--- parser3/src/targets/cgi/parser3.C 2001/03/21 16:59:07 1.29 +++ parser3/src/targets/cgi/parser3.C 2001/10/22 16:44:43 1.124 @@ -1,92 +1,130 @@ /** @file Parser: scripting and CGI main. - Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com) + Copyright(c) 2001 ArtLebedev Group(http://www.artlebedev.com) + Author: Alexander Petrosyan (http://design.ru/paf) - Author: Alexander Petrosyan (http://design.ru/paf) - - $Id: parser3.C,v 1.29 2001/03/21 16:59:07 paf Exp $ + $Id: parser3.C,v 1.124 2001/10/22 16:44:43 parser Exp $ */ -#ifdef HAVE_CONFIG_H -# include "pa_config.h" -#endif - +#include "pa_config_includes.h" #ifdef WIN32 # include -# include -#else -# include #endif -//\ifwin32 -#include -//#include -//\endifwin32 - -#include -#include -#include -#include - +#include "pa_sapi.h" +#include "classes.h" #include "pa_common.h" -#include "pa_globals.h" #include "pa_request.h" +#include "pa_socks.h" +#include "pa_version.h" + +#ifdef XML +#include +#endif -Pool pool; // global pool +//#define DEBUG_POOL_MALLOC + +// consts + +extern const char *main_RCSIds[]; +#ifdef USE_SMTP +extern const char *smtp_RCSIds[]; +#endif +extern const char *gd_RCSIds[]; +extern const char *classes_RCSIds[]; +extern const char *types_RCSIds[]; +extern const char *parser3_RCSIds[]; +#ifdef XML +extern const char *xalan_patched_RCSIds[]; +#endif +const char **RCSIds[]={ + main_RCSIds, +#ifdef USE_SMTP + smtp_RCSIds, +#endif + gd_RCSIds, + classes_RCSIds, + types_RCSIds, + parser3_RCSIds, +#ifdef XML + xalan_patched_RCSIds, +#endif + 0 +}; + +/// IIS refuses to read bigger chunks +const size_t READ_POST_CHUNK_SIZE=0x400*0x400; // 1M + +const char *argv0; +Pool pool(0); // global pool [dont describe to doxygen: it confuses it with param names] bool cgi; ///< we were started as CGI? -#ifdef WIN32 -# if _MSC_VER -// intercept global system errors -LONG WINAPI TopLevelExceptionFilter ( - struct _EXCEPTION_POINTERS *ExceptionInfo - ) { +// SAPI + +static void log(const char *fmt, va_list args) { + bool opened; + FILE *f=0; + + if(argv0) { + // beside by binary + char file_spec[MAX_STRING]; + strncpy(file_spec, argv0, MAX_STRING-1); file_spec[MAX_STRING-1]=0; // filespec of my binary + rsplit(file_spec, '/'); rsplit(file_spec, '\\');// strip filename + strcat(file_spec, "/parser3.log"); + f=fopen(file_spec, "at"); + } + opened=f!=0; + if(!opened) + f=stderr; + + // prefix + time_t t=time(0); + const char *stamp=ctime(&t); + fprintf(f, "[%.*s] ", strlen(stamp)-1, stamp); + // message + char buf[MAX_STRING]; - if(ExceptionInfo && ExceptionInfo->ExceptionRecord) { - struct _EXCEPTION_RECORD *r=ExceptionInfo->ExceptionRecord; - - int printed=0; - printed+=snprintf(buf+printed, MAX_STRING-printed, "Exception 0x%X at 0x%p", - r->ExceptionCode, - r->ExceptionAddress); - for(unsigned int i=0; iNumberParameters; i++) - printed+=snprintf(buf+printed, MAX_STRING-printed, ", 0x%X", - r->ExceptionInformation[i]); - } else - strcpy(buf, "Exception "); - - PTHROW(0, 0, - 0, - buf); + size_t size=vsnprintf(buf, MAX_STRING, fmt, args); + remove_crlf(buf, buf+size); - return EXCEPTION_EXECUTE_HANDLER; // never reached + fwrite(buf, size, 1, f); + // newline + fprintf(f, "\n"); + + if(opened) + fclose(f); + else + fflush(f); } -# endif -#endif - -//\if -void fix_slashes(char *s) { - if(s) - for(; *s; s++) - if(*s=='\\') - *s='/'; +// appends to parser3.log located beside my binary if openable, to stderr otherwize +void SAPI::log(Pool& , const char *fmt, ...) { + va_list args; + va_start(args,fmt); + ::log(fmt, args); + va_end(args); } -//\endif -// service funcs +void SAPI::die(const char *fmt, ...) { + va_list args; + va_start(args,fmt); + ::log(fmt, args); + va_end(args); + + exit(1); +} -const char *get_env(Pool& pool, const char *name) { - return getenv(name); +const char *SAPI::get_env(Pool& , const char *name) { + return