--- parser3/src/classes/file.C 2012/02/27 11:08:28 1.216 +++ parser3/src/classes/file.C 2016/07/21 18:30:10 1.243 @@ -1,12 +1,10 @@ /** @file Parser: @b file parser class. - Copyright (c) 2001-2009 ArtLebedev Group (http://www.artlebedev.com) + Copyright (c) 2001-2015 Art. Lebedev Studio (http://www.artlebedev.com) Author: Alexandr Petrosian (http://paf.design.ru) */ -static const char * const IDENT_FILE_C="$Date: 2012/02/27 11:08:28 $"; - #include "pa_config_includes.h" #include "classes.h" @@ -27,18 +25,38 @@ static const char * const IDENT_FILE_C=" #include "pa_vregex.h" #include "pa_version.h" +volatile const char * IDENT_FILE_C="$Id: file.C,v 1.243 2016/07/21 18:30:10 moko Exp $"; + // defines #define STDIN_EXEC_PARAM_NAME "stdin" #define CHARSET_EXEC_PARAM_NAME "charset" #define NAME_NAME "name" +#define KEEP_EMPTY_DIRS_NAME "keep-empty-dirs" +#define SUPPRESS_EXCEPTION_NAME "exception" // externs extern String sql_limit_name; extern String sql_offset_name; +// helpers + +class File_list_table_template_columns: public ArrayString { +public: + File_list_table_template_columns() { + *this+=new String("name"); + *this+=new String("dir"); + *this+=new String("size"); + *this+=new String("cdate"); + *this+=new String("mdate"); + *this+=new String("adate"); + } +}; + +Table file_list_table_template(new File_list_table_template_columns); + // class class MFile: public Methoded { @@ -50,7 +68,7 @@ public: // global variable -DECLARE_CLASS_VAR(file, new MFile, 0); +DECLARE_CLASS_VAR(file, new MFile); // consts @@ -128,26 +146,52 @@ static void _save(Request& r, MethodPara static void _delete(Request& r, MethodParams& params) { const String& file_name=params.as_string(0, FILE_NAME_MUST_NOT_BE_CODE); + bool keep_empty_dirs=false; + bool fail_on_problem=true; + + if(params.count()>1) + if(HashStringValue* options=params.as_hash(1)){ + int valid_options=0; + if(Value* vkeep_empty_dirs=options->get(KEEP_EMPTY_DIRS_NAME)){ + keep_empty_dirs=r.process_to_value(*vkeep_empty_dirs).as_bool(); + valid_options++; + } + if(Value* vsuppress_exception=options->get(SUPPRESS_EXCEPTION_NAME)){ + fail_on_problem=r.process_to_value(*vsuppress_exception).as_bool(); + valid_options++; + } + if(valid_options != options->count()) + throw Exception(PARSER_RUNTIME, 0, CALLED_WITH_INVALID_OPTION); + } // unlink - file_delete(r.absolute(file_name)); + file_delete(r.absolute(file_name), fail_on_problem, keep_empty_dirs); } static void _move(Request& r, MethodParams& params) { Value& vfrom_file_name=params.as_no_junction(0, "from file name must not be code"); Value& vto_file_name=params.as_no_junction(1, "to file name must not be code"); + bool keep_empty_dirs=false; + + if(params.count()>2) + if(HashStringValue* options=params.as_hash(2)){ + int valid_options=0; + if(Value* vkeep_empty_dirs=options->get(KEEP_EMPTY_DIRS_NAME)){ + keep_empty_dirs=r.process_to_value(*vkeep_empty_dirs).as_bool(); + valid_options++; + } + if(valid_options != options->count()) + throw Exception(PARSER_RUNTIME, 0, CALLED_WITH_INVALID_OPTION); + } // move file_move( r.absolute(vfrom_file_name.as_string()), - r.absolute(vto_file_name.as_string())); + r.absolute(vto_file_name.as_string()), + keep_empty_dirs); } -static void copy_process_source( - struct stat& , - int from_file, - const String& , const char* /*fname*/, bool, - void *context) { +static void copy_process_source(struct stat& , int from_file, const String&, void *context) { int& to_file=*static_cast(context); int nCount=0; @@ -166,7 +210,7 @@ static void copy_process_source( static void copy_open_target(int f, void *from_spec) { String& file_spec=*static_cast(from_spec); file_read_action_under_lock(file_spec, "copy", copy_process_source, &f); -}; +} static void _copy(Request& r, MethodParams& params) { Value& vfrom_file_name=params.as_no_junction(0, "from file name must not be code"); @@ -234,11 +278,9 @@ static void _load(Request& r, MethodPara if(Value* remote_content_type=file.headers->get(HTTP_CONTENT_TYPE_UPPER)) vcontent_type=new VString(*new String(remote_content_type->as_string().cstr())); } - - VFile& self=GET_SELF(r, VFile); - self.set(true/*tainted*/, file.str, file.length, user_file_name, vcontent_type, &r); - self.set_mode(as_text); + VFile& self=GET_SELF(r, VFile); + self.set(true/*tainted*/, as_text, file.str, file.length, user_file_name, vcontent_type, &r); if(file.headers){ file.headers->for_each(_load_pass_param, &self.fields()); @@ -249,9 +291,9 @@ static void _load(Request& r, MethodPara file_stat(lfile_name, size, atime, mtime, ctime); HashStringValue& ff=self.fields(); - ff.put(adate_name, new VDate(atime)); - ff.put(mdate_name, new VDate(mtime)); - ff.put(cdate_name, new VDate(ctime)); + ff.put(adate_name, new VDate((pa_time_t)atime)); + ff.put(mdate_name, new VDate((pa_time_t)mtime)); + ff.put(cdate_name, new VDate((pa_time_t)ctime)); } } @@ -308,22 +350,16 @@ static void _create(Request& r, MethodPa VFile& self=GET_SELF(r, VFile); if(const String* content_str=vcontent.get_string()){ - String::Body body=content_str->cstr_to_string_body_untaint(String::L_AS_IS); // explode content, honor tainting changes + String::Body body=content_str->cstr_to_string_body_untaint(String::L_AS_IS, r.connection(false), &r.charsets); // explode content, honor tainting changes if(asked_charset && is_text) body=Charset::transcode(body, r.charsets.source(), *asked_charset); - self.set(true/*tainted*/, body.cstr(), body.length()); - self.set_mode(is_text); + self.set(true/*tainted*/, is_text, body.cstrm(), body.length(), file_name, vcontent_type, &r); } else { if(asked_charset) throw Exception(PARSER_RUNTIME, 0, "charset option can not be used with file-content"); - self.set(*vcontent.as_vfile(String::L_AS_IS)); - if(mode) - self.set_mode(is_text); + self.set(*vcontent.as_vfile(String::L_AS_IS), mode != 0, is_text, file_name, vcontent_type, &r); } - self.set_name(file_name); - - self.set_content_type(vcontent_type, file_name, &r); } static void _stat(Request& r, MethodParams& params) { @@ -337,17 +373,17 @@ static void _stat(Request& r, MethodPara VFile& self=GET_SELF(r, VFile); - self.set(true/*tainted*/, 0/*no bytes*/, size, &lfile_name, 0, &r); + self.set_binary(true/*tainted*/, 0/*no bytes*/, size, &lfile_name, 0, &r); HashStringValue& ff=self.fields(); - ff.put(adate_name, new VDate(atime)); - ff.put(mdate_name, new VDate(mtime)); - ff.put(cdate_name, new VDate(ctime)); + ff.put(adate_name, new VDate((pa_time_t)atime)); + ff.put(mdate_name, new VDate((pa_time_t)mtime)); + ff.put(cdate_name, new VDate((pa_time_t)ctime)); } static bool is_safe_env_key(const char* key) { for(const char* validator=key; *validator; validator++) { char c=*validator; - if(!(c>='A' && c<='Z' || c>='0' && c<='9' || c=='_' || c=='-')) + if(!