--- parser3/src/main/pa_common.C 2003/01/31 12:34:36 1.143.2.10 +++ parser3/src/main/pa_common.C 2003/03/20 11:11:55 1.143.2.21.2.6 @@ -5,19 +5,40 @@ Author: Alexandr Petrosian (http://paf.design.ru) */ -static const char* IDENT_COMMON_C="$Date: 2003/01/31 12:34:36 $"; +static const char* IDENT_COMMON_C="$Date: 2003/03/20 11:11:55 $"; #include "pa_common.h" #include "pa_exception.h" -#include "pa_value_includes.h" #include "pa_hash.h" +#include "pa_globals.h" + +//#define PA_HTTP + +#ifdef PA_HTTP #include "pa_vstring.h" #include "pa_vint.h" -#ifdef WIN32 -# include +#ifdef CYGWIN +#define _GNU_H_WINDOWS32_SOCKETS +// for PASCAL +#include +// SOCKET +typedef u_int SOCKET; +int PASCAL closesocket(SOCKET); #else -# define closesocket close +# if defined(WIN32) +# include +# else +# define closesocket close +# endif +#endif + +#else + +# if defined(WIN32) +# include +# endif + #endif // some maybe-undefined constants @@ -84,8 +105,33 @@ static int unlock(int fd) { FLOCK(F_TLOC #endif #endif +// defines for globals + +#define FILE_STATUS_NAME "status" + +// globals + +const String file_status_name(FILE_STATUS_NAME); + +// defines for statics + +#define HTTP_METHOD_NAME "method" +#define HTTP_TIMEOUT_NAME "timeout" +#define HTTP_HEADERS_NAME "headers" +#define HTTP_ANY_STATUS_NAME "any-status" + +// 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); + +// defines + #define DEFAULT_USER_AGENT "parser3" +// functions void fix_line_breaks(char* buf, size_t& size) { if(size==0) @@ -115,17 +161,17 @@ void fix_line_breaks(char* buf, size_t& memcpy(dest, bol, eob-bol); } -char* file_read_text(Pool& pool, Charset& charset, - StringPtr file_spec, +char* file_read_text(Charset& charset, + const String& file_spec, bool fail_on_read_problem, - HashStringValue *params, HashStringValue** out_fields) { - char *result; size_t size; - File_read_result file_read_result=file_read(pool, charset, file_spec, result, size, true, params, out_fields, fail_on_read_problem); - return file_read_result.success?result:0; + HashStringValue* params/*, HashStringValue* * out_fields*/) { + File_read_result file= + file_read(charset, file_spec, true, params, fail_on_read_problem); + return file.success?file.data:0; } +#ifdef PA_HTTP //http request stuff -/* ************************ http stuff *********************** */ static bool set_addr(struct sockaddr_in *addr, const char* host, const short port){ memset(addr, 0, sizeof(*addr)); @@ -141,20 +187,20 @@ static bool set_addr(struct sockaddr_in return true; } -static int http_read_response(Pool& pool, String& response, int sock, bool fail_on_status_ne_200){ +static int http_read_responseString& response, int sock, bool fail_on_status_ne_200){ int result=0; ssize_t EOLat=0; while(true) { - char *buf=(char *)pool.malloc(MAX_STRING); + char *buf=new(PointerFreeGC) char[MAX_STRING]; ssize_t size=recv(sock, buf, MAX_STRING, 0); if(size<=0) break; response.APPEND_TAINTED(buf, size, "remote HTTP server response", 0); if(!result && (EOLat=response.pos("\r\n", 2))>=0) { // checking status in first response - StringPtr status_line=response.mid(0, (size_t)EOLat); - Array astatus; + const String& status_line=response.mid(0, (size_t)EOLat); + ArrayString astatus; status_line->split(astatus, 0, " ", 1); - StringPtr status_code=astatus.get(1); + const String& status_code=astatus.get(1); result=status_code->as_int(); if(fail_on_status_ne_200 && result!=200) @@ -167,7 +213,7 @@ static int http_read_response(Pool& pool return result; else throw Exception("http.response", - Exception::undefined_source, + 0, "bad response from host - no status found (size=%lu)", response.size()); } @@ -184,9 +230,9 @@ static void timeout_handler(int sig){ } #endif -static int http_request(Pool& pool, +static int http_request( String& response, - StringPtr origin_string, + const String* origin_string, const char* host, int port, const char* request, int timeout, @@ -231,7 +277,7 @@ static int http_request(Pool& pool, origin_string, "error sending request: %s (%d)", strerror(errno), errno); - result=http_read_response(pool, response, sock, fail_on_status_ne_200); + result=http_read_response(response, sock, fail_on_status_ne_200); closesocket(sock); #ifdef WE_CAN_USE_ALARM alarm(0); @@ -244,13 +290,13 @@ static int http_request(Pool& pool, #ifdef WE_CAN_USE_ALARM alarm(0); #endif - /*re*/throw; + rethrow; } } #ifndef DOXYGEN struct Http_pass_header_info { - Pool* pool; + ; Charset* charset; String* request; bool user_agent_specified; @@ -259,29 +305,34 @@ struct Http_pass_header_info { 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(*info->pool, value, String::UL_HTTP_HEADER, false) + << *attributed_meaning_to_string(*info->value, String::UL_HTTP_HEADER, false) <<"\n"; - if(*key->change_case(*info->pool, *info->charset, String::CC_UPPER)=="USER-AGENT") + if(*key->change_case(*info->*info->charset, String::CC_UPPER)=="USER-AGENT") info->user_agent_specified=true; } -static HashStringValuePtr file_read_http(Pool& pool, Charset& charset, StringPtr file_spec, - char*& data, size_t& data_size, - HashStringValue *options=0, bool return_outfields=false) { - HashStringValuePtr result(0); +#ifndef DOXYGEN +struct File_read_http_result { + char *data; size_t size; + HashStringValue* headers; +}; +#endif +static File_read_http_result file_read_httpCharset& charset, const String& file_spec, + HashStringValue *options=0) { + File_read_http_result result; char host[MAX_STRING]; const char* uri; int port; - CharPtr method(0); + const char* method(0); int timeout=2; bool fail_on_status_ne_200=true; ValuePtr vheaders(0); - StringPtr connect_string(new String()); + const String& connect_string(new String()); // not in ^sql{... UL_SQL ...} spirit, but closer to ^file::load one connect_string->append(*file_spec, String::UL_URI); // tainted pieces -> URI pieces - CharPtr connect_string_cstr=connect_string->cstr(String::UL_UNSPECIFIED); + const char* connect_string_cstr=connect_string->cstr(String::UL_UNSPECIFIED); const char* current=connect_string_cstr; if(strncmp(current, "http://", 7)!=0) throw Exception(0, @@ -316,7 +367,7 @@ static HashStringValuePtr file_read_http if(valid_options!=options->count()) throw Exception("parser.runtime", - Exception::undefined_source, + 0, "invalid option passed"); } @@ -330,7 +381,7 @@ static HashStringValuePtr file_read_http bool user_agent_specified=false; if(vheaders && !vheaders->is_string()) { // allow empty if(HashStringValue *headers=vheaders->get_hash(connect_string)) { - Http_pass_header_info info={&pool, &charset, &request}; + Http_pass_header_info info={&&charset, &request}; headers->for_each(http_pass_header, &info); user_agent_specified=info.user_agent_specified; } else @@ -344,8 +395,8 @@ static HashStringValuePtr file_read_http //sending request String response; - CharPtr request_cstr=request.cstr(String::UL_UNSPECIFIED); - int status_code=http_request(pool, response, + const char* request_cstr=request.cstr(String::UL_UNSPECIFIED); + int status_code=http_request(response, connect_string, host, port, request_cstr, timeout, fail_on_status_ne_200); @@ -356,60 +407,59 @@ static HashStringValuePtr file_read_http connect_string, "bad response from host - no headers found"); } - StringPtr header_block=response.mid(0, pos); - StringPtr body=response.mid(pos+4, response.size()); + const String& header_block=response.mid(0, pos); + const String& body=response.mid(pos+4, response.size()); - Array aheaders; - HashStringValuePtr headers(new HashStringValue); + ArrayString aheaders; + result.headers=HashStringValue* (new HashStringValue); header_block->split(aheaders, 0, "\r\n", 2); //processing headers for(int i=1; ipos(": ", 2); if(pos<1) throw Exception("http.response", connect_string, - "bad response from host - bad header \"%s\"", line->cstr()); + "bad response from host - bad header \"%s\"", line->cstr().get()); - headers->put( - line->mid(0, pos)->change_case(pool, charset, String::CC_UPPER), + result.headers->put( + line->mid(0, pos)->change_case(charset, String::CC_UPPER), ValuePtr(new VString(line->mid(pos+2, line->size())))); } // output response - data=new(pool) char[data_size=body->size()]; - body->store_to(data, String::UL_AS_IS); - if(return_outfields) { - headers->put(file_status_name, ValuePtr(new