|
|
| version 1.211, 2005/11/22 11:31:58 | version 1.244, 2009/05/13 07:35:41 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: commonly functions. | Parser: commonly functions. |
| Copyright(c) 2001-2005 ArtLebedev Group (http://www.artlebedev.com) | Copyright(c) 2001-2009 ArtLebedev Group (http://www.artlebedev.com) |
| Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) | Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) |
| * BASE64 part | * BASE64 part |
| Line 32 static const char * const IDENT_COMMON_C | Line 32 static const char * const IDENT_COMMON_C |
| #include "pa_exception.h" | #include "pa_exception.h" |
| #include "pa_hash.h" | #include "pa_hash.h" |
| #include "pa_globals.h" | #include "pa_globals.h" |
| #include "pa_request_charsets.h" | |
| #include "pa_charsets.h" | #include "pa_charsets.h" |
| #include "pa_http.h" | |
| #define PA_HTTP | #include "pa_request_charsets.h" |
| #include "pcre.h" | |
| #ifdef PA_HTTP | #include "pa_request.h" |
| #include "pa_vstring.h" | |
| #include "pa_vint.h" | |
| #include "pa_vhash.h" | |
| #include "pa_vtable.h" | |
| #include "pa_socks.h" | |
| #ifdef CYGWIN | |
| #define _GNU_H_WINDOWS32_SOCKETS | |
| // for PASCAL | |
| #include <windows.h> | |
| // SOCKET | |
| typedef u_int SOCKET; | |
| int PASCAL closesocket(SOCKET); | |
| #else | |
| # if defined(WIN32) | |
| # include <windows.h> | |
| # else | |
| # define closesocket close | |
| # endif | |
| #endif | |
| #else | |
| # if defined(WIN32) | |
| # include <windows.h> | |
| # endif | |
| #endif | |
| // some maybe-undefined constants | // some maybe-undefined constants |
| Line 86 int PASCAL closesocket(SOCKET); | Line 57 int PASCAL closesocket(SOCKET); |
| # endif | # endif |
| #endif | #endif |
| # ifndef INADDR_NONE | |
| # define INADDR_NONE ((ulong) -1) | |
| # endif | |
| // defines for globals | // defines for globals |
| #define FILE_STATUS_NAME "status" | #define FILE_STATUS_NAME "status" |
| Line 98 int PASCAL closesocket(SOCKET); | Line 65 int PASCAL closesocket(SOCKET); |
| const String file_status_name(FILE_STATUS_NAME); | const String file_status_name(FILE_STATUS_NAME); |
| // defines | |
| #define HTTP_METHOD_NAME "method" | |
| #define HTTP_FORM_NAME "form" | |
| #define HTTP_BODY_NAME "body" | |
| #define HTTP_TIMEOUT_NAME "timeout" | |
| #define HTTP_HEADERS_NAME "headers" | |
| #define HTTP_ANY_STATUS_NAME "any-status" | |
| #define HTTP_CHARSET_NAME "charset" | |
| #define HTTP_TABLES_NAME "tables" | |
| #define HTTP_USER "user" | |
| #define HTTP_PASSWORD "password" | |
| // defines | |
| #define DEFAULT_USER_AGENT "parser3" | |
| // functions | // functions |
| void fix_line_breaks(char *str, size_t& length) { | void fix_line_breaks(char *str, size_t& length) { |
| Line 144 void fix_line_breaks(char *str, size_t& | Line 94 void fix_line_breaks(char *str, size_t& |
| } | } |
| char* file_read_text(Request_charsets& charsets, | char* file_read_text(Request_charsets& charsets, |
| const String& file_spec, | const String& file_spec, |
| bool fail_on_read_problem, | bool fail_on_read_problem, |
| HashStringValue* params/*, HashStringValue* * out_fields*/) { | HashStringValue* params, |
| bool transcode_result) { | |
| File_read_result file= | |
| file_read(charsets, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result); | |
| return file.success?file.str:0; | |
| } | |
| char* file_load_text(Request& r, | |
| const String& file_spec, | |
| bool fail_on_read_problem, | |
| HashStringValue* params, | |
| bool transcode_result) { | |
| File_read_result file= | File_read_result file= |
| file_read(charsets, file_spec, true, params, fail_on_read_problem); | file_load(r, file_spec, true, params, fail_on_read_problem, 0, 0, 0, transcode_result); |
| return file.success?file.str:0; | return file.success?file.str:0; |
| } | } |
| /// these options were handled but not checked elsewhere, now check them | /// these options were handled but not checked elsewhere, now check them |
| static int get_valid_file_options_count(HashStringValue& options) | int pa_get_valid_file_options_count(HashStringValue& options) { |
| { | |
| int result=0; | int result=0; |
| if(options.get(PA_SQL_LIMIT_NAME)) | if(options.get(PA_SQL_LIMIT_NAME)) |
| result++; | result++; |
| Line 164 static int get_valid_file_options_count( | Line 124 static int get_valid_file_options_count( |
| result++; | result++; |
| if(options.get(PA_COLUMN_ENCLOSER_NAME)) | if(options.get(PA_COLUMN_ENCLOSER_NAME)) |
| result++; | result++; |
| if(options.get(PA_CHARSET_NAME)) | |
| result++; | |
| return result; | return result; |
| } | } |
| //http request stuff | |
| #ifdef PA_HTTP | |
| #undef CRLF | |
| #define CRLF "\r\n" | |
| static bool set_addr(struct sockaddr_in *addr, const char* host, const short port){ | |
| memset(addr, 0, sizeof(*addr)); | |
| addr->sin_family=AF_INET; | |
| addr->sin_port=htons(port); | |
| if(host) { | |
| ulong packed_ip=inet_addr(host); | |
| if(packed_ip!=INADDR_NONE) | |
| memcpy(&addr->sin_addr, &packed_ip, sizeof(packed_ip)); | |
| else { | |
| struct hostent *hostIP=gethostbyname(host); | |
| if(hostIP) | |
| memcpy(&addr->sin_addr, hostIP->h_addr, hostIP->h_length); | |
| else | |
| return false; | |
| } | |
| } else | |
| addr->sin_addr.s_addr=INADDR_ANY; | |
| return true; | |
| } | |
| size_t guess_content_length(char* buf) { | |
| char* ptr; | |
| if((ptr=strstr(buf, "Content-Length:"))) // Apache | |
| goto found; | |
| if((ptr=strstr(buf, "content-length:"))) // Parser 3 | |
| goto found; | |
| if((ptr=strstr(buf, "Content-length:"))) // maybe 1 | |
| goto found; | |
| if((ptr=strstr(buf, "CONTENT-LENGTH:"))) // maybe 2 | |
| goto found; | |
| return 0; | |
| found: | |
| char *error_pos; | |
| size_t result=(size_t)strtol(ptr+15/*strlen("CONTENT-LENGTH:")*/, &error_pos, 0); | |
| const size_t reasonable_initial_max=0x400*0x400*10 /*10M*/; | |
| if(result>reasonable_initial_max) // sanity check | |
| return reasonable_initial_max; | |
| return 0;//result; | |
| } | |
| static int http_read_response(char*& response, size_t& response_size, int sock, bool fail_on_status_ne_200) { | |
| int result=0; | |
| // fetching some to local buffer, guessing on possible content-length | |
| response_size=0x400*20; // initial size if content-length could not be determined | |
| const size_t preview_size=0x400*20; | |
| char preview_buf[preview_size+1/*terminator*/]; // 20K buffer to preview headers | |
| ssize_t received_size=recv(sock, preview_buf, preview_size, 0); | |
| if(received_size==0) | |
| goto done; | |
| if(received_size<0) { | |
| if(int no=pa_socks_errno()) | |
| throw Exception("http.timeout", | |
| 0, | |
