--- parser3/src/main/pa_common.C 2003/11/19 08:20:02 1.165 +++ parser3/src/main/pa_common.C 2004/03/05 10:32:17 1.177 @@ -1,11 +1,11 @@ /** @file Parser: commonly functions. - Copyright(c) 2001-2003 ArtLebedev Group (http://www.artlebedev.com) + Copyright(c) 2001-2004 ArtLebedev Group (http://www.artlebedev.com) Author: Alexandr Petrosian (http://paf.design.ru) */ -static const char* IDENT_COMMON_C="$Date: 2003/11/19 08:20:02 $"; +static const char * const IDENT_COMMON_C="$Date: 2004/03/05 10:32:17 $"; #include "pa_common.h" #include "pa_exception.h" @@ -64,6 +64,10 @@ int PASCAL closesocket(SOCKET); # endif #endif +# ifndef INADDR_NONE +# define INADDR_NONE ((ulong) -1) +# endif + // defines for globals #define FILE_STATUS_NAME "status" @@ -83,13 +87,6 @@ const String file_status_name(FILE_STATU // statics -static const String http_method_name(HTTP_METHOD_NAME); -static const String http_timeout_name(HTTP_TIMEOUT_NAME); -static const String http_headers_name(HTTP_HEADERS_NAME); -static const String http_any_status_name(HTTP_ANY_STATUS_NAME); -static const String http_charset_name(HTTP_CHARSET_NAME); -static const String http_tables_name(HTTP_TABLES_NAME); - // defines #define DEFAULT_USER_AGENT "parser3" @@ -142,7 +139,11 @@ static bool set_addr(struct sockaddr_in addr->sin_family=AF_INET; addr->sin_port=htons(port); if(host) { - if(struct hostent *hostIP=gethostbyname(host)) + ulong packed_ip=inet_addr(host); + struct hostent *hostIP = packed_ip==INADDR_NONE? + gethostbyname(host) + : gethostbyaddr((char *)&packed_ip, sizeof(packed_ip), AF_INET); + if(hostIP) memcpy(&addr->sin_addr, hostIP->h_addr, hostIP->h_length); else return false; @@ -151,17 +152,23 @@ static bool set_addr(struct sockaddr_in return true; } -static int http_read_response(String& response, 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){ + response=(char*)pa_malloc_atomic(1/*terminator*/); // setting memory block type + response[(response_size=0)]=0; int result=0; - size_t EOLat=0; + char* EOLat=0; while(true) { - char *buf=new(PointerFreeGC) char[MAX_STRING]; - ssize_t size=recv(sock, buf, MAX_STRING, 0); - if(size<=0) + char buf[MAX_STRING*10]; + ssize_t received_size=recv(sock, buf, sizeof(buf), 0); + if(received_size<=0) break; - response.append_strdup(buf, size, String::L_TAINTED); - if(!result && (EOLat=response.pos(CRLF, 2))!=STRING_NOT_FOUND) { // checking status in first response - const String& status_line=response.mid(0, (size_t)EOLat); + response=(char*)pa_realloc(response, response_size+received_size+1/*terminator*/); + memcpy(response+response_size, buf, received_size); + response_size+=received_size; + response[response_size]=0; + + if(!result && (EOLat=strstr(response, CRLF))) { // 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, " "); @@ -179,7 +186,7 @@ static int http_read_response(String& re else throw Exception("http.response", 0, - "bad response from host - no status found (size=%lu)", response.length()); + "bad response from host - no status found (size=%u)", response_size); } /* ********************** request *************************** */ @@ -195,10 +202,14 @@ static void timeout_handler(int sig){ } #endif -static int http_request(String& response, +static int http_request(char*& response, size_t& response_size, const char* host, int port, const char* request, - int timeout, + int +#ifdef PA_USE_ALARM + timeout +#endif + , bool fail_on_status_ne_200) { if(!host) throw Exception("http.host", @@ -246,7 +257,7 @@ static int http_request(String& response 0, "error sending request: %s (%d)", strerror(errno), errno); - result=http_read_response(response, sock, fail_on_status_ne_200); + result=http_read_response(response, response_size, sock, fail_on_status_ne_200); closesocket(sock); #ifdef PA_USE_ALARM alarm(0); @@ -326,7 +337,8 @@ struct File_read_http_result { /// @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, - HashStringValue *options=0) { + bool as_text, + HashStringValue *options=0) { File_read_http_result result; char host[MAX_STRING]; const char* uri; @@ -339,22 +351,22 @@ static File_read_http_result file_read_h if(options) { int