VInt(status_code))); - return headers; - } else - return HashStringValuePtr(0); + result.data=new(PointerFreeGC) char[1/*as text puts terminating zero later*/+(result.size=body->size())]; + body->store_to(result.data, String::UL_AS_IS); + result.headers->put(file_status_name, ValuePtr(new VInt(status_code))); + return result; } +#endif + #ifndef DOXYGEN struct File_read_action_info { char **data; size_t *data_size; }; #endif -static void file_read_action(Pool& pool, +static void file_read_action( struct stat& finfo, int f, - StringPtr 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) { - *info.data=pool.malloc(to_read_size+(as_text?1:0)); + *info.data=new(PointerFreeGC) char[to_read_size+(as_text?1:0)]; *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) throw Exception(0, - file_spec, + &file_spec, "read failed: actually read %lu bytes count not in [0..%lu] valid range", *info.data_size, to_read_size); } else { // empty file if(as_text) { - *info.data=pool.malloc(1); + *info.data=new(PointerFreeGC) char[1]; *(char*)(*info.data)=0; } else *info.data=0; @@ -417,45 +467,51 @@ static void file_read_action(Pool& pool, return; } } -File_read_result file_read(Pool& pool, Charset& charset, StringPtr file_spec, - char*& data, size_t& data_size, - bool as_text, HashStringValue *params, HashStringValue** out_fields, +File_read_result file_read(Charset& charset, const String& file_spec, + bool as_text, HashStringValue *params, bool fail_on_read_problem) { File_read_result result; - if(file_spec->starts_with("http://", 7)) { +#ifdef PA_HTTP + if(file_spec.starts_with("http://", 7)) { // fail on read problem - result.fields=file_read_http(pool, charset, file_spec, data, data_size, params); + File_read_http_result http=file_read_http(charset, file_spec, params); result.success=true; + result.data=http.data; + result.size=http.size; + result.headers=http.headers; } else { - File_read_action_info info={&data, &data_size}; - result.success=file_read_action_under_lock(pool, file_spec, +#endif + File_read_action_info info={&result.data, &result.size}; + 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(data_size>=3) { - char *in=(char *)data; + if(result.size>=3) { + char *in=(char *)result.data; if((in[0] == '\xEF') && (in[1] == '\xBB') && (in[2] == '\xBF')) { - data=in+3; data_size-=3;// skip prefix + result.data=in+3; result.size-=3;// skip prefix } } - fix_line_breaks((char *)(data), data_size); + fix_line_breaks((char *)(result.data), result.size); // note: after fixing - ((char*&)(data))[data_size]=0; + ((char*&)(result.data))[result.size]=0; } return result; } -bool file_read_action_under_lock(Pool& pool, StringPtr file_spec, +bool file_read_action_under_lock(const String& file_spec, const char* action_name, File_read_action action, void *context, bool as_text, bool fail_on_read_problem) { - CharPtr fname=file_spec->cstr(String::UL_FILE_SPEC); + const char* fname=file_spec.cstr(String::UL_FILE_SPEC); int f; // first open, next stat: @@ -471,31 +527,31 @@ bool file_read_action_under_lock(Pool& p try { if(lock_shared_blocking(f)!=0) throw Exception("file.lock", - file_spec, + &file_spec, "shared lock failed: %s (%d), actual filename '%s'", - strerror(errno), errno, fname); + strerror(errno), errno, fname); struct stat finfo; if(stat(fname, &finfo)!=0) throw Exception("file.missing", // hardly possible: we just opened it OK - file_spec, + &file_spec, "stat failed: %s (%d), actual filename '%s'", - strerror(errno), errno, fname); + strerror(errno), errno, fname); #ifdef PA_SAFE_MODE if(finfo.st_uid/*foreign?*/!=geteuid() && finfo.st_gid/*foreign?*/!