have zero-terminated buffers | // invariant: all Strings should have zero-terminated buffers |
| // 2. we use that out-of-size byte to detect if our content-length guess was wrong | // 2. we use that out-of-size byte to detect if our Content-Length guess was wrong |
| // when recv gets more than we expected | // when recv gets more than we expected |
| // a) we know that the content-length guess was wrong | // a) we know that the Content-Length guess was wrong |
| // b) we have space to put the first byte of extra data | // b) we have space to put the first byte of extra data |
| // c) we use less code to detect normal situation: on last while-cycle recv expected to just return 0 | // c) we use less code to detect normal situation: on last while-cycle recv expected to just return 0 |
| while(true) { | while(true) { |
| Line 148 static int http_read_response(char*& res | Line 158 static int http_read_response(char*& res |
| } | } |
| if(received_size<0) { | if(received_size<0) { |
| if(int no=pa_socks_errno()) | if(int no=pa_socks_errno()) |
| throw Exception("http.timeout", | throw Exception("http.timeout", 0, "error receiving response body: %s (%d)", pa_socks_strerr(no), no); |
| 0, | |
| "error receiving response body: %s (%d)", pa_socks_strerr(no), no); | |
| break; | break; |
| } | } |
| // they've touched the terminator? | // they've touched the terminator? |
| Line 176 done: | Line 184 done: |
| return result; | return result; |
| } | } |
| else | else |
| throw Exception("http.response", | throw Exception("http.response", 0, "bad response from host - no status found (size=%u)", response_size); |
| 0, | |
| "bad response from host - no status found (size=%u)", response_size); | |
| } | } |
| /* ********************** request *************************** */ | /* ********************** request *************************** */ |
| Line 194 static void timeout_handler(int /*sig*/) | Line 200 static void timeout_handler(int /*sig*/) |
| } | } |
| #endif | #endif |
| static size_t file_untaint(const char* str, size_t len) { | |
| // untaint file from L_FILE_POST encoding | |
| char* j=(char *)str; | |
| const char* end=str+len-1; | |
| for(const char* i=str; i<=end; i++, j++){ | |
| if(*i=='\\' && i!=end){ | |
| switch(*(i+1)){ | |
| case '0': | |
| *j='\0'; | |
| i++; | |
| continue; | |
| case '\\': | |
| *j='\\'; | |
| i++; | |
| continue; | |
| } | |
| } | |
| if(i!=j) | |
| *j=*i; | |
| } | |
| return j-str; // new length | |
| } | |
| static int http_request(char*& response, size_t& response_size, | static int http_request(char*& response, size_t& response_size, |
| const char* host, short port, | const char* host, short port, |
| const char* request, size_t request_size, | const char* request, size_t request_size, |
| int timeout_secs, | int timeout_secs, |
| bool fail_on_status_ne_200) { | bool fail_on_status_ne_200) { |
| if(!host) | if(!host) |
| throw Exception("http.host", | throw Exception("http.host", 0, "zero hostname"); //never |
| 0, | |
| "zero hostname"); //never | |
| volatile // to prevent makeing it register variable, because it will be clobbered by longjmp [thanks gcc warning] | volatile // to prevent makeing it register variable, because it will be clobbered by longjmp [thanks gcc warning] |
| int sock=-1; | int sock=-1; |
| Line 239 static int http_request(char*& response, | Line 220 static int http_request(char*& response, |
| // rewritten simplier [athough duplicating closesocket code] | // rewritten simplier [athough duplicating closesocket code] |
| if(sock>=0) | if(sock>=0) |
| closesocket(sock); | closesocket(sock); |
| throw Exception("http.timeout", | throw Exception("http.timeout", 0, "timeout occured while retrieving document"); |
| 0, | |
| "timeout occured while retrieving document"); | |
| return 0; // never | return 0; // never |
| } else { | } else { |
| alarm(timeout_secs); | alarm(timeout_secs); |
| Line 251 static int http_request(char*& response, | Line 230 static int http_request(char*& response, |
| struct sockaddr_in dest; | struct sockaddr_in dest; |
| if(!set_addr(&dest, host, port)) | if(!set_addr(&dest, host, port)) |
| throw Exception("http.host", | throw Exception("http.host", 0, "can not resolve hostname \"%s\"", host); |
| 0, | |
| "can not resolve hostname \"%s\"", host); | |
| if((sock=socket(AF_INET, SOCK_STREAM, IPPROTO_TCP/*0*/))<0) { | if((sock=socket(AF_INET, SOCK_STREAM, IPPROTO_TCP/*0*/))<0) { |
| int no=pa_socks_errno(); | int no=pa_socks_errno(); |
| throw Exception("http.connect", | throw Exception("http.connect", 0, "can not make socket: %s (%d)", pa_socks_strerr(no), no); |
| 0, | |
| "can not make socket: %s (%d)", pa_socks_strerr(no), no); | |
| } | } |
| // To enable SO_DONTLINGER (that is, disable SO_LINGER) | // To enable SO_DONTLINGER (that is, disable SO_LINGER) |
| Line 278 static int http_request(char*& response, | Line 253 static int http_request(char*& response, |
