|
|
| version 1.185, 2004/04/06 10:08:15 | version 1.217, 2006/02/03 15:08:04 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: commonly functions. | Parser: commonly functions. |
| Copyright(c) 2001-2004 ArtLebedev Group (http://www.artlebedev.com) | Copyright(c) 2001-2005 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 | |
| * Authors: Michael Zucchi <notzed@ximian.com> | |
| * Jeffrey Stedfast <fejj@ximian.com> | |
| * | |
| * Copyright 2000-2004 Ximian, Inc. (www.ximian.com) | |
| * | |
| * This program is free software; you can redistribute it and/or modify | |
| * it under the terms of the GNU General Public License as published by | |
| * the Free Software Foundation; either version 2 of the License, or | |
| * (at your option) any later version. | |
| * | |
| * This program is distributed in the hope that it will be useful, | |
| * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
| * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
| * GNU General Public License for more details. | |
| * | |
| * You should have received a copy of the GNU General Public License | |
| * along with this program; if not, write to the Free Software | |
| * Foundation, Inc., 59 Temple Street #330, Boston, MA 02111-1307, USA. | |
| * | |
| */ | |
| static const char * const IDENT_COMMON_C="$Date$"; | static const char * const IDENT_COMMON_C="$Date$"; |
| Line 11 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 | |
| #ifdef PA_HTTP | |
| #include "pa_vstring.h" | |
| #include "pa_vint.h" | |
| #include "pa_vhash.h" | |
| #include "pa_vtable.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 64 int PASCAL closesocket(SOCKET); | Line 54 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 76 int PASCAL closesocket(SOCKET); | Line 62 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 130 char* file_read_text(Request_charsets& c | Line 99 char* file_read_text(Request_charsets& c |
| return file.success?file.str:0; | return file.success?file.str:0; |
| } | } |
| //http request stuff | /// these options were handled but not checked elsewhere, now check them |
| #ifdef PA_HTTP | int pa_get_valid_file_options_count(HashStringValue& options) |
| #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; | |
| } | |
| 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; | |
| char* EOLat=0; | |
| while(true) { | |
| char buf[MAX_STRING*10]; | |
| ssize_t received_size=recv(sock, buf, sizeof(buf), 0); | |
| if(received_size<=0) | |
| break; | |
| 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, " "); | |
| 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"); | |
| } | |
| } | |
| if(result) | |
| 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, int port, | |
| const char* request, | |
| int | |
| #ifdef PA_USE_ALARM | |
| timeout | |
| #endif | |
| , | |
| bool fail_on_status_ne_200) { | |
| if(!host) | |
| throw Exception("http.host", | |
| 0, | |
| "zero hostname"); //never | |
| #ifdef PA_USE_ALARM | |
| signal(SIGALRM, timeout_handler); | |
| #endif | |
| int sock=-1; | |
| #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 [though duplicating closesocket code] | |
| if(sock>=0) | |
| closesocket(sock); | |
| throw Exception("http.timeout", | |
| origin_string, | |
| "timeout occured while retrieving document"); | |
| return 0; // never | |
| } else { | |
| alarm(timeout); | |
| #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) | |
| throw Exception("http.connect", | |
| 0, | |
| "can not make socket: %s (%d)", strerror(errno), errno); | |
| if(connect(sock, (struct sockaddr *)&dest, sizeof(dest))) | |
| throw Exception("http.connect", | |
| 0, | |
| "can not connect to host \"%s\": %s (%d)", host, strerror(errno), errno); | |
| size_t request_size=strlen(request); | |
| if(send(sock, request, request_size, 0)!=(ssize_t)request_size) | |
| throw Exception("http.connect", | |
| 0, | |
| "error sending request: %s (%d)", strerror(errno), errno); | |
| 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(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()) | int result=0; |