=getegid()) throw Exception("parser.runtime", - file_spec, + &file_spec, "parser is in safe mode: reading files of foreign group and user disabled [recompile parser with --disable-safe-mode configure option], actual filename '%s'", - fname); + fname); #endif - action(pool, finfo, f, file_spec, fname, as_text, context); + action(finfo, f, file_spec, fname, as_text, context); } catch(...) { unlock(f);close(f); if(fail_on_read_problem) - /*re*/throw; + rethrow; return false; } @@ -504,30 +560,30 @@ bool file_read_action_under_lock(Pool& p } else { if(fail_on_read_problem) throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0, - file_spec, + &file_spec, "%s failed: %s (%d), actual filename '%s'", - action_name, strerror(errno), errno, fname); + action_name, strerror(errno), errno, fname); return false; } } -static void create_dir_for_file(StringPtr file_spec) { +static void create_dir_for_file(const String& file_spec) { size_t pos_after=1; int pos_before; - while((pos_before=file_spec->pos("/", 1, pos_after))>=0) { - mkdir(file_spec->mid(0, pos_before)->cstr(String::UL_FILE_SPEC), 0775); + while((pos_before=file_spec.pos("/", pos_after))>=0) { + mkdir(file_spec.mid(0, pos_before).cstr(String::UL_FILE_SPEC), 0775); pos_after=pos_before+1; } } bool file_write_action_under_lock( - StringPtr file_spec, + const String& file_spec, const char* action_name, File_write_action action, void *context, bool as_text, bool do_append, bool do_block, bool fail_on_lock_problem) { - CharPtr fname=file_spec->cstr(String::UL_FILE_SPEC); + const char* fname=file_spec.cstr(String::UL_FILE_SPEC); int f; if(access(fname, W_OK)!=0) // no create_dir_for_file(file_spec); @@ -538,9 +594,9 @@ bool file_write_action_under_lock( |(do_append?O_APPEND:PA_O_TRUNC), 0664))>=0) { if((do_block?lock_exclusive_blocking(f):lock_exclusive_nonblocking(f))!=0) { Exception e("file.lock", - file_spec, + &file_spec, "shared lock failed: %s (%d), actual filename '%s'", - strerror(errno), errno, fname); + strerror(errno), errno, fname); close(f); if(fail_on_lock_problem) throw e; @@ -555,7 +611,7 @@ bool file_write_action_under_lock( ftruncate(f, lseek(f, 0, SEEK_CUR)); // one can not use O_TRUNC, read lower #endif unlock(f);close(f); - /*re*/throw; + rethrow; } #ifdef HAVE_FTRUNCATE @@ -566,15 +622,15 @@ bool file_write_action_under_lock( return true; } else throw Exception(errno==EACCES?"file.access":0, - file_spec, + &file_spec, "%s failed: %s (%d), actual filename '%s'", - action_name, strerror(errno), errno, fname); + action_name, strerror(errno), errno, fname); // here should be nothing, see rethrow above } #ifndef DOXYGEN struct File_write_action_info { - const void *data; size_t size; + const char* data; size_t size; }; #endif static void file_write_action(int f, void *context) { @@ -583,13 +639,13 @@ static void file_write_action(int f, voi int written=write(f, info.data, info.size); if(written<0) throw Exception(0, - Exception::undefined_source, + 0, "write failed: %s (%d)", strerror(errno), errno); } } void file_write( - StringPtr file_spec, - const void *data, size_t size, + const String& file_spec, + const char* data, size_t size, bool as_text, bool do_append) { File_write_action_info info={data, size}; @@ -601,38 +657,38 @@ void file_write( } // throws nothing! [this is required in file_move & file_delete] -static void rmdir(StringPtr file_spec, size_t pos_after) { +static void rmdir(const String& file_spec, size_t pos_after) { int pos_before; - if((pos_before=file_spec->pos("/", 1, pos_after))>=0) + if((pos_before=file_spec.pos("/", pos_after))>=0) rmdir(file_spec, pos_before+1); - rmdir(file_spec->mid(0, pos_after-1/* / */)->cstr(String::UL_FILE_SPEC)); + rmdir(file_spec.mid(0, pos_after-1/* / */).cstr(String::UL_FILE_SPEC)); } -bool file_delete(StringPtr file_spec, bool fail_on_read_problem) { - CharPtr fname=file_spec->cstr(String::UL_FILE_SPEC); +bool file_delete(const String& file_spec, bool fail_on_read_problem) { + const char* fname=file_spec.cstr(String::UL_FILE_SPEC); if(unlink(fname)!=0) if(fail_on_read_problem) throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0, - file_spec, + &file_spec, "unlink failed: %s (%d), actual filename '%s'", - strerror(errno), errno, fname); + strerror(errno), errno, fname); else return false; rmdir(file_spec, 1); return true; } -void file_move(StringPtr old_spec, StringPtr new_spec) { - CharPtr old_spec_cstr=old_spec->cstr(String::UL_FILE_SPEC); - CharPtr new_spec_cstr=new_spec->cstr(String::UL_FILE_SPEC); +void file_move(const String& old_spec, const String& new_spec) { + const char* old_spec_cstr=old_spec.cstr(String::UL_FILE_SPEC); + const char* new_spec_cstr=new_spec.cstr(String::UL_FILE_SPEC); create_dir_for_file(new_spec); if(rename(old_spec_cstr, new_spec_cstr)!