| if(connect(sock, (struct sockaddr *)&dest, sizeof(dest))) { | if(connect(sock, (struct sockaddr *)&dest, sizeof(dest))) { |
| int no=pa_socks_errno(); | int no=pa_socks_errno(); |
| throw Exception("http.connect", | throw Exception("http.connect", 0, "can not connect to host \"%s\": %s (%d)", host, pa_socks_strerr(no), no); |
| 0, | |
| "can not connect to host \"%s\": %s (%d)", host, pa_socks_strerr(no), no); | |
| } | } |
| if(send(sock, request, request_size, 0)!=(ssize_t)request_size) { | if(send(sock, request, request_size, 0)!=(ssize_t)request_size) { |
| int no=pa_socks_errno(); | int no=pa_socks_errno(); |
| throw Exception("http.timeout", | throw Exception("http.timeout", 0, "error sending request: %s (%d)", pa_socks_strerr(no), no); |
| 0, | |
| "error sending request: %s (%d)", pa_socks_strerr(no), no); | |
| } | } |
| result=http_read_response(response, response_size, sock, fail_on_status_ne_200); | result=http_read_response(response, response_size, sock, fail_on_status_ne_200); |
| Line 313 static int http_request(char*& response, | Line 284 static int http_request(char*& response, |
| struct Http_pass_header_info { | struct Http_pass_header_info { |
| Request_charsets* charsets; | Request_charsets* charsets; |
| String* request; | String* request; |
| bool user_agent_specified; | bool* user_agent_specified; |
| bool content_type_specified; | bool* content_type_specified; |
| bool* content_type_url_encoded; | |
| }; | }; |
| #endif | #endif |
| static void http_pass_header(HashStringValue::key_type name, | |
| HashStringValue::value_type value, | char *pa_http_safe_header_name(const char *name) { |
| char *result=pa_strdup(name); | |
| char *n=result; | |
| if(!pa_isalpha((unsigned char)*n)) | |
| *n++ = '_'; | |
| for(; *n; ++n) { | |
| if (!pa_isalnum((unsigned char)*n) && *n != '-' && *n != '_') | |
| *n = '_'; | |
| } | |
| return result; | |
| } | |
| static void http_pass_header(HashStringValue::key_type aname, | |
| HashStringValue::value_type avalue, | |
| Http_pass_header_info *info) { | Http_pass_header_info *info) { |
| String aname=String(name, String::L_URI); | const char* name_cstr=aname.cstr(); |
| *info->request << aname << ": " | if(strcasecmp(name_cstr, HTTP_CONTENT_LENGTH)==0) |
| << attributed_meaning_to_string(*value, String::L_URI, false) | return; |
| << CRLF; | |
| String name=String(pa_http_safe_header_name(capitalize(name_cstr)), String::L_AS_IS); | |
| String value=attributed_meaning_to_string(*avalue, String::L_HTTP_HEADER, true); | |
| *info->request << name << ": " << value << CRLF; | |
| const String::Body name_upper=aname.change_case(info->charsets->source(), String::CC_UPPER); | if(strcasecmp(name_cstr, HTTP_USER_AGENT)==0) |
| if(name_upper==HTTP_USER_AGENT_UPPER) | *info->user_agent_specified=true; |
| info->user_agent_specified=true; | if(strcasecmp(name_cstr, HTTP_CONTENT_TYPE)==0){ |
| if(name_upper==HTTP_CONTENT_TYPE_UPPER) | *info->content_type_specified=true; |
| info->content_type_specified=true; | *info->content_type_url_encoded=pa_strncasecmp(value.cstr(), HTTP_CONTENT_TYPE_FORM_URLENCODED)==0; |
| } | |
| } | } |
| static void http_pass_cookie(HashStringValue::key_type name, | static void http_pass_cookie(HashStringValue::key_type name, |
| Line 360 static const String* basic_authorization | Line 350 static const String* basic_authorization |
| return result; | return result; |
| } | } |
| static void form_string_value2string( | static void form_string_value2string(HashStringValue::key_type key, const String& value, String& result) { |
| HashStringValue::key_type key, | |
| const String& value, | |
| String& result) | |
| { | |
| result << String(key, String::L_URI) << "=" << String(value, String::L_URI) << "&"; | result << String(key, String::L_URI) << "=" << String(value, String::L_URI) << "&"; |
| } | } |
| Line 380 struct Form_table_value2string_info { | Line 366 struct Form_table_value2string_info { |
| static void form_table_value2string(Table::element_type row, Form_table_value2string_info* info) { | static void form_table_value2string(Table::element_type row, Form_table_value2string_info* info) { |
| form_string_value2string(info->key, *row->get(0), info->result); | form_string_value2string(info->key, *row->get(0), info->result); |
| } | } |
| static void form_value2string( | |
| HashStringValue::key_type key, | static void form_value2string(HashStringValue::key_type key, HashStringValue::value_type value, String* result) { |
| HashStringValue::value_type value, | |
| String* result) | |
| { | |
| if(const String* svalue=value->get_string()) | if(const String* svalue=value->get_string()) |
| form_string_value2string(key, *svalue, *result); | form_string_value2string(key, *svalue, *result); |
| else if(Table* tvalue=value->get_table()) { | else if(Table* tvalue=value->get_table()) { |
| Form_table_value2string_info info(key, *result); | Form_table_value2string_info info(key, *result); |