| "error receiving response header: %s (%d)", pa_socks_strerr(no), no); | |
| goto done; | |
| } | |
| // detecting response_size | |
| { | |
| preview_buf[received_size]=0; // terminator | |
| if(size_t content_length=guess_content_length(preview_buf)) | |
| response_size=preview_size+content_length; // a little more than needed, will adjust response_size by actual received size later | |
| } | |
| // [gcc is happier this way, see goto above] | |
| { | |
| // allocating initial buf | |
| response=(char*)pa_malloc_atomic(response_size+1/*terminator*/); // just setting memory block type | |
| char* ptr=response; | |
| size_t todo_size=response_size; | |
| // coping part of already received body | |
| memcpy(ptr, preview_buf, received_size); | |
| ptr+=received_size; | |
| todo_size-=received_size; | |
| // 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? | |
| // 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 | |
| // when recv gets more than we expected | |
| // a) we know that the content-length guess was wrong | |
| // 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 | |
| while(true) { | |
| received_size=recv(sock, ptr, todo_size+1/*there is always a place for terminator*/, 0); | |
| if(received_size==0) { | |
| response_size-=todo_size; // in case we received less than expected, cut down the reported size | |
| break; | |
| } | |
| if(received_size<0) { | |
| if(int no=pa_socks_errno()) | |
| throw Exception("http.timeout", | |
| 0, | |
| "error receiving response body: %s (%d)", pa_socks_strerr(no), no); | |
| break; | |
| } | |
| // they've touched the terminator? | |
| if((size_t)received_size>todo_size) | |
| { | |
| // that means that our guessed response_size was not big enough | |
| const size_t grow_chunk_size=0x400*0x400; // 1M | |
| response_size+=grow_chunk_size; | |
| size_t ptr_offset=ptr-response; | |
| response=(char*)pa_realloc(response, response_size+1/*terminator*/); | |
| ptr=response+ptr_offset; | |
| todo_size+=grow_chunk_size; | |
| } | |
| // can't do this before realloc: we need <todo_size check | |
| ptr+=received_size; | |
| todo_size-=received_size; | |
| char* EOLat=0; | |
| if(!result && (EOLat=strstr(response, "\n"))) { // checking status in first response | |
| const String status_line(pa_strdup(response, EOLat-response)); | |
| ArrayString astatus; | |
| size_t pos_after=0; | |
| status_line.split(astatus, pos_after, " "); | |
| const String& status_code=*astatus.get(astatus.count()>1?1:0); | |
| result=status_code.as_int(); | |
| if(fail_on_status_ne_200 && result!=200) | |
| throw Exception("http.status", | |
| &status_code, | |
| "invalid HTTP response status"); | |
| } | |
| } | |
| } | |
| done: | |
| if(result) | |
| { | |
| response[response_size]=0; | |
| return result; | |
| } | |
| else | |
| throw Exception("http.response", | |
| 0, | |
| "bad response from host - no status found (size=%u)", response_size); | |
| } | |
| /* ********************** request *************************** */ | |
| #if defined(SIGALRM) && defined(HAVE_SIGSETJMP) && defined(HAVE_SIGLONGJMP) | |
| # define PA_USE_ALARM | |
| #endif | |
| #ifdef PA_USE_ALARM | |
| static sigjmp_buf timeout_env; | |
| static void timeout_handler(int /*sig*/){ | |
| siglongjmp(timeout_env, 1); | |
| } | |
| #endif | |
| static int http_request(char*& response, size_t& response_size, | |
| const char* host, short port, | |
| const char* request, | |
| int timeout_secs, | |
| bool fail_on_status_ne_200) { | |
| if(!host) | |
| throw Exception("http.host", | |
| 0, | |
| "zero hostname"); //never | |
| volatile // to prevent makeing it register variable, because it will be clobbered by longjmp [thanks gcc warning] | |
| int sock=-1; | |
| #ifdef PA_USE_ALARM | |
| signal(SIGALRM, timeout_handler); | |
| #endif | |
| #ifdef PA_USE_ALARM | |
| if(sigsetjmp(timeout_env, 1)) { | |
| // stupid gcc [2.95.4] generated bad code | |
| // which failed to handle sigsetjmp+throw: crashed inside of pre-throw code. | |
| // rewritten simplier [athough duplicating closesocket code] | |
| if(sock>=0) | |
| closesocket(sock); | |
| throw Exception("http.timeout", | |
| 0, | |
| "timeout occured while retrieving document"); | |
| return 0; // never | |
| } else { | |
| alarm(timeout_secs); | |
| #endif | |
| try { | |
| int result; | |
| struct sockaddr_in dest; | |
| if(!set_addr(&dest, host, port)) | |
| throw Exception("http.host", | |
| 0, | |
| "can not resolve hostname \"%s\"", host); | |
| if((sock=socket(AF_INET, SOCK_STREAM, IPPROTO_TCP/*0*/))<0) { | |
| int no=pa_socks_errno(); | |
| throw Exception("http.connect", | |
| 0, | |
| "can not make socket: %s (%d)", pa_socks_strerr(no), no); | |
| } | |
| // To enable SO_DONTLINGER (that is, disable SO_LINGER) | |
| // l_onoff should be set to zero and setsockopt should be called | |
| linger dont_linger={0,0}; | |
| setsockopt(sock, SOL_SOCKET, SO_LINGER, (const char *)&dont_linger, sizeof(dont_linger)); | |
| #ifdef WIN32 | |
| // SO_*TIMEO can be defined in .h but not implemlemented in protocol, | |
| // failing subsequently with Option not supported by protocol (99) message | |
| // could not suppress that, so leaving this only for win32 | |
| int timeout_ms=timeout_secs*1000; | |
| setsockopt(sock, SOL_SOCKET, SO_SNDTIMEO, (const char*)&timeout_ms, sizeof(timeout_ms)); | |
| setsockopt(sock, SOL_SOCKET, SO_RCVTIMEO, (const char*)&timeout_ms, sizeof(timeout_ms)); | |
| #endif | |
| if(connect(sock, (struct sockaddr *)&dest, sizeof(dest))) { | |
| int no=pa_socks_errno(); | |
| throw Exception("http.connect", | |
| 0, | |
| "can not connect to host \"%s\": %s (%d)", host, pa_socks_strerr(no), no); | |
| } | |
| size_t request_size=strlen(request); | |
| if(send(sock, request, request_size, 0)!=(ssize_t)request_size) { | |
| int no=pa_socks_errno(); | |
| throw Exception("http.timeout", | |
| 0, | |
| "error sending request: %s (%d)", pa_socks_strerr(no), no); | |
| } | |
| result=http_read_response(response, response_size, sock, fail_on_status_ne_200); | |
| closesocket(sock); | |
| #ifdef PA_USE_ALARM | |
| alarm(0); | |
| #endif | |
| return result; | |
| } catch(...) { | |
| #ifdef PA_USE_ALARM | |
| alarm(0); | |
| #endif | |
| if(sock>=0) | |
| closesocket(sock); | |
| rethrow; | |
| } | |
| #ifdef PA_USE_ALARM | |
| } | |
| #endif | |
| } | |
| #ifndef DOXYGEN | |
| struct Http_pass_header_info { | |
| Request_charsets* charsets; | |
| String* request; | |
| bool user_agent_specified; | |
| }; | |
| #endif | |
| static void http_pass_header(HashStringValue::key_type key, | |
| HashStringValue::value_type value, | |
| Http_pass_header_info *info) { | |
| *info->request <<key<<": " | |