valid_options=0; - if(Value* vmethod=options->get(http_method_name)) { + if(Value* vmethod=options->get(HTTP_METHOD_NAME)) { valid_options++; method=vmethod->as_string().cstr(); } - if(Value* vtimeout=options->get(http_timeout_name)) { + if(Value* vtimeout=options->get(HTTP_TIMEOUT_NAME)) { valid_options++; timeout=vtimeout->as_int(); } - if(vheaders=options->get(http_headers_name)) { + if((vheaders=options->get(HTTP_HEADERS_NAME))) { valid_options++; } - if(Value* vany_status=options->get(http_any_status_name)) { + 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)) { + 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)); @@ -394,7 +406,6 @@ static File_read_http_result file_read_h char* error_pos=0; port=port_cstr?strtol(port_cstr, &error_pos, 0):80; - //making request String request; if(method) @@ -406,7 +417,7 @@ static File_read_http_result file_read_h bool user_agent_specified=false; if(vheaders && !vheaders->is_string()) { // allow empty if(HashStringValue *headers=vheaders->get_hash()) { - Http_pass_header_info info={&charsets, &request}; + Http_pass_header_info info={&charsets, &request, false}; headers->for_each(http_pass_header, &info); user_agent_specified=info.user_agent_specified; } else @@ -428,25 +439,30 @@ static File_read_http_result file_read_h *asked_remote_charset); //sending request - String response; - int status_code=http_request(response, + char* response; + size_t response_size; + int status_code=http_request(response, response_size, host, port, request_cstr, timeout, fail_on_status_ne_200); //processing results - size_t pos=response.pos(CRLF CRLF, 4); - if(pos==STRING_NOT_FOUND || pos<1) { + char* raw_body; size_t raw_body_size; + char* headers_end_at=strstr(response, CRLF CRLF /*change '4' below along!*/); + if(headers_end_at) { + raw_body=headers_end_at+4; + raw_body_size=response_size-(raw_body-response); + } else throw Exception("http.response", &connect_string, "bad response from host - no headers found"); - } - const String& header_block=response.mid(0, pos); - const String& raw_body=response.mid(pos+4, response.length()); + + *headers_end_at=0; + const String header_block(response, headers_end_at-response, true); ArrayString aheaders; result.headers=new HashStringValue; VHash* vtables=new VHash; - result.headers->put(http_tables_name, vtables); + result.headers->put(HTTP_TABLES_NAME, vtables); HashStringValue& tables=vtables->hash(); size_t pos_after=0; @@ -457,7 +473,7 @@ static File_read_http_result file_read_h Charset* real_remote_charset=0; // undetected, yet for(size_t i=1; i(real_body.str); // hacking a little result.length=real_body.length; @@ -520,7 +534,7 @@ struct File_read_action_info { static void file_read_action( struct stat& finfo, int f, - const String& file_spec, const char* fname, bool as_text, + const String& file_spec, const char* /*fname*/, bool as_text, void *context) { File_read_action_info& info=*static_cast(context); if(size_t to_read_size=(size_t)finfo.st_size) { @@ -530,7 +544,7 @@ static void file_read_action( if(ssize_t(*info.data_size)<0 || *info.data_size>to_read_size) throw Exception(0, &file_spec, - "read failed: actually read %l bytes count not in [0..%lu] valid range", + "read failed: actually read %u bytes count not in [0..%u] valid range", *info.data_size, to_read_size); } else { // empty file if(as_text) { @@ -545,11 +559,11 @@ static void file_read_action( File_read_result file_read(Request_charsets& charsets, const String& file_spec, bool as_text, HashStringValue *params, bool fail_on_read_problem) { - File_read_result result={false}; + File_read_result result={false, 0, 0, 0}; #ifdef PA_HTTP if(file_spec.starts_with("http://")) { // fail on read problem - File_read_http_result http=file_read_http(charsets, file_spec, params); + File_read_http_result http=file_read_http(charsets, file_spec, as_text, params); result.success=true; result.str=http.str; result.length=http.length; @@ -579,8 +593,6 @@ File_read_result file_read(Request_chars } fix_line_breaks((char *)(result.str), result.length); - // note: after fixing - ((char*&)(result.str))[result.length]=0; } return result; @@ -1049,6 +1061,11 @@ int __vsnprintf(char* b, size_t s, const // note: on win32& maybe somewhere else // vsnprintf do not writes terminating 0 in 'buffer full' case, reducing --s; + + // clients do not check for negative 's', feature: ignore such prints + if((ssize_t)s<0) + return 0; + #if _MSC_VER /* win32: @@ -1074,7 +1091,7 @@ int __vsnprintf(char* b, size_t s, const if(r<0) r=0; - else if(r>s) + else if((size_t)r>s) r=s; #endif b[r]=0;