| tvalue->for_each(form_table_value2string, &info); | tvalue->for_each(form_table_value2string, &info); |
| } else | } else |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, new String(key, String::L_TAINTED), |
| new String(key, String::L_TAINTED), | "is %s, " HTTP_FORM_NAME " option value can be string or table only (file is allowed for $." HTTP_METHOD_NAME "[POST] + $." HTTP_FORM_ENCTYPE_NAME "[" HTTP_CONTENT_TYPE_MULTIPART_FORMDATA "])", value->type()); |
| "is %s, "HTTP_FORM_NAME" option value can be string or table only (file is allowed for $."HTTP_METHOD_NAME"[POST] + $."HTTP_FORM_ENCTYPE_NAME"["HTTP_CONTENT_TYPE_MULTIPART_FORMDATA"])", value->type()); | |
| } | } |
| const char* pa_form2string(HashStringValue& form, Request_charsets& charsets) { | const char* pa_form2string(HashStringValue& form, Request_charsets& charsets) { |
| String string; | String string; |
| form.for_each<String*>(form_value2string, &string); | form.for_each<String*>(form_value2string, &string); |
| return string.untaint_cstr(String::L_AS_IS, 0, &charsets); | return string.untaint_and_transcode_cstr(String::L_URI, &charsets); |
| } | } |
| struct FormPart { | struct FormPart { |
| Request* r; | Request* r; |
| const char* boundary; | const char* boundary; |
| String string; | String* string; |
| Form_table_value2string_info* info; | Form_table_value2string_info* info; |
| struct BinaryBlock{ | |
| const char* ptr; | |
| size_t length; | |
| BinaryBlock(String* astring, Request* r): ptr(astring->untaint_and_transcode_cstr(String::L_AS_IS, &r->charsets)), length(strlen(ptr)){} | |
| BinaryBlock(const char* aptr, size_t alength): ptr(aptr), length(alength){} | |
| }; | |
| Array<BinaryBlock> blocks; | |
| FormPart(Request* ar, const char* aboundary): r(ar), boundary(aboundary), string(new String()){} | |
| const char *post(size_t &length){ | |
| if(blocks.count()){ | |
| blocks+=BinaryBlock(string, r); | |
| length=0; | |
| for(size_t i=0; i<blocks.count(); i++) | |
| length+=blocks[i].length; | |
| char *result=(char *)pa_malloc_atomic(length); | |
| char *ptr=result; | |
| for(size_t i=0; i<blocks.count(); i++){ | |
| memcpy(ptr, blocks[i].ptr, blocks[i].length); | |
| ptr+=blocks[i].length; | |
| } | |
| return result; | |
| } else { | |
| BinaryBlock result(string, r); | |
| length=result.length; | |
| return result.ptr; | |
| } | |
| } | |
| }; | }; |
| static void form_part_boundary_header(FormPart& part, String::Body name, const char* file_name=0){ | static void form_part_boundary_header(FormPart& part, String::Body name, const char* file_name=0) { |
| part.string << "--" << part.boundary | *part.string << "--" << part.boundary << CRLF CONTENT_DISPOSITION_CAPITALIZED ": form-data; name=\"" << name << "\""; |
| << CRLF HTTP_CONTENT_DISPOSITION ": form-data; name=\"" | |
| << Charset::transcode(name, part.r->charsets.source(), part.r->charsets.client()) | |
| << "\""; | |
| if(file_name){ | if(file_name){ |
| if(strcmp(file_name, NONAME_DAT)!=0) | if(strcmp(file_name, NONAME_DAT)!=0) |
| part.string << "; filename=\"" << file_name << "\""; | *part.string << "; filename=\"" << file_name << "\""; |
| part.string << CRLF HTTP_CONTENT_TYPE ": " << part.r->mime_type_of(file_name); | *part.string << CRLF HTTP_CONTENT_TYPE_CAPITALIZED ": " << part.r->mime_type_of(file_name); |
| } | } |
| part.string << CRLF CRLF; | *part.string << CRLF CRLF; |
| } | } |
| static void form_string_value2part( | static void form_string_value2part(HashStringValue::key_type key, const String& value, FormPart& part) { |
| HashStringValue::key_type key, | |
| const String& value, | |
| FormPart& part) | |
| { | |
| form_part_boundary_header(part, key); | form_part_boundary_header(part, key); |
| part.string << Charset::transcode(value, part.r->charsets.source(), part.r->charsets.client()) << CRLF; | *part.string << value << CRLF; |
| } | } |
| static void form_file_value2part( | static void form_file_value2part(HashStringValue::key_type key, VFile& vfile, FormPart& part) { |
| HashStringValue::key_type key, | |
| VFile& vfile, | |
| FormPart& part) | |
| { | |
| form_part_boundary_header(part, key, vfile.fields().get(name_name)->as_string().cstr()); | form_part_boundary_header(part, key, vfile.fields().get(name_name)->as_string().cstr()); |
| part.string.append_know_length(vfile.value_ptr(), vfile.value_size(), String::L_FILE_POST); | part.blocks+=FormPart::BinaryBlock(part.string, part.r); |
| part.string << CRLF; | part.blocks+=FormPart::BinaryBlock(vfile.value_ptr(), vfile.value_size()); |
| part.string=new String(); | |
| *part.string << CRLF; | |
| } | } |
| static void form_table_value2part(Table::element_type row, FormPart* part) { | static void form_table_value2part(Table::element_type row, FormPart* part) { |
| form_string_value2part(part->info->key, *row->get(0), *part); | form_string_value2part(part->info->key, *row->get(0), *part); |
| } | } |