| << attributed_meaning_to_string(*value, String::L_HTTP_HEADER, false) | |
| << CRLF; | |
| if(String(key, String::L_TAINTED).change_case(info->charsets->source(), String::CC_UPPER)=="USER-AGENT") | |
| info->user_agent_specified=true; | |
| } | |
| static Charset* detect_charset(Charset& source_charset, const String& content_type_value) { | |
| const String::Body CONTENT_TYPE_VALUE= | |
| content_type_value.change_case(source_charset, String::CC_UPPER); | |
| // content-type: xxx/xxx; source_charset=WE-NEED-THIS | |
| // content-type: xxx/xxx; source_charset="WE-NEED-THIS" | |
| // content-type: xxx/xxx; source_charset="WE-NEED-THIS"; | |
| size_t before_charseteq_pos=CONTENT_TYPE_VALUE.pos("CHARSET="); | |
| if(before_charseteq_pos!=STRING_NOT_FOUND) { | |
| size_t charset_begin=before_charseteq_pos+8/*CHARSET="*/; | |
| size_t open_quote_pos=CONTENT_TYPE_VALUE.pos('"', charset_begin); | |
| bool quoted=open_quote_pos==charset_begin; | |
| if(quoted) | |
| charset_begin++; // skip opening '"' | |
| size_t charset_end=CONTENT_TYPE_VALUE.length(); | |
| if(quoted) { | |
| size_t close_quote_pos=CONTENT_TYPE_VALUE.pos('"', charset_begin); | |
| if(close_quote_pos!=STRING_NOT_FOUND) | |
| charset_end=close_quote_pos; | |
| } else { | |
| size_t delim_pos=CONTENT_TYPE_VALUE.pos(';', charset_begin); | |
| if(delim_pos!=STRING_NOT_FOUND) | |
| charset_end=delim_pos; | |
| } | |
| const String::Body CHARSET_NAME_BODY= | |
| CONTENT_TYPE_VALUE.mid(charset_begin, charset_end); | |
| return &charsets.get(CHARSET_NAME_BODY); | |
| } | |
| return 0; | |
| } | |
| static const String* basic_authorization_field(const char* user, const char* pass) { | |
| if(!user&& !pass) | |
| return 0; | |
| String combined; | |
| if(user) | |
| combined<<user; | |
| combined<<":"; | |
| if(pass) | |
| combined<<pass; | |
| String* result=new String("Basic "); *result<<pa_base64_encode(combined.cstr(), combined.length()); | |
| return result; | |
| } | |
| static void form_string_value2string( | |
| HashStringValue::key_type key, | |
| const String& value, | |
| String& result) | |
| { | |
| result << String(key, String::L_TAINTED) << "="; | |
| result.append(value, String::L_URI, true); | |
| result<< "&"; | |
| } | |
| #ifndef DOXYGEN | |
| struct Form_table_value2string_info { | |
| HashStringValue::key_type key; | |
| String& result; | |
| Form_table_value2string_info(HashStringValue::key_type akey, String& aresult): | |
| key(akey), result(aresult) {} | |
| }; | |
| #endif | |
| static void form_table_value2string(Table::element_type row, Form_table_value2string_info* info) { | |
| form_string_value2string(info->key, *row->get(0), info->result); | |
| } | |
| static void form_value2string( | |
| HashStringValue::key_type key, | |
| HashStringValue::value_type value, | |
| String* result) | |
| { | |
| if(const String* svalue=value->get_string()) | |
| form_string_value2string(key, *svalue, *result); | |
| else if(Table* tvalue=value->get_table()) { | |
| Form_table_value2string_info info(key, *result); | |
| tvalue->for_each(form_table_value2string, &info); | |
| } else | |
| throw Exception(0, | |
| new String(key, String::L_TAINTED), | |
| "is %s, "HTTP_FORM_NAME" option value must either string or table", value->type()); | |
| } | |
| static const char* form2string(HashStringValue& form) { | |
| String string; | |
| form.for_each(form_value2string, &string); | |
| return string.cstr(String::L_UNSPECIFIED); | |
| } | |
| #ifndef DOXYGEN | |
| struct File_read_http_result { | |
| char *str; size_t length; | |
| HashStringValue* headers; | |
| }; | |
| #endif | |
| static void find_headers_end(char* p, | |
| char*& headers_end_at, | |
| char*& raw_body) | |
| { | |
| raw_body=p; | |
| // \n\n | |
| // \r\n\r\n | |
| while((p=strchr(p, '\n'))) { | |
| headers_end_at=++p; // \n>.< | |
| if(*p=='\r') // \r\n>\r?<\n | |
| p++; | |
| if(*p=='\n') { // \r\n\r>\n?< | |
| raw_body=p+1; | |
| return; | |
| } | |
| } | |
| headers_end_at=0; | |
| } | |
| /// @todo build .cookies field. use ^file.tables.SET-COOKIES.menu{ for now | |
| static File_read_http_result file_read_http(Request_charsets& charsets, | |
| const String& file_spec, | |
| bool as_text, | |
| HashStringValue *options=0) { | |
| File_read_http_result result; | |
| char host[MAX_STRING]; | |
| const char* uri; | |
| short port; | |
| const char* method="GET"; bool method_is_get; | |
| HashStringValue* form=0; | |
| const char* body_cstr=0; | |
| int timeout_secs=2; | |
| bool fail_on_status_ne_200=true; | |
| Value* vheaders=0; | |
| Charset *asked_remote_charset=0; | |
| const char* user_cstr=0; | |
| const char* password_cstr=0; | |
| if(options) { | |
| int valid_options=get_valid_file_options_count(*options); | |
| if(Value* vmethod=options->get(HTTP_METHOD_NAME)) { | |
| valid_options++; | |
| method=vmethod->as_string().cstr(); | |
| } | |
| if(Value* vform=options->get(HTTP_FORM_NAME)) { | |
| valid_options++; | |
| form=vform->get_hash(); | |
| } | |
| if(Value* vbody=options->get(HTTP_BODY_NAME)) { | |
| valid_options++; | |
| body_cstr=vbody->as_string().cstr(String::L_UNSPECIFIED); | |
| } | |
| if(Value* vtimeout=options->get(HTTP_TIMEOUT_NAME)) { | |
| valid_options++; | |
| timeout_secs=vtimeout->as_int(); | |
| } | |
| if((vheaders=options->get(HTTP_HEADERS_NAME))) { | |
| valid_options++; | |
| } | |
| if(Value* vany_status=options->get(HTTP_ANY_STATUS_NAME)) { | |
| valid_options++; | |
| fail_on_status_ne_200=!vany_status->as_bool(); | |
| } | |
| if(Value* vcharset_name=options->get(HTTP_CHARSET_NAME)) { | |
| valid_options++; | |
| asked_remote_charset=&::charsets.get(vcharset_name->as_string(). | |
| change_case(charsets.source(), String::CC_UPPER)); | |
| } | |
| if(Value* vuser=options->get(HTTP_USER)) { | |
| valid_options++; | |
| user_cstr=vuser->as_string().cstr(); | |
| } | |
| if(Value* vpassword=options->get(HTTP_PASSWORD)) { | |
| valid_options++; | |
| password_cstr=vpassword->as_string().cstr(); | |
| } | |
| if(valid_options!=options->count()) | |
| throw Exception("parser.runtime", | |
| 0, | |
| "invalid option passed"); | |
| } | |
| if(!asked_remote_charset) // defaulting to $request:charset | |
| asked_remote_charset=&charsets.source(); | |
| method_is_get=strcmp(method, "GET")==0; | |
| if(method_is_get && body_cstr) | |
| throw Exception("parser.runtime", | |
| 0, | |
| "you can not use $."HTTP_BODY_NAME" option with method GET"); | |
| //preparing request | |
| String& connect_string=*new String; | |
| // not in ^sql{... L_SQL ...} spirit, but closer to ^file::load one | |
| connect_string.append(file_spec, String::L_URI); // tainted pieces -> URI pieces | |
| String request_head_and_body; | |
