--- parser3/src/classes/file.C 2008/09/03 15:15:32 1.178 +++ parser3/src/classes/file.C 2009/04/22 04:41:32 1.185 @@ -1,11 +1,11 @@ /** @file Parser: @b file parser class. - Copyright (c) 2001-2005 ArtLebedev Group (http://www.artlebedev.com) + Copyright (c) 2001-2009 ArtLebedev Group (http://www.artlebedev.com) Author: Alexandr Petrosian (http://paf.design.ru) */ -static const char * const IDENT_FILE_C="$Date: 2008/09/03 15:15:32 $"; +static const char * const IDENT_FILE_C="$Date: 2009/04/22 04:41:32 $"; #include "pa_config_includes.h" @@ -24,6 +24,7 @@ static const char * const IDENT_FILE_C=" #include "pa_charsets.h" #include "pa_sql_connection.h" #include "pa_md5.h" +#include "pa_vregex.h" // defines @@ -153,10 +154,10 @@ static void _move(Request& r, MethodPara } static void copy_process_source( - struct stat& , - int from_file, - const String& , const char* /*fname*/, bool, - void *context) { + struct stat& , + int from_file, + const String& , const char* /*fname*/, bool, + void *context) { int& to_file=*static_cast(context); int nCount=0; @@ -165,7 +166,7 @@ static void copy_process_source( nCount = file_block_read(from_file, buffer, sizeof(buffer)); int written=write(to_file, buffer, nCount); if( written < 0 ) - throw Exception(0, + throw Exception("file.access", 0, "write failed: %s (%d)", strerror(errno), errno); @@ -192,46 +193,55 @@ static void _copy(Request& r, MethodPara } static void _load_pass_param( - HashStringValue::key_type key, - HashStringValue::value_type value, - HashStringValue *dest) { + HashStringValue::key_type key, + HashStringValue::value_type value, + HashStringValue *dest) { dest->put(key, value); } + static void _load(Request& r, MethodParams& params) { - Value& vmode_name=params.as_no_junction(0, MODE_MUST_NOT_BE_CODE); + bool as_text=is_text_mode(params.as_no_junction(0, MODE_MUST_NOT_BE_CODE).as_string()); const String& lfile_name=r.absolute(params.as_no_junction(1, FILE_NAME_MUST_NOT_BE_CODE).as_string()); - Value* third_param=params.count()>2?¶ms.as_no_junction(2, "filename or options must not be code") - :0; - HashStringValue* third_param_hash=third_param?third_param->get_hash():0; - size_t alt_filename_param_index=2; - if(third_param_hash) - alt_filename_param_index++; - HashStringValue* options=third_param_hash; + size_t param_index=params.count()-1; + Value* param_value=param_index>1?¶ms.as_no_junction(param_index, "filename or options must not be code"):0; + + HashStringValue* options=0; + const char *user_file_name=0; + + if(param_value){ + options=param_value->get_hash(); + if(options || param_index>2) + param_index--; + if(param_index>1){ + const String& luser_file_name=params.as_string(param_index, FILE_NAME_MUST_BE_STRING); + if(!luser_file_name.is_empty()) + user_file_name=luser_file_name.cstr(String::L_FILE_SPEC); + } + } + if(!user_file_name) + user_file_name=lfile_name.cstr(String::L_FILE_SPEC); + size_t offset=0; size_t limit=0; - if(options) { + + if(options){ options=new HashStringValue(*options); - if(Value *voffset=(Value *)options->get(sql_offset_name)) { + if(Value *voffset=(Value *)options->get(sql_offset_name)){ offset=r.process_to_value(*voffset).as_int(); } - if(Value *vlimit=(Value *)options->get(sql_limit_name)) { + if(Value *vlimit=(Value *)options->get(sql_limit_name)){ limit=r.process_to_value(*vlimit).as_int(); } // no check on options count here, see file_read } - File_read_result file=file_read(r.charsets, lfile_name, - is_text_mode(vmode_name.as_string()), - options, true, 0, offset, limit + File_read_result file=file_load(r, lfile_name, + as_text, options, true, 0, offset, limit ); - const char *user_file_name=params.count()>alt_filename_param_index? - params.as_string(alt_filename_param_index, FILE_NAME_MUST_BE_STRING).cstr() - :lfile_name.cstr(String::L_FILE_SPEC); - Value* vcontent_type=0; if(file.headers){ - if(Value* remote_content_type=file.headers->get("CONTENT-TYPE")) + if(Value* remote_content_type=file.headers->get(HTTP_CONTENT_TYPE_UPPER)) vcontent_type=new VString(*new String(remote_content_type->as_string().cstr())); } if(!vcontent_type) @@ -321,9 +331,9 @@ struct Append_env_pair_info { }; #endif static void append_env_pair( - HashStringValue::key_type akey, - HashStringValue::value_type avalue, - Append_env_pair_info *info) { + HashStringValue::key_type akey, + HashStringValue::value_type avalue, + Append_env_pair_info *info) { if(akey==STDIN_EXEC_PARAM_NAME) { info->vstdin=avalue; } else if(akey==CHARSET_EXEC_PARAM_NAME) { @@ -344,15 +354,15 @@ struct Pass_cgi_header_attribute_info { }; #endif static void pass_cgi_header_attribute( - ArrayString::element_type astring, - Pass_cgi_header_attribute_info* info) { + ArrayString::element_type astring, + Pass_cgi_header_attribute_info* info) { size_t colon_pos=astring->pos(':'); if(colon_pos!=STRING_NOT_FOUND) { const String& key=astring->mid(0, colon_pos).change_case( *info->charset, String::CC_UPPER); Value* value=new VString(astring->mid(colon_pos+1, astring->length()).trim()); info->fields->put(key, value); - if(key=="CONTENT-TYPE") + if(key==HTTP_CONTENT_TYPE_UPPER) info->content_type=value; } } @@ -459,9 +469,9 @@ static void _exec_cgi(Request& r, Method // argv from params ArrayString argv; if(param_index < params.count()) { - // influence tainting - // main target -- URLencoding of tainted pieces to String::L_URI lang - Temp_client_charset temp(r.charsets, charset? *charset: r.charsets.source()); + // influence tainting + // main target -- URLencoding of tainted pieces to String::L_URI lang + Temp_client_charset temp(r.charsets, charset? *charset: r.charsets.source()); for(size_t i=param_index; i1) { - regexp=¶ms.as_no_junction(1, "regexp must not be code").as_string(); - - const char* pattern=regexp->cstr(String::L_UNSPECIFIED); - const char* errptr; - int erroffset; - int options=PCRE_EXTRA | PCRE_DOTALL; - if(r.charsets.source().isUTF8()) - options=options|PCRE_UTF8; - - regexp_code=pcre_compile(pattern, options, - &errptr, &erroffset, - r.charsets.source().pcre_tables); - - if(!regexp_code) - throw Exception(PCRE_EXCEPTION_TYPE, - ®exp->mid(erroffset, regexp->length()), - "regular expression syntax error - %s", errptr); + VRegex* vregex; + VRegexCleaner vrcleaner; + if(params.count()>1){ + Value& regexp=params.as_no_junction(1, "regexp must not be code"); + if(Value* value=regexp.as(VREGEX_TYPE, false)){ + vregex=static_cast(value); + } else { + vregex=new VRegex(r.charsets.source(), ®exp.as_string(), 0/*options*/); + vrcleaner.vregex=vregex; + } + vregex->study(); } else { - regexp=0; // not used, just to calm down compiler - regexp_code=0; + vregex=0; } - const char* absolute_path_cstr=r.absolute(relative_path.as_string()).cstr(String::L_FILE_SPEC); Table::columns_type columns(new ArrayString); *columns+=new String("name"); Table& table=*new Table(columns); + const int ovector_size=(1/*match*/)*3; + int ovector[ovector_size]; + LOAD_DIR(absolute_path_cstr, const char* file_name_cstr=ffblk.ff_name; size_t file_name_size=strlen(file_name_cstr); - bool suits=true; - if(regexp_code) { - int exec_result=pcre_exec(regexp_code, 0, - ffblk.ff_name, file_name_size, 0, - 0, ovector, ovecsize); - - if(exec_result==PCRE_ERROR_NOMATCH) - suits=false; - else if(exec_result<0) { - (*pcre_free)(regexp_code); - throw Exception(PCRE_EXCEPTION_TYPE, - regexp, - print_pcre_exec_error_text(exec_result), - exec_result); - } - } - if(suits) { + if(!vregex || vregex->exec(ffblk.ff_name, file_name_size, ovector, ovector_size)>=0) { Table::element_type row(new ArrayString); - *row+=new String(pa_strdup(file_name_cstr, file_name_size), file_name_size, true); + *row+=new String(pa_strdup(file_name_cstr, file_name_size), file_name_size, true/*tainted*/); table+=row; } ); - if(regexp_code) - pcre_free(regexp_code); - // write