( (c>='A' && c<='Z') || (c>='0' && c<='9') || (c=='_' || c=='-') )) return false; } #ifdef PA_SAFE_MODE @@ -410,7 +446,7 @@ static void pass_cgi_header_attribute( static void append_to_argv(Request& r, ArrayString& argv, const String* str){ if(!str->is_empty()) - argv+=new String(str->cstr_to_string_body_untaint(String::L_AS_IS, 0, &r.charsets), String::L_AS_IS); + argv+=new String(str->cstr_to_string_body_untaint(String::L_AS_IS, r.connection(false), &r.charsets), String::L_AS_IS); } /// @todo fix `` in perl - they produced flipping consoles and no output to perl @@ -429,24 +465,14 @@ static void _exec_cgi(Request& r, Method const String& script_name=r.absolute(params.as_string(param_index++, FILE_NAME_MUST_NOT_BE_CODE)); HashStringString env; - #define ECSTR(name, value_cstr) \ - if(value_cstr) \ - env.put( \ - String::Body(#name), \ - String::Body(*value_cstr?value_cstr:0)); \ - // passing SAPI::environment - if(const char *const *pairs=SAPI::environment(r.sapi_info)) { - while(const char* pair=*pairs++) - if(const char* eq_at=strchr(pair, '=')) - if(eq_at[1]) // has value - env.put( - pa_strdup(pair, eq_at-pair), - pa_strdup(eq_at+1, 0)); - } + #define ECSTR(name, value_cstr) if(value_cstr) env.put(#name, value_cstr); + // passing environment + for(SAPI::Env::Iterator i(r.sapi_info); i; i.next() ) + env.put(i.key(), i.value() ); // const ECSTR(GATEWAY_INTERFACE, "CGI/1.1"); - ECSTR("PARSER_VERSION", PARSER_VERSION); + ECSTR(PARSER_VERSION, PARSER_VERSION); // from Request.info ECSTR(DOCUMENT_ROOT, r.request_info.document_root); ECSTR(PATH_TRANSLATED, r.request_info.path_translated); @@ -456,15 +482,13 @@ static void _exec_cgi(Request& r, Method ECSTR(CONTENT_TYPE, r.request_info.content_type); ECSTR(CONTENT_LENGTH, format(r.request_info.content_length, "%u")); // SCRIPT_* - env.put(String::Body("SCRIPT_NAME"), script_name); - //env.put(String::Body("SCRIPT_FILENAME"), ??&script_name); + env.put("SCRIPT_NAME", script_name); // environment & stdin from param String *in=new String(); Charset *charset=0; // default script works raw_in 'source' charset = no transcoding needed if(param_index < params.count()) { - Value& venv=params.as_no_junction(param_index++, "env must not be code"); - if(HashStringValue* user_env=venv.get_hash()) { + if(HashStringValue* user_env=params.as_hash(param_index++, "env")) { // $.charset [previewing to handle URI pieces] if(Value* vcharset=user_env->get(CHARSET_EXEC_PARAM_NAME)) charset=&charsets.get(vcharset->as_string() @@ -482,7 +506,7 @@ static void _exec_cgi(Request& r, Method if(info.vstdin) { if(const String* sstdin=info.vstdin->get_string()) { // untaint stdin - in = new String(sstdin->cstr_to_string_body_untaint(String::L_AS_IS), String::L_AS_IS); + in = new String(sstdin->cstr_to_string_body_untaint(String::L_AS_IS, r.connection(false), &r.charsets), String::L_AS_IS); } else if(VFile* vfile=static_cast(info.vstdin->as("file"))) in->append_know_length((const char* )vfile->value_ptr(), vfile->value_size(), String::L_TAINTED); @@ -606,7 +630,7 @@ static void _exec_cgi(Request& r, Method file_out->length -= headersize; // $body - self.set(false/*not tainted*/, file_out->str, file_out->length); + self.set(false/*not tainted*/, is_text, file_out->str, file_out->length); // $fields << header if(header) { @@ -622,19 +646,15 @@ static void _exec_cgi(Request& r, Method } } else { // ^file::exec // $body - self.set(false/*not tainted*/, file_out->str, file_out->length); + self.set(false/*not tainted*/, is_text, file_out->str, file_out->length); } - self.set_mode(is_text); - // $status self.fields().put(file_status_name, new VInt(execution.status)); // $stderr if(!real_err->is_empty()) - self.fields().put( - String::Body("stderr"), - new VString(*real_err)); + self.fields().put("stderr", new VString(*real_err)); } static void _exec(Request& r, MethodParams& params) { _exec_cgi(r, params, false); @@ -646,26 +666,49 @@ static void _cgi(Request& r, MethodParam static void _list(Request& r, MethodParams& params) { Value& relative_path=params.as_no_junction(0, "path must