=0) throw Exception(errno==EACCES?"file.access":errno==ENOENT?"file.missing":0, - old_spec, + &old_spec, "rename failed: %s (%d), actual filename '%s' to '%s'", - strerror(errno), errno, old_spec_cstr, new_spec_cstr); + strerror(errno), errno, old_spec_cstr, new_spec_cstr); rmdir(old_spec, 1); } @@ -646,13 +702,13 @@ bool entry_exists(const char* fname, str return result; } -bool entry_exists(StringPtr file_spec) { - CharPtr fname=file_spec->cstr(String::UL_FILE_SPEC); +bool entry_exists(const String& file_spec) { + const char* fname=file_spec.cstr(String::UL_FILE_SPEC); return entry_exists(fname, 0); } -static bool entry_readable(StringPtr file_spec, bool need_dir) { - CharPtr fname=file_spec->cstr(String::UL_FILE_SPEC); +static bool entry_readable(const String& file_spec, bool need_dir) { + char* fname=file_spec.cstrm(String::UL_FILE_SPEC); if(need_dir) { size_t size=strlen(fname); while(size) { @@ -670,36 +726,36 @@ static bool entry_readable(StringPtr fil } return false; } -bool file_readable(StringPtr file_spec) { +bool file_readable(const String& file_spec) { return entry_readable(file_spec, false); } -bool dir_readable(StringPtr file_spec) { +bool dir_readable(const String& file_spec) { return entry_readable(file_spec, true); } -StringPtr file_readable(StringPtr path, const String& name) { - StringPtr result(new String(*path)); - *result << "/"; - *result << name; - return file_readable(result)?result:StringPtr(0); +const String* file_readable(String& path, String& name) { + String& result=*new String(path); + result << "/"; + result << name; + return file_readable(result)?&result:0; } -bool file_executable(StringPtr file_spec) { - return access(file_spec->cstr(String::UL_FILE_SPEC), X_OK)==0; +bool file_executable(const String& file_spec) { + return access(file_spec.cstr(String::UL_FILE_SPEC), X_OK)==0; } -bool file_stat(StringPtr file_spec, +bool file_stat(const String& file_spec, size_t& rsize, time_t& ratime, time_t& rmtime, time_t& rctime, bool fail_on_read_problem) { - CharPtr fname=file_spec->cstr(String::UL_FILE_SPEC); + const char* fname=file_spec.cstr(String::UL_FILE_SPEC); struct stat finfo; if(stat(fname, &finfo)!=0) if(fail_on_read_problem) throw Exception("file.missing", - file_spec, + &file_spec, "getting file size failed: %s (%d), real filename '%s'", - strerror(errno), errno, fname); + strerror(errno), errno, fname); else return false; rsize=finfo.st_size; @@ -751,7 +807,7 @@ char* rsplit(char* string, char delim) { } /// @todo less stupid type detection -const char* format(Pool& pool, double value, char* fmt) { +const char* format(double value, char* fmt) { char local_buf[MAX_NUMBER]; size_t size; @@ -766,13 +822,13 @@ const char* format(Pool& pool, double va else size=snprintf(local_buf, sizeof(local_buf), "%d", (int)value); - return pool.copy(local_buf, size+1); + return pa_strdup(local_buf, size); } size_t stdout_write(const void *buf, size_t size) { #ifdef WIN32 do{ - int chunk_written=fwrite(buf, 1, min(8*0x400, size), stdout); + int chunk_written=fwrite(buf, 1, min((size_t)8*0x400, size), stdout); if(chunk_written<=0) break; size-=chunk_written; @@ -785,8 +841,8 @@ size_t stdout_write(const void *buf, siz #endif } -char* unescape_chars(Pool& pool, const char* cp, int len) { - char* s=pool.malloc(len + 1); +char* unescape_chars(const char* cp, int len) { + char* s=new(PointerFreeGC) char[len + 1]; enum EscapeState { EscapeRest, EscapeFirst, @@ -907,7 +963,7 @@ int __vsnprintf(char* b, size_t s, const win32: mk:@MSITStore:C:\Program%20Files\Microsoft%20Visual%20Studio\MSDN\2001APR\1033\vccore.chm::/html/_crt__vsnprintf.2c_._vsnwprintf.htm - if the number of bytes to write exceeds buffer, then count bytes are written and –1 is returned + if the number of bytes to write exceeds buffer, then count bytes are written and Ö1 is returned */ r=_vsnprintf(b, s, f, l); if(r<0)