| { | |
| // influence URLencoding of tainted pieces to String::L_URI lang | |
| Temp_client_charset temp(charsets, *asked_remote_charset); | |
| const char* connect_string_cstr=connect_string.cstr(String::L_UNSPECIFIED); | |
| const char* current=connect_string_cstr; | |
| if(strncmp(current, "http://", 7)!=0) | |
| throw Exception(0, | |
| &connect_string, | |
| "does not start with http://"); //never | |
| current+=7; | |
| strncpy(host, current, sizeof(host)-1); host[sizeof(host)-1]=0; | |
| char* host_uri=lsplit(host, '/'); | |
| uri=host_uri?current+(host_uri-1-host):"/"; | |
| char* port_cstr=lsplit(host, ':'); | |
| char* error_pos=0; | |
| port=port_cstr?(short)strtol(port_cstr, &error_pos, 0):80; | |
| if(strchr(uri, '?') && form) | |
| throw Exception("parser.runtime", | |
| 0, | |
| "use either uri with ?params or $."HTTP_FORM_NAME" option"); | |
| //making request head | |
| String head; | |
| head << method; | |
| head << " " << uri; | |
| if(form) | |
| if(method_is_get) | |
| head << "?" << form2string(*form); | |
| head <<" HTTP/1.0" CRLF | |
| "host: "<< host << CRLF; | |
| if(form && !method_is_get) { | |
| head << "content-type: application/x-www-form-urlencoded" CRLF; | |
| body_cstr = form2string(*form); | |
| } | |
| // http://www.ietf.org/rfc/rfc2617.txt | |
| if(const String* authorization_field_value=basic_authorization_field(user_cstr, password_cstr)) | |
| head<<"authorization: "<<*authorization_field_value<<CRLF; | |
| bool user_agent_specified=false; | |
| if(vheaders && !vheaders->is_string()) { // allow empty | |
| if(HashStringValue *headers=vheaders->get_hash()) { | |
| Http_pass_header_info info={&charsets, &head, false}; | |
| headers->for_each(http_pass_header, &info); | |
| user_agent_specified=info.user_agent_specified; | |
| } else | |
| throw Exception("parser.runtime", | |
| &connect_string, | |
| "headers param must be hash"); | |
| }; | |
| if(!user_agent_specified) // defaulting | |
| head << "user-agent: " DEFAULT_USER_AGENT CRLF; | |
| if(body_cstr) { | |
| // recode those pieces which are not in String::L_URI lang | |
| // [those violating HTTP standard, but widly used] | |
| body_cstr=Charset::transcode( | |
| String::C(body_cstr, strlen(body_cstr)), | |
| charsets.source(), | |
| *asked_remote_charset); | |
| head << "content-length: " << format(strlen(body_cstr), "%u") << CRLF; | |
| } | |
| const char* head_cstr=head.cstr(String::L_UNSPECIFIED); | |
| // recode those pieces which are not in String::L_URI lang | |
| // [those violating HTTP standard, but widly used] | |
| head_cstr=Charset::transcode( | |
| String::C(head_cstr, strlen(head_cstr)), | |
| charsets.source(), | |
| *asked_remote_charset); | |
| // head + end of header | |
| request_head_and_body << head_cstr << CRLF; | |
| // body | |
| if(body_cstr) | |
| request_head_and_body << body_cstr; | |
| } | |
| //sending request | |
| char* response; | |
| size_t response_size; | |
| int status_code=http_request(response, response_size, | |
| host, port, request_head_and_body.cstr(), | |
| timeout_secs, fail_on_status_ne_200); | |
| //processing results | |
| char* raw_body; size_t raw_body_size; | |
| char* headers_end_at; | |
| find_headers_end(response, | |
| headers_end_at, | |
| raw_body); | |
| raw_body_size=response_size-(raw_body-response); | |
| result.headers=new HashStringValue; | |
| VHash* vtables=new VHash; | |
| result.headers->put(HTTP_TABLES_NAME, vtables); | |
| Charset* real_remote_charset=0; // undetected, yet | |
| if(headers_end_at) { | |
| *headers_end_at=0; | |
| const String header_block(String::C(response, headers_end_at-response), true); | |
| ArrayString aheaders; | |
| HashStringValue& tables=vtables->hash(); | |
| size_t pos_after=0; | |
| 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(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=="CONTENT-TYPE") | |
| real_remote_charset=detect_charset(charsets.source(), header_value); | |
| // 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)); | |
| } | |
| } | |
| // output response | |
| String::C real_body=String::C(raw_body, raw_body_size); | |
| if(as_text && 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, charsets.source()); | |
| } | |
| result.str=const_cast<char *>(real_body.str); // hacking a little | |
| result.length=real_body.length; | |
| result.headers->put(file_status_name, new VInt(status_code)); | |
| return result; | |
| } | |
| #endif | |
| #ifndef DOXYGEN | #ifndef DOXYGEN |
| struct File_read_action_info { | struct File_read_action_info { |
| char **data; size_t *data_size; | char **data; size_t *data_size; |
| Line 808 struct File_read_action_info { | Line 136 struct File_read_action_info { |
| }; | }; |
| #endif | #endif |
| static void file_read_action( | static void file_read_action( |
| struct stat& finfo, | struct stat& finfo, |
| int f, | int f, |
| const String& file_spec, const char* /*fname*/, bool as_text, | const String& file_spec, const char* /*fname*/, bool as_text, |
| void *context) { | void *context) { |
| File_read_action_info& info=*static_cast<File_read_action_info *>(context); | File_read_action_info& info=*static_cast<File_read_action_info *>(context); |
| size_t to_read_size=info.count; | size_t to_read_size=info.count; |
| if(!to_read_size) | if(!to_read_size) |
| Line 826 static void file_read_action( | Line 154 static void file_read_action( |
| *info.data_size=(size_t)read(f, *info.data, to_read_size); | *info.data_size=(size_t)read(f, *info.data, to_read_size); |
| if(ssize_t(*info.data_size)<0 || *info.data_size>to_read_size) | if(ssize_t(*info.data_size)<0 || *info.data_size>to_read_size) |
| throw Exception(0, | throw Exception("file.read", |
| &file_spec, | &file_spec, |
| "read failed: actually read %u bytes count not in [0..%u] valid range", | "read failed: actually read %u bytes count not in [0..%u] valid range", |
| *info.data_size, to_read_size); | *info.data_size, to_read_size); |
| Line 838 static void file_read_action( | Line 166 static void file_read_action( |
| return; | return; |
| } | } |
| } | } |
| File_read_result file_read(Request_charsets& charsets, const String& file_spec, | File_read_result file_read(Request_charsets& charsets, const String& file_spec, |
| bool as_text, HashStringValue *params, | bool as_text, HashStringValue *params, |
| bool fail_on_read_problem, | bool fail_on_read_problem, |
| char* buf, size_t offset, size_t count) { | char* buf, size_t offset, size_t count, bool transcode_text_result) { |
| File_read_result result={false, 0, 0, 0}; | |
| if(params){ | |
| int valid_options=pa_get_valid_file_options_count(*params); | |
| if(valid_options!=params->count()) | |