| static void form_value2part( | static void form_value2part(HashStringValue::key_type key, HashStringValue::value_type value, FormPart& part) { |
| HashStringValue::key_type key, | |
| HashStringValue::value_type value, | |
| FormPart& part) | |
| { | |
| if(const String* svalue=value->get_string()) | if(const String* svalue=value->get_string()) |
| form_string_value2part(key, *svalue, part); | form_string_value2part(key, *svalue, part); |
| else if(Table* tvalue=value->get_table()) { | else if(Table* tvalue=value->get_table()) { |
| Form_table_value2string_info info(key, part.string); | Form_table_value2string_info info(key, *part.string); |
| part.info = &info; | part.info = &info; |
| tvalue->for_each(form_table_value2part, &part); | tvalue->for_each(form_table_value2part, &part); |
| } else if(VFile* vfile=static_cast<VFile *>(value->as("file", false))){ | } else if(VFile* vfile=static_cast<VFile *>(value->as("file"))){ |
| form_file_value2part(key, *vfile, part); | form_file_value2part(key, *vfile, part); |
| } else | } else |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, new String(key, String::L_TAINTED), "is %s, " HTTP_FORM_NAME " option value can be string, table or file only", value->type()); |
| new String(key, String::L_TAINTED), | |
| "is %s, "HTTP_FORM_NAME" option value can be string, table or file only", value->type()); | |
| } | } |
| const char* pa_form2string_multipart(HashStringValue& form, Request& r, const char* boundary, size_t& post_size){ | const char* pa_form2string_multipart(HashStringValue& form, Request& r, const char* boundary, size_t& post_size){ |
| FormPart formpart; | FormPart formpart(&r, boundary); |
| formpart.r=&r; | |
| formpart.boundary=boundary; | |
| formpart.info=NULL; | |
| form.for_each<FormPart&>(form_value2part, formpart); | form.for_each<FormPart&>(form_value2part, formpart); |
| formpart.string << "--" << boundary << "--"; | *formpart.string << "--" << boundary << "--"; |
| post_size=formpart.string.length(); // very surprizing, but it calculates correct post_size even with binary files! | // @todo: return binary blocks here to save memory in pa_internal_file_read_http |
| return formpart.string.untaint_cstr(String::L_AS_IS); // without transcoding | return formpart.post(post_size); |
| } | } |
| static void find_headers_end(char* p, | static void find_headers_end(char* p, char*& headers_end_at, char*& raw_body) { |
| char*& headers_end_at, | |
| char*& raw_body) | |
| { | |
| raw_body=p; | raw_body=p; |
| // \n\n | // \n\n |
| // \r\n\r\n | // \r\n\r\n |
| Line 494 static void find_headers_end(char* p, | Line 491 static void find_headers_end(char* p, |
| headers_end_at=0; | headers_end_at=0; |
| } | } |
| // Set-Cookie: name=value; Domain=docs.foo.com; Path=/accounts; Expires=Wed, 13-Jan-2021 22:23:01 GMT; Secure; HttpOnly | |
| static ArrayString* parse_cookie(Request& r, const String& cookie) { | |
| char *current=pa_strdup(cookie.cstr()); | |
| const String* name=0; | |
| const String* value=&String::Empty; | |
| const String* expires=&String::Empty; | |
| const String* max_age=&String::Empty; | |
| const String* path=&String::Empty; | |
| const String* domain=&String::Empty; | |
| const String* httponly=&String::Empty; | |
| const String* secure=&String::Empty; | |
| bool first_pair=true; | |
| do { | |
| if(char *meaning=search_stop(current, ';')) | |
| if(char *attribute=search_stop(meaning, '=')) { | |
| const String* sname=new String(unescape_chars(attribute, strlen(attribute), &r.charsets.source(), true/*don't convert '"' to space*/), String::L_TAINTED); | |
| const String* smeaning=0; | |
| if(meaning) | |
| smeaning=new String(unescape_chars(meaning, strlen(meaning), &r.charsets.source(), true/*don't convert '"' to space*/), String::L_TAINTED); | |
| if(first_pair) { | |
| // name + value | |
| name=sname; | |
| value=smeaning; | |
| first_pair=false; | |
| } else { | |
| const String& slower=sname->change_case(r.charsets.source(), String::CC_LOWER); | |
| if(slower == "expires") | |
| expires=smeaning; | |
| else if(slower == "max-age") | |
| max_age=smeaning; | |
| else if(slower == "domain") | |
| domain=smeaning; | |
| else if(slower == "path") | |
| path=smeaning; | |
| else if(slower == "httponly") | |
| httponly=new String("1", String::L_CLEAN); | |
| else if(slower == "secure") | |
| secure=new String("1", String::L_CLEAN); | |
| else { | |
| // todo@ ? | |
| } | |
| } | |
| } | |
| } while(current); | |
| if(!name) | |
| return 0; | |
| ArrayString* result=new ArrayString(8); | |
| *result+=name; | |
| *result+=value; | |
| *result+=expires; | |
| *result+=max_age; | |
| *result+=domain; | |
| *result+=path; | |
| *result+=httponly; | |
| *result+=secure; | |
| return result; | |
| } | |
| Table* parse_cookies(Request& r, Table *cookies){ | |
| Table& result=*new Table(new Cookies_table_template_columns); | |
| for(Array_iterator<Table::element_type> i(*cookies); i.has_next(); ) | |
| if(ArrayString* row=parse_cookie(r, *i.next()->get(0))) | |