getenv(name); } -int read_post(char *buf, int max_bytes) { - int read_size=0; +size_t SAPI::read_post(Pool& , char *buf, size_t max_bytes) { + size_t read_size=0; do { - int chunk_size=read - (fileno(stdin), buf+read_size, min(0x400*0x400, max_bytes-read_size)); + int chunk_size=read(fileno(stdin), + buf+read_size, min(READ_POST_CHUNK_SIZE, max_bytes-read_size)); if(chunk_size<0) break; read_size+=chunk_size; @@ -95,34 +133,226 @@ int read_post(char *buf, int max_bytes) return read_size; } -void output_header_attribute(const char *key, const char *value) { +void SAPI::add_header_attribute(Pool& , const char *key, const char *value) { if(cgi) printf("%s: %s\n", key, value); } -void output_body(const char *buf, size_t size) { - if(cgi) // header | body delimiter +/// @todo intelligent cache-control +void SAPI::send_header(Pool& ) { + if(cgi) { + puts("expires: Fri, 23 Mar 2001 09:32:23 GMT"); + + // header | body delimiter puts(""); + } +} +void SAPI::send_body(Pool& , const void *buf, size_t size) { stdout_write(buf, size); } -// main +// -int main(int argc, char *argv[]) { - // TODO:umask(2); -//\#ifdef WIN32 - setmode(fileno(stdin), _O_BINARY); - setmode(fileno(stdout), _O_BINARY); - setmode(fileno(stderr), _O_BINARY); -//\#endif +char *full_file_spec(char *file_name) { + if(file_name && !strchr(file_name, '/')) { + static char cwd[MAX_STRING]; getcwd(cwd, MAX_STRING); + static char buf[MAX_STRING]; + snprintf(buf, MAX_STRING, "%s/%s", cwd, file_name); + return buf; + } + return file_name; +} - // Service funcs - service_funcs.get_env=get_env; - service_funcs.read_post=read_post; - service_funcs.output_header_attribute=output_header_attribute; - service_funcs.output_body=output_body; +/** +main workhorse + + @todo + IIS: remove trailing default-document[index.html] from $request.uri. + to do that we need to consult metabase, + wich is tested but seems slow. +*/ +void real_parser_handler( + const char *filespec_to_process, + const char *request_method, bool header_only) { + // init socks + init_socks(pool); + +#ifdef XML + /** + * Initialize Xerces and Xalan. + * + * Should be called only once per process before making + * any other API calls. + */ + //_asm int 3; + XalanInitialize(); +#endif + + // init global classes + init_methoded_array(pool); + // init global variables + pa_globals_init(pool); + + if(!filespec_to_process) + throw Exception(0, 0, + 0, + "Parser/%s", PARSER_VERSION); + + // Request info + Request::Info request_info; + if(cgi) { + if(const char *env_document_root=SAPI::get_env(pool, "DOCUMENT_ROOT")) + request_info.document_root=env_document_root; + else if(const char *path_info=SAPI::get_env(pool, "PATH_INFO")) { + // IIS + size_t len=strlen(filespec_to_process)-strlen(path_info); + char *buf=(char *)pool.malloc(len+1); + memcpy(buf, filespec_to_process, len); buf[len]=0; + request_info.document_root=buf; + } else + throw Exception(0, 0, + 0, + "CGI: no PATH_INFO defined(in reinventing DOCUMENT_ROOT)"); + } else { + static char buf[MAX_STRING]; + strncpy(buf, filespec_to_process, MAX_STRING-1); buf[MAX_STRING-1]=0; + if(rsplit(buf, '/') || rsplit(buf, '\\')) // strip filename + request_info.document_root=buf; + else + request_info.document_root=""; + } + request_info.path_translated=filespec_to_process; + request_info.method=request_method ? request_method : "GET"; + const char *query_string=SAPI::get_env(pool, "QUERY_STRING"); + request_info.query_string=query_string; + if(cgi) { + if(const char *env_request_uri=SAPI::get_env(pool, "REQUEST_URI")) + request_info.uri=env_request_uri; + else if(const char *path_info=SAPI::get_env(pool, "PATH_INFO")) + if(query_string) { + char *reconstructed_uri=(char *)pool.malloc( + strlen(path_info)+1/*'?'