| throw Exception(PARSER_RUNTIME, | |
| 0, | |
| "invalid option passed"); | |
| } | |
| File_read_action_info info={&result.str, &result.length, buf, offset, count}; | |
| result.success=file_read_action_under_lock(file_spec, | |
| "read", file_read_action, &info, | |
| as_text, fail_on_read_problem); | |
| if(as_text){ | |
| if(result.success){ | |
| if(result.length>=3 && strncmp(result.str, "\xEF\xBB\xBF", 3)==0){ | |
| // skip UTF-8 signature (BOM code) | |
| result.str+=3; | |
| result.length-=3; | |
| } | |
| if(result.length && transcode_text_result && params){ // must be checked because transcode returns CONST string in case length==0, which contradicts hacking few lines below | |
| if(Value* vcharset_name=params->get(PA_CHARSET_NAME)){ | |
| Charset asked_charset=::charsets.get(vcharset_name->as_string(). | |
| change_case(charsets.source(), String::CC_UPPER)); | |
| String::C body=String::C(result.str, result.length); | |
| body=Charset::transcode(body, asked_charset, charsets.source()); | |
| result.str=const_cast<char*>(body.str); // hacking a little | |
| result.length=body.length; | |
| } | |
| } | |
| } | |
| if(result.length) | |
| fix_line_breaks(result.str, result.length); | |
| } | |
| return result; | |
| } | |
| File_read_result file_load(Request& r, const String& file_spec, | |
| bool as_text, HashStringValue *params, | |
| bool fail_on_read_problem, | |
| char* buf, size_t offset, size_t count, bool transcode_text_result) { | |
| File_read_result result={false, 0, 0, 0}; | File_read_result result={false, 0, 0, 0}; |
| #ifdef PA_HTTP | |
| if(file_spec.starts_with("http://")) { | if(file_spec.starts_with("http://")) { |
| if(offset || count) | if(offset || count) |
| throw Exception("parser.runtime", | throw Exception(PARSER_RUNTIME, |
| 0, | 0, |
| "offset and load options are not supported for HTTP:// file load"); | "offset and load options are not supported for HTTP:// file load"); |
| // fail on read problem | // fail on read problem |
| File_read_http_result http=file_read_http(charsets, file_spec, as_text, params); | File_read_http_result http=pa_internal_file_read_http(r, file_spec, as_text, params, transcode_text_result); |
| result.success=true; | result.success=true; |
| result.str=http.str; | result.str=http.str; |
| result.length=http.length; | result.length=http.length; |
| result.headers=http.headers; | result.headers=http.headers; |
| } else { | } else |
| #endif | result= |
| if(params) { | file_read(r.charsets, file_spec, as_text, params, fail_on_read_problem, buf, offset, count, transcode_text_result); |
| int valid_options=get_valid_file_options_count(*params); | |
| if(valid_options!=params->count()) | |
| throw Exception("parser.runtime", | |
| 0, | |
| "invalid option passed"); | |
| } | |
| File_read_action_info info={&result.str, &result.length, | |
| buf, offset, count}; | |
| result.success=file_read_action_under_lock(file_spec, | |
| "read", file_read_action, &info, | |
| as_text, fail_on_read_problem); | |
| #ifdef PA_HTTP | |
| } | |
| #endif | |
| if(result.success && as_text) { | |
| // UTF-8 signature: EF BB BF | |
| if(result.length>=3) { | |
| char *in=(char *)result.str; | |
| if(strncmp(in, "\xEF\xBB\xBF", 3)==0) { | |
| result.str=in+3; result.length-=3;// skip prefix | |
| } | |
| } | |
| fix_line_breaks((char *)(result.str), result.length); | |
| } | |
| return result; | return result; |
| } | } |
| Line 894 File_read_result file_read(Request_chars | Line 243 File_read_result file_read(Request_chars |
| void check_safe_mode(struct stat finfo, const String& file_spec, const char* fname) { | void check_safe_mode(struct stat finfo, const String& file_spec, const char* fname) { |
| if(finfo.st_uid/*foreign?*/!=geteuid() | if(finfo.st_uid/*foreign?*/!=geteuid() |
| && finfo.st_gid/*foreign?*/!=getegid()) | && finfo.st_gid/*foreign?*/!=getegid()) |
| throw Exception("parser.runtime", | throw Exception(PARSER_RUNTIME, |
| &file_spec, | &file_spec, |
| "parser is in safe mode: " | "parser is in safe mode: " |
| "reading files of foreign group and user disabled " | "reading files of foreign group and user disabled " |
| Line 923 bool file_read_action_under_lock(const S | Line 272 bool file_read_action_under_lock(const S |
| // they delay update till open, so we would receive "!^test[" string | // they delay update till open, so we would receive "!^test[" string |
| // if would do stat, next open. | // if would do stat, next open. |
| // later: it seems, even this does not help sometimes | // later: it seems, even this does not help sometimes |
| if((f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) { | if((f=open(fname, O_RDONLY|(as_text?_O_TEXT:_O_BINARY)))>=0) { |
| try { | try { |
| if(pa_lock_shared_blocking(f)!=0) | if(pa_lock_shared_blocking(f)!=0) |
| throw Exception("file.lock", | throw Exception("file.lock", |
| Line 952 bool file_read_action_under_lock(const S | Line 301 bool file_read_action_under_lock(const S |
| pa_unlock(f);close(f); | pa_unlock(f);close(f); |
| return true; | return true; |
| } else { | } else { |
| if(fail_on_read_problem) | if(fail_on_read_problem) |
| throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0, | throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0, |
| &file_spec, | &file_spec, |
| Line 973 void create_dir_for_file(const String& f | Line 322 void create_dir_for_file(const String& f |
| bool file_write_action_under_lock( | bool file_write_action_under_lock( |
| const String& file_spec, | const String& file_spec, |
| const char* action_name, File_write_action action, void *context, | const char* action_name, |
| File_write_action action, | |
| void *context, | |
| bool as_text, | bool as_text, |
| bool do_append, | bool do_append, |
| bool do_block, | bool do_block, |
| Line 1033 static void file_write_action(int f, voi | Line 384 static void file_write_action(int f, voi |
| if(info.length) { | if(info.length) { |
| int written=write(f, info.str, info.length); | int written=write(f, info.str, info.length); |
| if(written<0) | if(written<0) |
| throw Exception(0, | throw Exception("file.access", |
| 0, | 0, |
| "write failed: %s (%d)", strerror(errno), errno); | "write failed: %s (%d)", strerror(errno), errno); |
| } | } |
| Line 1044 void file_write( | Line 395 void file_write( |
| bool as_text, | bool as_text, |
| bool do_append) { | bool do_append) { |
| File_write_action_info info={data, size}; | File_write_action_info info={data, size}; |
| file_write_action_under_lock( | file_write_action_under_lock( |
| file_spec, | file_spec, |
| "write", file_write_action, &info, | "write", |
| file_write_action, | |
| &info, | |