| result+=row; | |
| return &result; | |
| } | |
| void *tables_update(HashStringValue& tables, const String::Body name, const String& value){ | |
| Table *table; | |
| if(Value *valready=tables.get(name)) { | |
| // second+ appearence | |
| table=valready->get_table(); | |
| } else { | |
| // first appearence | |
| Table::columns_type columns=new ArrayString(1); | |
| *columns+=new String("value"); | |
| table=new Table(columns); | |
| tables.put(name, new VTable(table)); | |
| } | |
| // this string becomes next row | |
| ArrayString& row=*new ArrayString(1); | |
| row+=&value; | |
| *table+=&row; | |
| } | |
| /// @todo build .cookies field. use ^file.tables.SET-COOKIES.menu{ for now | /// @todo build .cookies field. use ^file.tables.SET-COOKIES.menu{ for now |
| File_read_http_result pa_internal_file_read_http(Request& r, | File_read_http_result pa_internal_file_read_http(Request& r, const String& file_spec, bool as_text, HashStringValue *options, bool transcode_text_result) { |
| const String& file_spec, | |
| bool as_text, | |
| HashStringValue *options, | |
| bool transcode_text_result) { | |
| File_read_http_result result; | File_read_http_result result; |
| char host[MAX_STRING]; | char host[MAX_STRING]; |
| const char *idna_host; | |
| const char* uri; | const char* uri; |
| short port; | short port=80; |
| const char* method="GET"; | const char* method="GET"; |
| bool method_is_get=true; | bool method_is_get=true; |
| HashStringValue* form=0; | HashStringValue* form=0; |
| const char* body_cstr=0; | |
| int timeout_secs=2; | int timeout_secs=2; |
| bool fail_on_status_ne_200=true; | bool fail_on_status_ne_200=true; |
| bool omit_post_charset=false; | bool omit_post_charset=false; |
| Value* vheaders=0; | Value* vheaders=0; |
| Value* vcookies=0; | Value* vcookies=0; |
| Value* vbody=0; | Value* vbody=0; |
| Charset *asked_remote_charset=0; | Charset* asked_remote_charset=0; |
| Charset* real_remote_charset=0; | |
| const char* user_cstr=0; | const char* user_cstr=0; |
| const char* password_cstr=0; | const char* password_cstr=0; |
| const char* encode=0; | const char* encode=0; |
| Line 558 File_read_http_result pa_internal_file_r | Line 646 File_read_http_result pa_internal_file_r |
| omit_post_charset=vomit_post_charset->as_bool(); | omit_post_charset=vomit_post_charset->as_bool(); |
| } | } |
| if(Value* vcharset_name=options->get(PA_CHARSET_NAME)) { | if(Value* vcharset_name=options->get(PA_CHARSET_NAME)) { |
| asked_remote_charset=&charsets.get(vcharset_name->as_string(). | asked_remote_charset=&charsets.get(vcharset_name->as_string().change_case(r.charsets.source(), String::CC_UPPER)); |
| change_case(r.charsets.source(), String::CC_UPPER)); | } |
| if(Value* vresponse_charset_name=options->get(PA_RESPONSE_CHARSET_NAME)) { | |
| valid_options++; | |
| real_remote_charset=&charsets.get(vresponse_charset_name->as_string().change_case(r.charsets.source(), String::CC_UPPER)); | |
| } | } |
| if(Value* vuser=options->get(HTTP_USER)) { | if(Value* vuser=options->get(HTTP_USER)) { |
| valid_options++; | valid_options++; |
| Line 571 File_read_http_result pa_internal_file_r | Line 662 File_read_http_result pa_internal_file_r |
| } | } |
| if(valid_options!=options->count()) | if(valid_options!=options->count()) |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, 0, CALLED_WITH_INVALID_OPTION); |
| 0, | |
| "invalid option passed"); | |
| } | } |
| if(!asked_remote_charset) // defaulting to $request:charset | if(!asked_remote_charset) // defaulting to $request:charset |
| asked_remote_charset=&(r.charsets).source(); | asked_remote_charset=&(r.charsets).source(); |
| if(encode){ | if(encode){ |
| if(method_is_get) | if(method_is_get) |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, 0, "you can not use $." HTTP_FORM_ENCTYPE_NAME " option with method GET"); |
| 0, | |
| "you can not use $."HTTP_FORM_ENCTYPE_NAME" option with method GET"); | |
| multipart=strcasecmp(encode, HTTP_CONTENT_TYPE_MULTIPART_FORMDATA)==0; | multipart=strcasecmp(encode, HTTP_CONTENT_TYPE_MULTIPART_FORMDATA)==0; |
| if(!multipart && strcasecmp(encode, HTTP_CONTENT_TYPE_FORM_URLENCODED)!=0) | if(!multipart && strcasecmp(encode, HTTP_CONTENT_TYPE_FORM_URLENCODED)!=0) |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, 0, "$." HTTP_FORM_ENCTYPE_NAME " option value can be " HTTP_CONTENT_TYPE_FORM_URLENCODED " or " HTTP_CONTENT_TYPE_MULTIPART_FORMDATA " only"); |
| 0, | |
| "$."HTTP_FORM_ENCTYPE_NAME" option value can be "HTTP_CONTENT_TYPE_FORM_URLENCODED" or "HTTP_CONTENT_TYPE_MULTIPART_FORMDATA" only"); | |
| } | } |
| if(vbody){ | if(vbody){ |
| if(method_is_get) | if(method_is_get) |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, 0, "you can not use $." HTTP_BODY_NAME " option with method GET"); |
| 0, | |
| "you can not use $."HTTP_BODY_NAME" option with method GET"); | |
| if(form) | if(form) |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, 0, "you can not use options $." HTTP_BODY_NAME " and $." HTTP_FORM_NAME " together"); |