*/+ + strlen(query_string)+1/*0*/); + strcpy(reconstructed_uri, path_info); + strcat(reconstructed_uri, "?"); + strcat(reconstructed_uri, query_string); + request_info.uri=reconstructed_uri; + } else + request_info.uri=path_info; + else + throw Exception(0, 0, + 0, + "CGI: no PATH_INFO defined(in reinventing REQUEST_URI)"); + + if(const char *script_name=SAPI::get_env(pool, "SCRIPT_NAME")) { + size_t script_name_len=strlen(script_name); + size_t uri_len=strlen(request_info.uri); + if(strncmp(request_info.uri,script_name, script_name_len)==0 && + script_name_len != uri_len) // under IIS they are the same + throw Exception(0, 0, + 0, + "CGI: illegal call"); + } + } else + request_info.uri=0; + + request_info.content_type=SAPI::get_env(pool, "CONTENT_TYPE"); + const char *content_length=SAPI::get_env(pool, "CONTENT_LENGTH"); + request_info.content_length=(content_length?atoi(content_length):0); + request_info.cookie=SAPI::get_env(pool, "HTTP_COOKIE"); + request_info.user_agent=SAPI::get_env(pool, "HTTP_USER_AGENT"); + + // prepare to process request + Request request(pool, + request_info, + cgi ? String::UL_USER_HTML : String::UL_AS_IS + ); + // some root-controlled location +#ifdef SYSCONFDIR + const char *root_config_filespec=SYSCONFDIR "/" CONFIG_FILE_NAME; +#else +# ifdef WIN32 + // c:\windows + char root_config_path[MAX_STRING]; + GetWindowsDirectory(root_config_path, MAX_STRING); + + char root_config_filespec[MAX_STRING]; + snprintf(root_config_filespec, MAX_STRING, + "%s/%s", + root_config_path, CONFIG_FILE_NAME); +# else +#error must be compiled either configure/make or MSVC++ +# endif +#endif + + // beside by binary + // @todo full path, not ./! + static char site_config_path[MAX_STRING]; + strncpy(site_config_path, argv0, MAX_STRING-1); site_config_path[MAX_STRING-1]=0; // filespec of my binary + if(!( + rsplit(site_config_path, '/') || + rsplit(site_config_path, '\\'))) { // strip filename + // no path, just filename + site_config_path[0]='.'; site_config_path[1]=0; + } + + char site_config_filespec[MAX_STRING]; + snprintf(site_config_filespec, MAX_STRING, + "%s/%s", + site_config_path, CONFIG_FILE_NAME); + + // process the request + request.core( + root_config_filespec, false, + site_config_filespec, false, + header_only); + + // + done_socks(); + +#ifdef DEBUG_POOL_MALLOC + extern void log_pool_stats(Pool& pool); + log_pool_stats(pool); +#endif +} + +void call_real_parser_handler__do_SEH( + const char *filespec_to_process, + const char *request_method, bool header_only) { +#if _MSC_VER & !defined(_DEBUG) + LPEXCEPTION_POINTERS system_exception=0; + __try { +#endif + real_parser_handler( + filespec_to_process, + request_method, header_only); + +#if _MSC_VER & !defined(_DEBUG) + } __except ( + (system_exception=GetExceptionInformation()), + EXCEPTION_EXECUTE_HANDLER) { + + if(system_exception) + if(_EXCEPTION_RECORD *er=system_exception->ExceptionRecord) + throw Exception(0, 0, + 0, + "Exception 0x%08X at 0x%08X", er->ExceptionCode, er->ExceptionAddress); + else + throw Exception(0, 0, 0, "Exception "); + else + throw Exception(0, 0, 0, "Exception "); + } +#endif +} + +int main(int argc, char *argv[]) { + int result; + argv0=argv[0]; + + umask(2); + // were we started as CGI? cgi= getenv("SERVER_SOFTWARE") || @@ -132,110 +362,72 @@ int main(int argc, char *argv[]) { if(!cgi) { if(argc<2) { - char *binary=argv[0]; - printf("Usage: %s \n", binary?binary:"parser3"); - exit(1); + printf( + "Parser/%s Copyright(c) 2001 ArtLebedev Group(http://www.artlebedev.com)\n" + "Author: Alexander Petrosyan (http://design.ru/paf)\n" + "\n" + "Usage: %s \n", + PARSER_VERSION, + argv0?argv0:"parser3"); + return 1; } } - char *filespec_to_process=cgi?getenv("PATH_TRANSLATED"):argv[1]; -//\#ifdef WIN32 - fix_slashes(filespec_to_process); -//\#endif - - PTRY { // global try - // must be first in PTRY{}PCATCH #ifdef WIN32 -# if _MSC_VER - SetUnhandledExceptionFilter(&TopLevelExceptionFilter); - //TODO: initSocks(); -# endif + setmode(fileno(stdin), _O_BINARY); + setmode(fileno(stdout), _O_BINARY); + setmode(fileno(stderr), _O_BINARY); #endif - // init global variables - globals_init(pool); - - if(!filespec_to_process) - PTHROW(0, 0, - 0, - "no file to process"); - - // Request info - Request::Info request_info; - const char *document_root=getenv("DOCUMENT_ROOT"); - if(!document_root) { - static char fake_document_root[MAX_STRING]; - strncpy(fake_document_root, filespec_to_process, MAX_STRING); - rsplit(fake_document_root, '/'); rsplit(fake_document_root, '\\');// strip filename - document_root=fake_document_root; - } - request_info.document_root=document_root; - request_info.path_translated=filespec_to_process; - request_info.method=getenv("REQUEST_METHOD"); - request_info.query_string=getenv("QUERY_STRING"); - request_info.uri=getenv("REQUEST_URI"); - request_info.content_type=getenv("CONTENT_TYPE"); - const char *content_length=getenv("CONTENT_LENGTH"); - request_info.content_length=(content_length?atoi(content_length):0); - request_info.cookie=getenv("HTTP_COOKIE"); - - // prepare to process request - Pool request_pool; - Request request(request_pool, - request_info, - cgi ? String::UL_HTML_TYPO : String::UL_NO - ); - - // some root-controlled location - char *sys_auto_path1; + char *filespec_to_process=cgi?getenv("PATH_TRANSLATED"):argv[1]; #ifdef WIN32 - // c:\windows - sys_auto_path1=(char *)pool.malloc(MAX_STRING); - GetWindowsDirectory(sys_auto_path1, MAX_STRING); - strcat(sys_auto_path1, PATH_DELIMITER_STRING); -#else - // ~nobody - sys_auto_path1=getenv("HOME"); + back_slashes_to_slashes(filespec_to_process); #endif - - // beside by binary - char *sys_auto_path2=(char *)pool.malloc(MAX_STRING); - strncpy(sys_auto_path2, argv[0], MAX_STRING); // filespec of my binary - rsplit(sys_auto_path2, '/'); rsplit(sys_auto_path2, '\\');// strip filename - strcat(sys_auto_path2, PATH_DELIMITER_STRING); - - // process the request - request.core(pool.exception(), - sys_auto_path1, - sys_auto_path2); - // no actions with request' data past this point - // request.exception not not handled here, but all - // request' data are associated with it's pool=exception + filespec_to_process=full_file_spec(filespec_to_process); + + const char *request_method=getenv("REQUEST_METHOD"); + bool header_only=request_method && strcasecmp(request_method, "HEAD")==0; + try { // global try + call_real_parser_handler__do_SEH( + filespec_to_process, + request_method, header_only); - // must be last in PTRY{}PCATCH -#ifdef WIN32 -# if _MSC_VER - SetUnhandledExceptionFilter(0); -# endif -#endif // successful finish - return 0; - } PCATCH(e) { // global problem + result=0; + } catch(const Exception& e) { // global problem + // don't allocate anything on pool here: + // possible pool' exception not catch-ed now + // and there could be out-of-memory exception + const char *body=e.comment(); + // log it + SAPI::log(pool, "exception in request exception handler: %s", body); + + // int content_length=strlen(body); - // header - (*service_funcs.output_header_attribute)("content-type", "text/plain"); + // prepare header + SAPI::add_header_attribute(pool, "content-type", "text/plain"); char content_length_cstr[MAX_NUMBER]; - snprintf(content_length_cstr, MAX_NUMBER, "%d", content_length); - (*service_funcs.output_header_attribute)("content-length", - content_length_cstr); + snprintf(content_length_cstr, MAX_NUMBER, "%u", content_length); + SAPI::add_header_attribute(pool, "content-length", content_length_cstr); + + // send header + SAPI::send_header(pool); // body - (*service_funcs.output_body)(body, content_length); + if(!header_only) + SAPI::send_body(pool, body, content_length); // unsuccessful finish - return 1; + result=1; } - PEND_CATCH + + +#ifndef WIN32 + // + if(!cgi) + SAPI::send_body(pool, "\n", 1); +#endif + return result; }