| as_text, | as_text, |
| do_append); | do_append); |
| } | } |
| Line 1059 static void rmdir(const String& file_spe | Line 413 static void rmdir(const String& file_spe |
| rmdir(file_spec.mid(0, pos_after-1/* / */).cstr(String::L_FILE_SPEC)); | rmdir(file_spec.mid(0, pos_after-1/* / */).cstr(String::L_FILE_SPEC)); |
| } | } |
| bool file_delete(const String& file_spec, bool fail_on_problem) { | bool file_delete(const String& file_spec, bool fail_on_problem) { |
| const char* fname=file_spec.cstr(String::L_FILE_SPEC); | const char* fname=file_spec.cstr(String::L_FILE_SPEC); |
| if(unlink(fname)!=0) | if(unlink(fname)!=0) |
| Line 1073 bool file_delete(const String& file_spec | Line 428 bool file_delete(const String& file_spec |
| rmdir(file_spec, 1); | rmdir(file_spec, 1); |
| return true; | return true; |
| } | } |
| void file_move(const String& old_spec, const String& new_spec) { | void file_move(const String& old_spec, const String& new_spec) { |
| const char* old_spec_cstr=old_spec.cstr(String::L_FILE_SPEC); | const char* old_spec_cstr=old_spec.cstr(String::L_FILE_SPEC); |
| const char* new_spec_cstr=new_spec.cstr(String::L_FILE_SPEC); | const char* new_spec_cstr=new_spec.cstr(String::L_FILE_SPEC); |
| Line 1121 static bool entry_readable(const String& | Line 477 static bool entry_readable(const String& |
| } | } |
| return false; | return false; |
| } | } |
| bool file_readable(const String& file_spec) { | |
| bool file_exist(const String& file_spec) { | |
| return entry_readable(file_spec, false); | return entry_readable(file_spec, false); |
| } | } |
| bool dir_readable(const String& file_spec) { | |
| bool dir_exists(const String& file_spec) { | |
| return entry_readable(file_spec, true); | return entry_readable(file_spec, true); |
| } | } |
| const String* file_readable(const String& path, const String& name) { | |
| const String* file_exist(const String& path, const String& name) { | |
| String& result=*new String(path); | String& result=*new String(path); |
| result << "/"; | result << "/"; |
| result << name; | result << name; |
| return file_readable(result)?&result:0; | return file_exist(result)?&result:0; |
| } | } |
| bool file_executable(const String& file_spec) { | bool file_executable(const String& file_spec) { |
| return access(file_spec.cstr(String::L_FILE_SPEC), X_OK)==0; | return access(file_spec.cstr(String::L_FILE_SPEC), X_OK)==0; |
| } | } |
| bool file_stat(const String& file_spec, | bool file_stat(const String& file_spec, |
| size_t& rsize, | size_t& rsize, |
| time_t& ratime, | time_t& ratime, |
| time_t& rmtime, | time_t& rmtime, |
| time_t& rctime, | time_t& rctime, |
| bool fail_on_read_problem) { | bool fail_on_read_problem) { |
| const char* fname=file_spec.cstr(String::L_FILE_SPEC); | const char* fname=file_spec.cstr(String::L_FILE_SPEC); |
| struct stat finfo; | struct stat finfo; |
| if(stat(fname, &finfo)!=0) | if(stat(fname, &finfo)!=0) |
| Line 1161 bool file_stat(const String& file_spec, | Line 521 bool file_stat(const String& file_spec, |
| } | } |
| char* getrow(char* *row_ref, char delim) { | char* getrow(char* *row_ref, char delim) { |
| char* result=*row_ref; | char* result=*row_ref; |
| if(result) { | if(result) { |
| *row_ref=strchr(result, delim); | *row_ref=strchr(result, delim); |
| if(*row_ref) | if(*row_ref) |
| *((*row_ref)++)=0; | *((*row_ref)++)=0; |
| else if(!*result) | else if(!*result) |
| return 0; | return 0; |
| } | } |
| return result; | return result; |
| } | } |
| char* lsplit(char* string, char delim) { | char* lsplit(char* string, char delim) { |
| if(string) { | if(string) { |
| char* v=strchr(string, delim); | char* v=strchr(string, delim); |
| if(v) { | if(v) { |
| *v=0; | *v=0; |
| return v+1; | return v+1; |
| } | } |
| } | } |
| return 0; | return 0; |
| } | } |
| char* lsplit(char* *string_ref, char delim) { | char* lsplit(char* *string_ref, char delim) { |
| char* result=*string_ref; | char* result=*string_ref; |
| char* next=lsplit(*string_ref, delim); | char* next=lsplit(*string_ref, delim); |
| *string_ref=next; | *string_ref=next; |
| return result; | return result; |
| } | } |
| char* rsplit(char* string, char delim) { | char* rsplit(char* string, char delim) { |
| if(string) { | if(string) { |
| char* v=strrchr(string, delim); | char* v=strrchr(string, delim); |
| if(v) { | if(v) { |
| *v=0; | *v=0; |
| return v+1; | return v+1; |
| } | } |
| } | } |
| return NULL; | return NULL; |
| } | } |
| /// @todo less stupid type detection | |
| // format: %[flags][width][.precision]type http://msdn.microsoft.com/ru-ru/library/56e442dc(en-us,VS.80).aspx | |
| // flags: '-', '+', ' ', '#', '0' http://msdn.microsoft.com/ru-ru/library/8aky45ct(en-us,VS.80).aspx | |
| // width, precision: non negative decimal number | |
| enum FormatType { | |
| FormatInvalid, | |
| FormatInt, | |
| FormatUInt, | |
| FormatDouble | |
| }; | |
| FormatType format_type(char* fmt){ | |
| enum FormatState { | |
| Percent, | |
| Flags, | |
| Width, | |
| Precision, | |
| Done | |
| } state=Percent; | |
| FormatType result=FormatInvalid; | |
| char* pos=fmt; | |
| while(char c=*(pos++)){ | |
| switch(state){ | |
| case Percent: | |
| if(c=='%'){ | |
| state=Flags; | |
| } else { | |
| return FormatInvalid; // 1st char must be '%' only | |
| } | |
| break; | |
| case Flags: | |
| if(strchr("-+ #0", c)!=0){ | |
| break; | |
| } | |
| // go to the next step | |
| case Width: | |
| if(c=='.'){ | |
| state=Precision; | |
| break; | |
| } | |
| // go to the next step | |
| case Precision: | |
| if(c>='0' && c<='9'){ | |
| if(state == Flags) state=Width; // no more flags | |
| break; | |
| } else if(c=='d' || c=='i'){ | |
| result=FormatInt; | |
| } else if(strchr("feEgG", c)!=0){ | |
| result=FormatDouble; | |
| } else if(strchr("uoxX", c)!=0){ | |
| result=FormatUInt; | |
| } else { | |
| return FormatInvalid; // invalid char | |
| } | |
| state=Done; | |
| break; | |
| case Done: | |
| return FormatInvalid; // no chars allowed after 'type' | |
| } | |
| } | |
| return result; | |
| } | |
| const char* format(double value, char* fmt) { | const char* format(double value, char* fmt) { |
| char local_buf[MAX_NUMBER]; | char local_buf[MAX_NUMBER]; |
| size_t size; | int size=-1; |
| if(fmt) | if(fmt && strlen(fmt)){ |
| if(strpbrk(fmt, "diouxX")) | switch(format_type(fmt)){ |
| if(strpbrk(fmt, "ouxX")) | case FormatDouble: |
| size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value); | size=snprintf(local_buf, sizeof(local_buf), fmt, value); |
| else | break; |
| case FormatInt: | |
| size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value); | size=snprintf(local_buf, sizeof(local_buf), fmt, (int)value); |