| 0, | |
| "you can not use options $."HTTP_BODY_NAME" and $."HTTP_FORM_NAME" together"); | |
| } | } |
| //preparing request | //preparing request |
| String& connect_string=*new String(file_spec); | String& connect_string=*new String(file_spec); |
| String request_head_and_body; | const char* request; |
| size_t request_size; | |
| { | { |
| // influence URLencoding of tainted pieces to String::L_URI lang | // influence URLencoding of tainted pieces to String::L_URI lang |
| Temp_client_charset temp(r.charsets, *asked_remote_charset); | Temp_client_charset temp(r.charsets, *asked_remote_charset); |
| const char* connect_string_cstr=connect_string.untaint_cstr(String::L_URI, 0, &(r.charsets)); | const char* connect_string_cstr=connect_string.untaint_and_transcode_cstr(String::L_URI, &(r.charsets)); |
| const char* current=connect_string_cstr; | const char* current=connect_string_cstr; |
| if(strncmp(current, "http://", 7)!=0) | if(strncmp(current, "http://", 7)!=0) |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, &connect_string, "does not start with http://"); //never |
| &connect_string, | |
| "does not start with http://"); //never | |
| current+=7; | current+=7; |
| strncpy(host, current, sizeof(host)-1); host[sizeof(host)-1]=0; | strncpy(host, current, sizeof(host)-1); host[sizeof(host)-1]=0; |
| char* host_uri=lsplit(host, '/'); | char* host_uri=lsplit(host, '/'); |
| uri=host_uri?current+(host_uri-1-host):"/"; | uri=host_uri?current+(host_uri-1-host):"/"; |
| char* port_cstr=lsplit(host, ':'); | char* port_cstr=lsplit(host, ':'); |
| char* error_pos=0; | |
| port=port_cstr?(short)strtol(port_cstr, &error_pos, 0):80; | if (port_cstr){ |
| char* error_pos=0; | |
| port=(short)strtol(port_cstr, &error_pos, 10); | |
| if(port==0 || *error_pos) | |
| throw Exception(PARSER_RUNTIME, &connect_string, "invalid port number '%s'", port_cstr); | |
| } | |
| idna_host=pa_idna_encode(host, r.charsets.source()); | |
| // making request head | // making request head |
| String head; | String head; |
| Line 634 File_read_http_result pa_internal_file_r | Line 721 File_read_http_result pa_internal_file_r |
| if(method_is_get && form) | if(method_is_get && form) |
| head << (strchr(uri, '?')!=0?"&":"?") << pa_form2string(*form, r.charsets); | head << (strchr(uri, '?')!=0?"&":"?") << pa_form2string(*form, r.charsets); |
| head <<" HTTP/1.0" CRLF "host: "<< host << CRLF; | head <<" HTTP/1.0" CRLF "Host: "<< idna_host; |
| if (port != 80) | |
| head << ":" << port_cstr; | |
| head << CRLF; | |
| char* boundary=0; | char* boundary= multipart ? get_uuid_boundary() : 0; |
| if(multipart){ | String user_headers; |
| uuid uuid=get_uuid(); | bool user_agent_specified=false; |
| const int boundary_bufsize=10+32+1/*for zero-teminator*/+1/*for faulty snprintfs*/; | bool content_type_specified=false; |
| boundary=new(PointerFreeGC) char[boundary_bufsize]; | bool content_type_url_encoded=false; |
| snprintf(boundary, boundary_bufsize, | if(vheaders && !vheaders->is_string()) { // allow empty |
| "----------%08X%04X%04X%02X%02X%02X%02X%02X%02X%02X%02X", | if(HashStringValue *headers=vheaders->get_hash()) { |
| uuid.time_low, uuid.time_mid, uuid.time_hi_and_version, | Http_pass_header_info info={ |
| uuid.clock_seq >> 8, uuid.clock_seq & 0xFF, | &(r.charsets), |
| uuid.node[0], uuid.node[1], uuid.node[2], | &user_headers, |
| uuid.node[3], uuid.node[4], uuid.node[5]); | &user_agent_specified, |
| } | &content_type_specified, |
| &content_type_url_encoded}; | |
| headers->for_each<Http_pass_header_info*>(http_pass_header, &info); | |
| } else | |
| throw Exception(PARSER_RUNTIME, 0, "headers param must be hash"); | |
| }; | |
| const char* request_body=0; | |
| size_t post_size=0; | size_t post_size=0; |
| if(form && !method_is_get) { | if(form && !method_is_get) { |
| head << HTTP_CONTENT_TYPE ": " << (multipart ? HTTP_CONTENT_TYPE_MULTIPART_FORMDATA : HTTP_CONTENT_TYPE_FORM_URLENCODED); | head << "Content-Type: " << (multipart ? HTTP_CONTENT_TYPE_MULTIPART_FORMDATA : HTTP_CONTENT_TYPE_FORM_URLENCODED); |
| if(!omit_post_charset) | if(!omit_post_charset) |
| head << "; charset=" << asked_remote_charset->NAME_CSTR(); | head << "; charset=" << asked_remote_charset->NAME_CSTR(); |
| if(multipart) { | if(multipart) { |
| head << "; boundary=" << boundary; | head << "; boundary=" << boundary; |
| body_cstr=pa_form2string_multipart(*form, r/*charsets & mime_type needed*/, boundary, post_size/*correct post_size returned here*/); | request_body=pa_form2string_multipart(*form, r/*charsets & mime_type needed*/, boundary, post_size/*correct post_size returned here*/); |
| } else { | } else { |
| body_cstr=pa_form2string(*form, r.charsets); | request_body=pa_form2string(*form, r.charsets); |
| post_size=strlen(body_cstr); | post_size=strlen(request_body); |
| } | } |
| head << CRLF; | head << CRLF; |