out result r.write_no_lang(*new VTable(&table)); } @@ -754,7 +735,7 @@ static void _find(Request& r, MethodPara static void _dirname(Request& r, MethodParams& params) { const String& file_spec=params.as_string(0, FILE_NAME_MUST_BE_STRING); - // /a/some.tar.gz > /a + // /a/some.tar.gz > /a // /a/b/ > /a int afterslash=lastposafter(file_spec, 0, "/", 1, true); if(afterslash>0) @@ -765,21 +746,21 @@ static void _dirname(Request& r, MethodP static void _basename(Request& r, MethodParams& params) { const String& file_spec=params.as_string(0, FILE_NAME_MUST_BE_STRING); - // /a/some.tar.gz > some.tar.gz + // /a/some.tar.gz > some.tar.gz int afterslash=lastposafter(file_spec, 0, "/", 1); r.write_assign_lang(file_spec.mid(afterslash, file_spec.length())); } static void _justname(Request& r, MethodParams& params) { const String& file_spec=params.as_string(0, FILE_NAME_MUST_BE_STRING); - // /a/some.tar.gz > some.tar + // /a/some.tar.gz > some.tar int afterslash=lastposafter(file_spec, 0, "/", 1); int afterdot=lastposafter(file_spec, afterslash, ".", 1); r.write_assign_lang(file_spec.mid(afterslash, afterdot!=afterslash?afterdot-1:file_spec.length())); } static void _justext(Request& r, MethodParams& params) { const String& file_spec=params.as_string(0, FILE_NAME_MUST_BE_STRING); - // /a/some.tar.gz > gz + // /a/some.tar.gz > gz int afterdot=lastposafter(file_spec, 0, ".", 1); if(afterdot>0) r.write_assign_lang(file_spec.mid(afterdot, file_spec.length())); @@ -928,26 +909,26 @@ static void _sql(Request& r, MethodParam static void _base64(Request& r, MethodParams& params) { bool dynamic = !(&r.get_self() == file_class); - if ( dynamic ){ - VFile& self=GET_SELF(r, VFile); - if(params.count()) { - // decode - const char* cstr=params.as_string(0, PARAMETER_MUST_BE_STRING).cstr(); - char* decoded_cstr=0; - size_t decoded_size=0; - pa_base64_decode(cstr, strlen(cstr), decoded_cstr, decoded_size); - if(decoded_cstr && decoded_size) - self.set(true/*tainted*/, decoded_cstr, decoded_size); - } else { - // encode - const char* encoded=pa_base64_encode(self.value_ptr(), self.value_size()); - r.write_assign_lang(*new String(encoded, 0, true/*once ?param=base64(something) was needed*/)); - } + if(dynamic){ + VFile& self=GET_SELF(r, VFile); + if(params.count()) { + // decode: ^file::base64[encoded] + const char* cstr=params.as_string(0, PARAMETER_MUST_BE_STRING).cstr(); + char* decoded=0; + size_t length=0; + pa_base64_decode(cstr, strlen(cstr), decoded, length); + if(decoded && length) + self.set(true/*tainted*/, decoded, length); + } else { + // encode: ^f.base64[] + const char* encoded=pa_base64_encode(self.value_ptr(), self.value_size()); + r.write_assign_lang(*new String(encoded, 0, true/*tainted. once ?param=base64(something) was needed**/)); + } } else { - // encode + // encode: ^file:base64[filespec] const String& file_spec=params.as_string(0, FILE_NAME_MUST_BE_STRING); const char* encoded=pa_base64_encode(r.absolute(file_spec)); - r.write_assign_lang(*new String(encoded, 0, true/*once ?param=base64(something) was needed*/)); + r.write_assign_lang(*new String(encoded, 0, true/*tainted. once ?param=base64(something) was needed*/)); } } @@ -973,10 +954,10 @@ static void _crc32(Request& r, MethodPar static void file_md5_file_action( - struct stat& finfo, - int f, - const String& , const char* /*fname*/, bool, - void *context) + struct stat& finfo, + int f, + const String& , const char* /*fname*/, bool, + void *context) { PA_MD5_CTX& md5context=*static_cast(context); if(finfo.st_size) { @@ -1052,7 +1033,7 @@ MFile::MFile(): Methoded("file") { // ^file::load[mode;disk-name] // ^file::load[mode;disk-name;user-name] - add_native_method("load", Method::CT_DYNAMIC, _load, 2, 3); + add_native_method("load", Method::CT_DYNAMIC, _load, 2, 4); // ^file::stat[disk-name] add_native_method("stat", Method::CT_DYNAMIC, _stat, 1, 1);