| else | break; |
| size=snprintf(local_buf, sizeof(local_buf), fmt, value); | case FormatUInt: |
| else | size=snprintf(local_buf, sizeof(local_buf), fmt, (uint)value); |
| size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value); | break; |
| case FormatInvalid: | |
| return pa_strdup(local_buf, size); | throw Exception(PARSER_RUNTIME, |
| 0, | |
| "Incorrect format string '%s' was specified.", fmt); | |
| } | |
| } else | |
| size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value); | |
| if(size < 0 || size >= MAX_NUMBER-1){ // on win32 we manually reduce max size while printing | |
| throw Exception(PARSER_RUNTIME, | |
| 0, | |
| "Error occure white executing snprintf with format string '%s'.", fmt); | |
| } | |
| return pa_strdup(local_buf, (size_t)size); | |
| } | } |
| size_t stdout_write(const void *buf, size_t size) { | size_t stdout_write(const void *buf, size_t size) { |
| Line 1237 size_t stdout_write(const void *buf, siz | Line 675 size_t stdout_write(const void *buf, siz |
| #endif | #endif |
| } | } |
| char* unescape_chars(const char* cp, int len) { | enum EscapeState { |
| char* s=new(PointerFreeGC) char[len + 1]; | EscapeRest, |
| enum EscapeState { | EscapeFirst, |
| EscapeRest, | EscapeSecond, |
| EscapeFirst, | EscapeUnicode |
| EscapeSecond | }; |
| } escapeState=EscapeRest; | |
| uchar escapedValue=0; | // @todo prescan for reduce required size (unescaped sting in 1 byte charset requires less memory usually) |
| char* unescape_chars(const char* cp, int len, Charset* charset, bool ignore_plus){ | |
| char* s=new(PointerFreeGC) char[len+1]; // must be enough (%uXXXX==6 bytes, max utf-8 char length==6 bytes) | |
| char* dst=s; | |
| EscapeState escapeState=EscapeRest; | |
| uint escapedValue=0; | |
| int srcPos=0; | int srcPos=0; |
| int dstPos=0; | short int jsCnt=0; |
| while(srcPos < len) { | while(srcPos<len){ |
| uchar ch=(uchar)cp[srcPos]; | uchar c=(uchar)cp[srcPos]; |
| switch(escapeState) { | if(c=='%'){ |
| case EscapeRest: | escapeState=EscapeFirst; |
| if(ch=='%') { | } else { |
| escapeState=EscapeFirst; | switch(escapeState) { |
| } else if(ch=='+') { | case EscapeRest: |
| s[dstPos++]=' '; | if(!ignore_plus && c=='+'){ |
| } else { | *dst++=' '; |
| s[dstPos++]=ch; | } else { |
| *dst++=c; | |
| } | |
| break; | |
| case EscapeFirst: | |
| if(charset && c=='u'){ | |
| // escaped unicode value: %u0430 | |
| jsCnt=0; | |
| escapedValue=0; | |
| escapeState=EscapeUnicode; | |
| } else { | |
| if(isxdigit(c)){ | |
| escapedValue=hex_value[c] << 4; | |
| escapeState=EscapeSecond; | |
| } else { | |
| *dst++=c; | |
| escapeState=EscapeRest; | |
| } | |
| } | |
| break; | |
| case EscapeSecond: | |
| if(isxdigit(c)){ | |
| escapedValue+=hex_value[c]; | |
| *dst++=(char)escapedValue; | |
| } | |
| escapeState=EscapeRest; | |
| break; | |
| case EscapeUnicode: | |
| if(isxdigit(c)){ | |
| escapedValue=(escapedValue << 4) + hex_value[c]; | |
| if(++jsCnt==4){ | |
| // transcode utf8 char to client charset (we can lost some chars here) | |
| charset->store_Char((XMLByte*&)dst, (XMLCh)escapedValue, '?'); | |
| escapeState=EscapeRest; | |
| } | |
| } else { | |
| // not full unicode value | |
| escapeState=EscapeRest; | |
| } | |
| break; | |
| } | } |
| break; | |
| case EscapeFirst: | |
| escapedValue=(uchar)(hex_value[ch] << 4); | |
| escapeState=EscapeSecond; | |
| break; | |
| case EscapeSecond: | |
| escapedValue +=hex_value[ch]; | |
| s[dstPos++]=escapedValue; | |
| escapeState=EscapeRest; | |
| break; | |
| } | } |
| srcPos++; | |
| srcPos++; | |
| } | } |
| s[dstPos]=0; | |
| *dst=0; // zero-termination | |
| return s; | return s; |
| } | } |
| Line 1292 void slashes_to_back_slashes(char* s) { | Line 766 void slashes_to_back_slashes(char* s) { |
| */ | */ |
| #endif | #endif |
| bool StrEqNc(const char* s1, const char* s2, bool strict) { | bool StrStartFromNC(const char* str, const char* substr, bool equal){ |
| while(true) { | while(true) { |
| if(!(*s1)) { | if(!(*substr)){ |
| if(!(*s2)) | if(!(*str)) |
| return true; | return true; |
| else | else |
| return !strict; | return !equal; |
| } else if(!(*s2)) | } |
| return !strict; | if(!(*str)) |
| if(isalpha((unsigned char)*s1)) { | return false; |
| if(tolower((unsigned char)*s1) !=tolower((unsigned char)*s2)) | if(isalpha((unsigned char)*str)) { |
| if(tolower((unsigned char)*str)!=tolower((unsigned char)*substr)) | |
| return false; | return false; |
| } else if((*s1) !=(*s2)) | } else if((*str) != (*substr)) |
| return false; | return false; |
| s1++; | str++; |
| s2++; | substr++; |
| } | |
| } | |
| size_t strpos(const char *str, const char *substr) { | |
| const char *p = strstr(str, substr); | |
| return (p==0)?STRING_NOT_FOUND:p-str; | |
| } | |
| // content-type: xxx; charset=WE-NEED-THIS | |
| // content-type: xxx; charset="WE-NEED-THIS" | |
| // content-type: xxx; charset="WE-NEED-THIS"; | |
| Charset* detect_charset(const char* content_type){ | |
| if(content_type){ | |
| size_t len=strlen(content_type); | |
| char* CONTENT_TYPE=new(PointerFreeGC) char[len+1]; | |
| memcpy(CONTENT_TYPE, content_type, len); | |
| for(char *p=CONTENT_TYPE; *p; p++) | |
| *p=(char)toupper((unsigned char)*p); | |
| if(const char* begin=strstr(CONTENT_TYPE, "CHARSET=")){ | |
| begin+=8; // skip "CHARSET=" | |
| char* end=0; | |
| if(*begin && (*begin=='"' || *begin =='\'')){ | |
| char quote=*begin; | |
| begin++; | |
| end=(char*)strchr(begin, quote); | |
| } | |
| if(!end) | |
| end=(char*)strchr(begin, ';'); | |
| if(end) | |
| *end=0; // terminator | |
| return &charsets.get(begin); | |
| } | |
| } | } |
| return 0; | |
| } | } |
| Charset* detect_charset(Charset& source_charset, const String& content_type){ | |
| const String& CONTENT_TYPE=content_type.change_case(source_charset, String::CC_UPPER); | |
| size_t begin=CONTENT_TYPE.pos("CHARSET="); | |
| if(begin!=STRING_NOT_FOUND) { | |
| begin+=8; // skip "CHARSET=" | |
| size_t end=STRING_NOT_FOUND; | |
| if(CONTENT_TYPE.pos('"', begin)==begin){ | |
| begin++; | |
| end=CONTENT_TYPE.pos('"', begin); | |
| } else if(CONTENT_TYPE.pos('\'', begin)==begin){ | |
| begin++; | |
| end=CONTENT_TYPE.pos('\'', begin); | |
| } | |
| if(end==STRING_NOT_FOUND) | |
| end=CONTENT_TYPE.pos(';', begin); | |
| if(end==STRING_NOT_FOUND) | |
| end=CONTENT_TYPE.length(); | |
| const String::Body NAME=CONTENT_TYPE.mid(begin, end); | |
| return &charsets.get(NAME); | |
| } | |
| return 0; | |
| } | |
| static bool isLeap(int year) { | static bool isLeap(int year) { |
| return !( | return !( |
| (year % 4) || ((year % 400) && !(year % 100)) | (year % 4) || ((year % 400) && !(year % 100)) |
| ); | ); |
| } | } |