| } else if (vbody) { | } else if(vbody) { |
| // transcode tainted pieces and then URI-encode them | // $.body was specified |
| body_cstr=vbody->as_string().untaint_cstr(String::L_AS_IS, 0, &(r.charsets)); | if(content_type_url_encoded){ |
| // transcode + url-encode | |
| // now transcode is needed only if own content-type was specified _and_ clean chars with code>127 are in the body | request_body=vbody->as_string().untaint_and_transcode_cstr(String::L_URI, &(r.charsets)); |
| // @todo: I don't like the current behaviour | } else { |
| body_cstr=Charset::transcode( | // content-type != application/x-www-form-urlencoded -> transcode only, don't url-encode! |
| String::C(body_cstr, strlen(body_cstr)), | const String &sbody=vbody->as_string(); |
| r.charsets.source(), | request_body=Charset::transcode(String::C(sbody.cstr(), sbody.length()), r.charsets.source(), *asked_remote_charset).str; |
| *asked_remote_charset | } |
| ); | post_size=strlen(request_body); |
| post_size=strlen(body_cstr); | |
| } | } |
| // http://www.ietf.org/rfc/rfc2617.txt | // http://www.ietf.org/rfc/rfc2617.txt |
| if(const String* authorization_field_value=basic_authorization_field(user_cstr, password_cstr)) | if(const String* authorization_field_value=basic_authorization_field(user_cstr, password_cstr)) |
| head<<"authorization: "<<*authorization_field_value<<CRLF; | head << "Authorization: " << *authorization_field_value << CRLF; |
| head << user_headers; | |
| bool user_agent_specified=false; | |
| bool content_type_specified=false; | |
| if(vheaders && !vheaders->is_string()) { // allow empty | |
| if(HashStringValue *headers=vheaders->get_hash()) { | |
| Http_pass_header_info info={&(r.charsets), &head, false}; | |
| headers->for_each<Http_pass_header_info*>(http_pass_header, &info); | |
| user_agent_specified=info.user_agent_specified; | |
| content_type_specified=info.content_type_specified; | |
| } else | |
| throw Exception(PARSER_RUNTIME, | |
| &connect_string, | |
| "headers param must be hash"); | |
| }; | |
| if(!user_agent_specified) // defaulting | if(!user_agent_specified) // defaulting |
| head << HTTP_USER_AGENT ": " DEFAULT_USER_AGENT CRLF; | head << "User-Agent: " DEFAULT_USER_AGENT CRLF; |
| if(form && !method_is_get && content_type_specified) // POST + form + content-type was specified | if(form && !method_is_get && content_type_specified) // POST + form + content-type was specified |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, 0, "$.content-type can't be specified with method POST"); |
| &connect_string, | |
| "$.content-type can't be specified with method POST"); | |
| if(vcookies && !vcookies->is_string()){ // allow empty | if(vcookies && !vcookies->is_string()){ // allow empty |
| if(HashStringValue* cookies=vcookies->get_hash()) { | if(HashStringValue* cookies=vcookies->get_hash()) { |
| head << "cookie: "; | head << "Cookie: "; |
| Http_pass_header_info info={&(r.charsets), &head, false}; | Http_pass_header_info info={&(r.charsets), &head, 0, 0, 0}; |
| cookies->for_each<Http_pass_header_info*>(http_pass_cookie, &info); | cookies->for_each<Http_pass_header_info*>(http_pass_cookie, &info); |
| head << CRLF; | head << CRLF; |
| } else | } else |
| throw Exception(PARSER_RUNTIME, | throw Exception(PARSER_RUNTIME, 0, "cookies param must be hash"); |
| &connect_string, | |
| "cookies param must be hash"); | |
| } | } |
| if(body_cstr) | if(request_body) |
| head << "content-length: " << format(post_size, "%u") << CRLF; | head << "Content-Length: " << format(post_size, "%u") << CRLF; |
| // head + end of header | head << CRLF; |
| request_head_and_body << head.untaint_cstr(String::L_AS_IS, 0, &(r.charsets)) << CRLF; | |
| const char *request_head=head.untaint_and_transcode_cstr(String::L_URI, &(r.charsets)); | |
| // body | if(request_body){ |
| if(body_cstr) | size_t head_size = strlen(request_head); |
| request_head_and_body << body_cstr; | request_size=post_size + head_size; |
| char *ptr=(char *)pa_malloc_atomic(request_size); | |
| memcpy(ptr, request_head, head_size); | |
| memcpy(ptr+head_size, request_body, post_size); | |
| request=ptr; | |
| } else { | |
| request_size=strlen(request_head); | |
| request=request_head; | |
| } | |
| } | } |
| const char* request_cstr=request_head_and_body.cstr(); | char* response_str; |
| size_t request_size=strlen(request_cstr); | |
| if(multipart) | |
| request_size=file_untaint(request_cstr, request_size); | |
| char* response; | |
| size_t response_size; | size_t response_size; |
| // sending request | // sending request |
| int status_code=http_request(response, response_size, | int status_code=http_request(response_str, response_size, idna_host, port, request, request_size, timeout_secs, fail_on_status_ne_200); |
| host, port, request_cstr, request_size, | |
| timeout_secs, fail_on_status_ne_200); | |
| // processing results | // processing results |
| char* raw_body; size_t raw_body_size; | char* raw_body; size_t raw_body_size; |