| form_string_value2string(key, *svalue, *result); | if(options.get(PA_SQL_LIMIT_NAME)) |
| else if(Table* tvalue=value->get_table()) { | result++; |
| Form_table_value2string_info info(key, *result); | if(options.get(PA_SQL_OFFSET_NAME)) |
| tvalue->for_each(form_table_value2string, &info); | result++; |
| } else | if(options.get(PA_COLUMN_SEPARATOR_NAME)) |
| throw Exception(0, | result++; |
| new String(key, String::L_TAINTED), | if(options.get(PA_COLUMN_ENCLOSER_NAME)) |
| "is %s, "HTTP_FORM_NAME" option value must either string or table", value->type()); | result++; |
| } | |
| 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 | |
| /// @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; | |
| int port; | |
| const char* method="GET"; bool method_is_get; | |
| HashStringValue* form=0; | |
| const char* body_cstr=0; | |
| int timeout=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=0; | |
| 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=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?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, fail_on_status_ne_200); | |
| //processing results | |
| 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"); | |
| *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); | |
| HashStringValue& tables=vtables->hash(); | |
| size_t pos_after=0; | |
| header_block.split(aheaders, pos_after, CRLF); | |
| //processing headers | |
| size_t aheaders_count=aheaders.count(); | |
| Charset* real_remote_charset=0; // undetected, yet | |
| for(size_t i=1; i<aheaders_count; i++) { | |
| const String& line=*aheaders.get(i); | |
| size_t pos=line.pos(": ", 2); | |
| 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+2, line.length()); | |
| if(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)); | |
| } | |
| // defaulting to used-asked charset [it's never empty!] | |
| if(!real_remote_charset) | |
| real_remote_charset=asked_remote_charset; | |
| // 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 | |
| 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; | 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; |
| char* buf; size_t offset; size_t count; | |
| }; | }; |
| #endif | #endif |
| static void file_read_action( | static void file_read_action( |
| Line 654 static void file_read_action( | Line 126 static void file_read_action( |
| 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); |
| if(size_t to_read_size=(size_t)finfo.st_size) { | size_t to_read_size=info.count; |
| *info.data=new(PointerFreeGC) char[to_read_size+(as_text?1:0)]; | if(!to_read_size) |
| to_read_size=(size_t)finfo.st_size; | |
| assert( !(info.buf && as_text) ); | |
| if(to_read_size) { | |
| if(info.offset) | |
| lseek(f, info.offset, SEEK_SET); | |
| *info.data=info.buf | |
| ? info.buf | |
| : new(PointerFreeGC) char[to_read_size+(as_text?1:0)]; | |
| *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) |
| Line 664 static void file_read_action( | Line 144 static void file_read_action( |
| "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); |
| } else { // empty file | } else { // empty file |
| if(as_text) { | // for both, text and binary: for text we need that terminator, for binary we need nonzero pointer to be able to save such files |
| *info.data=new(PointerFreeGC) char[1]; | *info.data=new(PointerFreeGC) char[1]; |
| *(char*)(*info.data)=0; | *(char*)(*info.data)=0; |
| } else | |
| *info.data=0; | |
| *info.data_size=0; | *info.data_size=0; |
| 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) { | |
| 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) | |
| throw Exception("parser.runtime", | |
| 0, | |
| "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(charsets, file_spec, as_text, params); |
| 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 | if(params) { |
| if(params && params->count()) | int valid_options=pa_get_valid_file_options_count(*params); |
| throw Exception("parser.runtime", | if(valid_options!=params->count()) |
| 0, | throw Exception("parser.runtime", |
| "invalid option passed"); | 0, |
| "invalid option passed"); | |
| } | |