| int getMonthDays(int year, int month) { | int getMonthDays(int year, int month) { |
| int monthDays[]={ | static int monthDays[]={ |
| 31, | 31, |
| isLeap(year) ? 29 : 28, | 28, |
| 31, | 31, |
| 30, | 30, |
| 31, | 31, |
| 30, | 30, |
| 31, | 31, |
| 31, | 31, |
| 30, | 30, |
| 31, | 31, |
| 30, | 30, |
| 31 | 31 |
| }; | }; |
| return monthDays[month]; | return (month == 1 /* january -- 0 */ && isLeap(year)) ? 29 : monthDays[month]; |
| } | } |
| void remove_crlf(char* start, char* end) { | int remove_crlf(char* start, char* end) { |
| for(char* p=start; p<end; p++) | char* from=start; |
| switch(*p) { | char* to=start; |
| case '\n': *p='|'; break; | bool skip=false; |
| case '\r': *p=' '; break; | while(from < end){ |
| switch(*from){ | |
| case '\n': | |
| case '\r': | |
| case '\t': | |
| case ' ': | |
| if(!skip){ | |
| *to=' '; | |
| to++; | |
| skip=true; | |
| } | |
| break; | |
| default: | |
| if(from != to) | |
| *to=*from; | |
| to++; | |
| skip=false; | |
| } | } |
| from++; | |
| } | |
| return to-start; | |
| } | } |
| Line 1393 int __vsnprintf(char* b, size_t s, const | Line 951 int __vsnprintf(char* b, size_t s, const |
| int __snprintf(char* b, size_t s, const char* f, ...) { | int __snprintf(char* b, size_t s, const char* f, ...) { |
| va_list l; | va_list l; |
| va_start(l, f); | va_start(l, f); |
| int r=__vsnprintf(b, s, f, l); | int r=__vsnprintf(b, s, f, l); |
| va_end(l); | va_end(l); |
| return r; | return r; |
| } | } |
| Line 1500 g_mime_utils_base64_encode_step (const u | Line 1058 g_mime_utils_base64_encode_step (const u |
| case 2: *saveout++ = *inptr++; | case 2: *saveout++ = *inptr++; |
| case 1: *saveout++ = *inptr++; | case 1: *saveout++ = *inptr++; |
| } | } |
| ((char *)save)[0] += inlen; | *(char *)save = *(char *)save+(char)inlen; |
| } | } |
| /*d(printf ("mode = %d\nc1 = %c\nc2 = %c\n", | /*d(printf ("mode = %d\nc1 = %c\nc2 = %c\n", |
| Line 1593 static unsigned char gmime_base64_rank[2 | Line 1151 static unsigned char gmime_base64_rank[2 |
| size_t | size_t |
| g_mime_utils_base64_decode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save) | g_mime_utils_base64_decode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save) |
| { | { |
| const register unsigned char *inptr; | const unsigned char *inptr; |
| register unsigned char *outptr; | unsigned char *outptr; |
| const unsigned char *inend; | const unsigned char *inend; |
| register guint32 saved; | int saved; |
| unsigned char c; | unsigned char c; |
| int i; | int i; |
| Line 1613 g_mime_utils_base64_decode_step (const u | Line 1171 g_mime_utils_base64_decode_step (const u |
| saved = (saved << 6) | c; | saved = (saved << 6) | c; |
| i++; | i++; |
| if (i == 4) { | if (i == 4) { |
| *outptr++ = saved >> 16; | *outptr++ = (unsigned char)(saved >> 16); |
| *outptr++ = saved >> 8; | *outptr++ = (unsigned char)(saved >> 8); |
| *outptr++ = saved; | *outptr++ = (unsigned char)(saved); |
| i = 0; | i = 0; |
| } | } |
| } | } |
| Line 1641 g_mime_utils_base64_decode_step (const u | Line 1199 g_mime_utils_base64_decode_step (const u |
| } | } |
| char* pa_base64_encode(const char *in, size_t in_size) | char* pa_base64_encode(const char *in, size_t in_size){ |
| { | |
| /* wont go to more than 2x size (overly conservative) */ | /* wont go to more than 2x size (overly conservative) */ |
| char* result=new(PointerFreeGC) char[in_size * 2 + 6]; | char* result=new(PointerFreeGC) char[in_size * 2 + 6]; |
| int state=0; | int state=0; |
| Line 1657 char* pa_base64_encode(const char *in, s | Line 1214 char* pa_base64_encode(const char *in, s |
| return result; | return result; |
| } | } |
| void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size) | |
| { | char* pa_base64_encode(const String& file_spec){ |
| unsigned char* base64=0; | |
| File_base64_action_info info={&base64}; | |
| file_read_action_under_lock(file_spec, | |
| "pa_base64_encode", file_base64_file_action, &info); | |
| return (char*)base64; | |
| } | |
| static void file_base64_file_action( | |
| struct stat& finfo, | |
| int f, | |
| const String&, const char* /*fname*/, bool, | |
| void *context) { | |
| if(finfo.st_size) { | |
| File_base64_action_info& info=*static_cast<File_base64_action_info *>(context); | |
| *info.base64=new(PointerFreeGC) unsigned char[finfo.st_size * 2 + 6]; | |
| unsigned char* base64 = *info.base64; | |
| int state=0; | |
| int save=0; | |
| int nCount; | |
| do { | |
| unsigned char buffer[FILE_BUFFER_SIZE]; | |
| nCount = file_block_read(f, buffer, sizeof(buffer)); | |
| if( nCount ){ | |
| size_t filled=g_mime_utils_base64_encode_step ((const unsigned char*)buffer, nCount, base64, &state, &save); | |
| base64+=filled; | |
| } | |
| } while(nCount > 0); | |
| g_mime_utils_base64_encode_close (0, 0, base64, &state, &save); | |
| } | |
| } | |
| void pa_base64_decode(const char *in, size_t in_size, char*& result, size_t& result_size) { | |
| /* wont go to more than had (overly conservative) */ | /* wont go to more than had (overly conservative) */ |
| result=new(PointerFreeGC) char[in_size+1/*terminator*/]; | result=new(PointerFreeGC) char[in_size+1/*terminator*/]; |
| int state=0; | int state=0; |
| Line 1669 void pa_base64_decode(const char *in, si | Line 1262 void pa_base64_decode(const char *in, si |
| assert(result_size <= in_size); | assert(result_size <= in_size); |
| result[result_size]=0; // for text files | result[result_size]=0; // for text files |
| } | } |
| int file_block_read(const int f, unsigned char* buffer, const size_t size){ | |
| int nCount = read(f, buffer, size); | |
| if (nCount < 0) | |
| throw Exception("file.read", | |
| 0, | |
| "read failed: %s (%d)", strerror(errno), errno); | |
| return nCount; | |
| } | |
| const unsigned long pa_crc32(const char *in, size_t in_size){ | |
| unsigned long crc32=0xFFFFFFFF; | |
| InitCrc32Table(); | |
| for(size_t i = 0; i<in_size; i++) | |
| CalcCrc32(in[i], crc32); | |
| return ~crc32; | |
| } | |
| const unsigned long pa_crc32(const String& file_spec){ | |
| unsigned long crc32=0xFFFFFFFF; | |
| file_read_action_under_lock(file_spec, "crc32", file_crc32_file_action, &crc32); | |
| return ~crc32; | |
| } | |
| static void file_crc32_file_action( | |
| struct stat& finfo, | |
| int f, | |
| const String&, const char* /*fname*/, bool, | |
| void *context) { | |
| unsigned long& crc32=*static_cast<unsigned long *>(context); | |
| if(finfo.st_size) { | |
| InitCrc32Table(); | |
| int nCount=0; | |
| do { | |
| unsigned char buffer[FILE_BUFFER_SIZE]; | |
| nCount = file_block_read(f, buffer, sizeof(buffer)); | |
| for(int i = 0; i < nCount; i++) CalcCrc32(buffer[i], crc32); | |
| } while(nCount > 0); | |
| } | |
| } | |