| char* headers_end_at; | char* headers_end_at; |
| find_headers_end(response, | find_headers_end(response_str, headers_end_at, raw_body); |
| headers_end_at, | raw_body_size=response_size-(raw_body-response_str); |
| raw_body); | |
| raw_body_size=response_size-(raw_body-response); | |
| result.headers=new HashStringValue; | result.headers=new HashStringValue; |
| VHash* vtables=new VHash; | VHash* vtables=new VHash; |
| result.headers->put(HTTP_TABLES_NAME, vtables); | result.headers->put("tables", vtables); |
| Charset* real_remote_charset=0; // undetected, yet | |
| ResponseHeaders response; | |
| if(headers_end_at) { | if(headers_end_at) { |
| *headers_end_at=0; | *headers_end_at=0; |
| const String header_block(String::C(response, headers_end_at-response), String::L_TAINTED); | const String header_block(String::C(response_str, headers_end_at-response_str), String::L_TAINTED); |
| ArrayString aheaders; | ArrayString aheaders; |
| HashStringValue& tables=vtables->hash(); | |
| size_t pos_after=0; | size_t pos_after=0; |
| header_block.split(aheaders, pos_after, "\n"); | header_block.split(aheaders, pos_after, "\n"); |
| // processing headers | |
| size_t aheaders_count=aheaders.count(); | |
| for(size_t i=1; i<aheaders_count; i++) { | |
| const String& line=*aheaders.get(i); | |
| size_t pos=line.pos(':'); | |
| if(pos==STRING_NOT_FOUND || pos<1) | |
| throw Exception("http.response", | |
| &connect_string, | |
| "bad response from host - bad header \"%s\"", line.cstr()); | |
| const String::Body HEADER_NAME=line.mid(0, pos).change_case(r.charsets.source(), String::CC_UPPER); | |
| const String& HEADER_VALUE=line.mid(pos+1, line.length()).trim(String::TRIM_BOTH, " \t\r"); | |
| if(as_text && HEADER_NAME==HTTP_CONTENT_TYPE_UPPER) | |
| real_remote_charset=detect_charset(HEADER_VALUE.cstr()); | |
| // tables | |
| { | |
| Value *valready=(Value *)tables.get(HEADER_NAME); | |
| bool existed=valready!=0; | |
| Table *table; | |
| if(existed) { | |
| // second+ appearence | |
| table=valready->get_table(); | |
| } else { | |
| // first appearence | |
| Table::columns_type columns=new ArrayString(1); | |
| *columns+=new String("value"); | |
| table=new Table(columns); | |
| } | |
| // this string becomes next row | |
| ArrayString& row=*new ArrayString(1); | |
| row+=&HEADER_VALUE; | |
| *table+=&row; | |
| // not existed before? add it | |
| if(!existed) | |
| tables.put(HEADER_NAME, new VTable(table)); | |
| } | |
| result.headers->put(HEADER_NAME, new VString(HEADER_VALUE)); | Array_iterator<const String*> i(aheaders); |
| i.next(); // skipping status | |
| for(;i.has_next();){ | |
| const char *line=i.next()->cstr(); | |
| if(!response.add_header(line)) | |
| throw Exception("http.response", &connect_string, "bad response from host - bad header \"%s\"", line); | |
| } | } |
| } | } |
| if(as_text && raw_body_size>=3 && strncmp(raw_body, "\xEF\xBB\xBF", 3)==0){ | if (!real_remote_charset && !response.content_type.is_empty()) |
| // skip UTF-8 signature (BOM code) | real_remote_charset= detect_charset(response.content_type.cstr()); |
| raw_body+=3; | |
| raw_body_size-=3; | if(as_text) |
| real_remote_charset=charsets.checkBOM(raw_body, raw_body_size, real_remote_charset); | |
| if (!real_remote_charset) | |
| real_remote_charset=asked_remote_charset; // never null | |
| for(Array_iterator<ResponseHeaders::Header> i(response.headers); i.has_next(); ){ | |
| ResponseHeaders::Header header=i.next(); | |
| header.transcode(*real_remote_charset, r.charsets.source()); | |
| String &header_value=*new String(header.value, String::L_TAINTED); | |
| tables_update(vtables->hash(), header.name, header_value); | |
| result.headers->put(header.name, new VString(header_value)); | |
| } | } |
| // filling $.cookies | |
| if(Value *vcookies=vtables->hash().get("SET-COOKIE")) | |
| result.headers->put(HTTP_COOKIES_NAME, new VTable(parse_cookies(r, vcookies->get_table()))); | |
| // output response | // output response |
| String::C real_body=String::C(raw_body, raw_body_size); | String::C real_body=String::C(raw_body, raw_body_size); |
| if(as_text && transcode_text_result && raw_body_size) { // raw_body_size must be checked because transcode returns CONST string in case length==0, which contradicts hacking few lines below | if(as_text && transcode_text_result && raw_body_size) { // raw_body_size must be checked because transcode returns CONST string in case length==0, which contradicts hacking few lines below |
| // defaulting to used-asked charset [it's never empty!] | |
| if(!real_remote_charset) | |
| real_remote_charset=asked_remote_charset; | |
| real_body=Charset::transcode(real_body, *real_remote_charset, r.charsets.source()); | real_body=Charset::transcode(real_body, *real_remote_charset, r.charsets.source()); |
| } | } |
| result.str=const_cast<char *>(real_body.str); // hacking a little | result.str=const_cast<char *>(real_body.str); // hacking a little |