| File_read_action_info info={&result.str, &result.length}; | File_read_action_info info={&result.str, &result.length, |
| buf, offset, count}; | |
| result.success=file_read_action_under_lock(file_spec, | result.success=file_read_action_under_lock(file_spec, |
| "read", file_read_action, &info, | "read", file_read_action, &info, |
| as_text, fail_on_read_problem); | as_text, fail_on_read_problem); |
| #ifdef PA_HTTP | |
| } | } |
| #endif | |
| if(result.success && as_text) { | if(result.success && as_text) { |
| // UTF-8 signature: EF BB BF | // UTF-8 signature: EF BB BF |
| Line 787 bool file_read_action_under_lock(const S | Line 271 bool file_read_action_under_lock(const S |
| } | } |
| } | } |
| static void create_dir_for_file(const String& file_spec) { | void create_dir_for_file(const String& file_spec) { |
| size_t pos_after=1; | size_t pos_after=1; |
| size_t pos_before; | size_t pos_before; |
| while((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) { | while((pos_before=file_spec.pos('/', pos_after))!=STRING_NOT_FOUND) { |
| Line 946 static bool entry_readable(const String& | Line 430 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; |
| Line 1047 const char* format(double value, char* f | Line 531 const char* format(double value, char* f |
| size_t stdout_write(const void *buf, size_t size) { | size_t stdout_write(const void *buf, size_t size) { |
| #ifdef WIN32 | #ifdef WIN32 |
| size_t to_write = size; | |
| do{ | do{ |
| int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout); | int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout); |
| if(chunk_written<=0) | if(chunk_written<=0) |
| Line 1055 size_t stdout_write(const void *buf, siz | Line 540 size_t stdout_write(const void *buf, siz |
| buf=((const char*)buf)+chunk_written; | buf=((const char*)buf)+chunk_written; |
| } while(size>0); | } while(size>0); |
| return size; | return to_write-size; |
| #else | #else |
| return fwrite(buf, 1, size, stdout); | return fwrite(buf, 1, size, stdout); |
| #endif | #endif |
| Line 1068 char* unescape_chars(const char* cp, int | Line 553 char* unescape_chars(const char* cp, int |
| EscapeFirst, | EscapeFirst, |
| EscapeSecond | EscapeSecond |
| } escapeState=EscapeRest; | } escapeState=EscapeRest; |
| int escapedValue=0; | uint escapedValue=0; |
| int srcPos=0; | int srcPos=0; |
| int dstPos=0; | int dstPos=0; |
| while(srcPos < len) { | while(srcPos < len) { |
| int ch=cp[srcPos]; | uchar ch=(uchar)cp[srcPos]; |
| switch(escapeState) { | switch(escapeState) { |
| case EscapeRest: | case EscapeRest: |
| if(ch=='%') { | if(ch=='%') { |
| Line 1084 char* unescape_chars(const char* cp, int | Line 569 char* unescape_chars(const char* cp, int |
| } | } |
| break; | break; |
| case EscapeFirst: | case EscapeFirst: |
| escapedValue=hex_value[ch] << 4; | escapedValue=hex_value[ch] << 4; |
| escapeState=EscapeSecond; | escapeState=EscapeSecond; |
| break; | break; |
| case EscapeSecond: | case EscapeSecond: |
| escapedValue +=hex_value[ch]; | escapedValue +=hex_value[ch]; |
| s[dstPos++]=escapedValue; | s[dstPos++]=(char)escapedValue; |
| escapeState=EscapeRest; | escapeState=EscapeRest; |
| break; | break; |
| } | } |
| Line 1125 bool StrEqNc(const char* s1, const char* | Line 610 bool StrEqNc(const char* s1, const char* |
| return !strict; | return !strict; |
| } else if(!(*s2)) | } else if(!(*s2)) |
| return !strict; | return !strict; |
| if(isalpha(*s1)) { | if(isalpha((unsigned char)*s1)) { |
| if(tolower(*s1) !=tolower(*s2)) | if(tolower((unsigned char)*s1) !=tolower((unsigned char)*s2)) |
| return false; | return false; |
| } else if((*s1) !=(*s2)) | } else if((*s1) !=(*s2)) |
| return false; | return false; |
| Line 1266 static char *base64_alphabet = | Line 751 static char *base64_alphabet = |
| static size_t | static size_t |
| g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save) | g_mime_utils_base64_encode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save) |
| { | { |
| const register unsigned char *inptr; | register const unsigned char *inptr; |
| register unsigned char *outptr; | register unsigned char *outptr; |
| if (inlen <= 0) | if (inlen <= 0) |
| Line 1324 g_mime_utils_base64_encode_step (const u | Line 809 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 1383 g_mime_utils_base64_encode_close (const | Line 868 g_mime_utils_base64_encode_close (const |
| return (outptr - out); | return (outptr - out); |
| } | } |
| char* pa_base64(const char *in, size_t len) | static unsigned char gmime_base64_rank[256] = { |
| 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, | |
| 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, | |
| 255,255,255,255,255,255,255,255,255,255,255, 62,255,255,255, 63, | |
| 52, 53, 54, 55, 56, 57, 58, 59, 60, 61,255,255,255, 0,255,255, | |
| 255, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, | |
| 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25,255,255,255,255,255, | |
| 255, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, | |
| 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51,255,255,255,255,255, | |
| 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, | |
| 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, | |
| 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, | |
| 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, | |
| 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, | |
| 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, | |
| 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, | |
| 255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255, | |
| }; | |
| /** | |
| * g_mime_utils_base64_decode_step: | |
| * @in: input stream | |
| * @inlen: max length of data to decode | |
| * @out: output stream | |
| * @state: holds the number of bits that are stored in @save | |
| * @save: leftover bits that have not yet been decoded | |
| * | |
| * Decodes a chunk of base64 encoded data. | |
| * | |
| * Returns the number of bytes decoded (which have been dumped in @out). | |
| **/ | |
| size_t | |
| g_mime_utils_base64_decode_step (const unsigned char *in, size_t inlen, unsigned char *out, int *state, int *save) | |
| { | |
| const unsigned char *inptr; | |
| unsigned char *outptr; | |
| const unsigned char *inend; | |
| int saved; | |
| unsigned char c; | |
| int i; | |
| inend = in + inlen; | |
| outptr = out; | |
| /* convert 4 base64 bytes to 3 normal bytes */ | |
| saved = *save; | |
| i = *state; | |
| inptr = in; | |
| while (inptr < inend) { | |
| c = gmime_base64_rank[*inptr++]; | |
| if (c != 0xff) { | |
| saved = (saved << 6) | c; | |
| i++; | |
| if (i == 4) { | |
| *outptr++ = (unsigned char)(saved >> 16); | |
| *outptr++ = (unsigned char)(saved >> 8); | |
| *outptr++ = (unsigned char)(saved); | |
| i = 0; | |
| } | |
| } | |
| } | |
| *save = saved; | |
| *state = i; | |
| /* quick scan back for '=' on the end somewhere */ | |
| /* fortunately we can drop 1 output char for each trailing = (upto 2) */ | |
| i = 2; | |
| while (inptr > in && i) { | |
| inptr--; | |
| if (gmime_base64_rank[*inptr] != 0xff) { | |
| if (*inptr == '=' && outptr > out) | |
| outptr--; | |
| i--; | |
| } | |
| } | |
| /* if i != 0 then there is a truncation error! */ | |
| return (outptr - out); | |
| } | |
| 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[len * 2 + 6]; | char* result=new(PointerFreeGC) char[in_size * 2 + 6]; |
| int state=0; | int state=0; |
| int save=0; | int save=0; |
| #ifndef NDEBUG | #ifndef NDEBUG |
| size_t filled= | size_t filled= |
| #endif | #endif |
| g_mime_utils_base64_encode_close ((const unsigned char*)in, len, | g_mime_utils_base64_encode_close ((const unsigned char*)in, in_size, |
| (unsigned char*)result, &state, &save); | (unsigned char*)result, &state, &save); |
| assert(filled <= len * 2 + 6); | assert(filled <= in_size * 2 + 6); |
| return result; | return result; |
| } | } |
| 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) */ | |
| result=new(PointerFreeGC) char[in_size+1/*terminator*/]; | |
| int state=0; | |
| int save=0; | |
| result_size= | |
| g_mime_utils_base64_decode_step ((const unsigned char*)in, in_size, | |
| (unsigned char*)result, &state, &save); | |
| assert(result_size <= in_size); | |
| result[result_size]=0; // for text files | |
| } |