not be code"); + bool stat=false; VRegex* vregex=0; VRegexCleaner vrcleaner; + if(params.count()>1){ - Value& regexp=params.as_no_junction(1, "regexp must not be code"); - if(regexp.is_defined()){ - if(Value* value=regexp.as(VREGEX_TYPE)){ - vregex=static_cast(value); + Value& voption=params.as_no_junction(1, "option must not be code"); + if(voption.is_defined()) { + Value* vfilter=0; + if(HashStringValue* options=voption.get_hash()) { + int valid_options=0; + if(Value* vstat=options->get("stat")) { + stat=r.process_to_value(*vstat).as_bool(); + valid_options++; + } + if(Value* value=options->get("filter")) { + vfilter=value; + valid_options++; + } + if(valid_options!=options->count()) + throw Exception(PARSER_RUNTIME, 0, CALLED_WITH_INVALID_OPTION); } else { - vregex=new VRegex(r.charsets.source(), ®exp.as_string(), 0/*options*/); - vregex->study(); - vrcleaner.vregex=vregex; + vfilter=&voption; + } + if(vfilter) { + if(Value* value=vfilter->as(VREGEX_TYPE)) { + vregex=static_cast(value); + } else if(vfilter->is_string()) { + if(!vfilter->get_string()->trim().is_empty()) { + vregex=new VRegex(r.charsets.source(), &vfilter->as_string(), 0/*options*/); + vregex->study(); + vrcleaner.vregex=vregex; + } + } else { + throw Exception(PARSER_RUNTIME, 0, "filter must be regex or string"); + } } } } const char* absolute_path_cstr=r.absolute(relative_path.as_string()).taint_cstr(String::L_FILE_SPEC); - Table::columns_type columns(new ArrayString); - *columns+=new String("name"); - Table& table=*new Table(columns); + Table::Action_options table_options; + Table& table=*new Table(file_list_table_template, table_options); const int ovector_size=(1/*match*/)*3; int ovector[ovector_size]; @@ -677,6 +720,13 @@ static void _list(Request& r, MethodPara 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), String::L_TAINTED); + *row+=new String(String::Body::Format(ffblk.is_dir(stat) ? 1 : 0), String::L_CLEAN); + if(stat) { + *row+=VDouble(ffblk.size()).get_string(); + *row+=new String(String::Body::Format((int)ffblk.c_timestamp()), String::L_CLEAN); + *row+=new String(String::Body::Format((int)ffblk.m_timestamp()), String::L_CLEAN); + *row+=new String(String::Body::Format((int)ffblk.a_timestamp()), String::L_CLEAN); + } table+=row; } ); @@ -691,11 +741,13 @@ struct Lock_execute_body_info { Value* body_code; }; #endif + static void lock_execute_body(int , void *ainfo) { Lock_execute_body_info& info=*static_cast(ainfo); // execute body info.r->write_assign_lang(info.r->process(*info.body_code)); -}; +} + static void _lock(Request& r, MethodParams& params) { const String& file_spec=r.absolute(params.as_string(0, FILE_NAME_MUST_BE_STRING)); Lock_execute_body_info info={ @@ -710,8 +762,19 @@ static void _lock(Request& r, MethodPara &info); } +static size_t afterlastslash(const String& str) { + size_t pos=str.strrpbrk("/\\"); + return pos!=STRING_NOT_FOUND?pos+1:0; +} + +static size_t afterlastslash(const String& str, size_t right) { + size_t pos=str.strrpbrk("/\\", 0, right); + return pos!=STRING_NOT_FOUND?pos+1:0; +} + static void _find(Request& r, MethodParams& params) { const String& file_name=params.as_string(0, FILE_NAME_MUST_NOT_BE_CODE); + Value* not_found_code=(params.count()==2)?¶ms.as_junction(1, "not-found param must be code"):0; const String* file_spec; @@ -727,19 +790,23 @@ static void _find(Request& r, MethodPara } // monkey way - int after_base_slash=lastposafter(*file_spec, 0, "/", 1); - const String* dirname=&file_spec->mid(0, after_base_slash); - const String& basename=file_spec->mid(after_base_slash, file_spec->length()); - - int after_monkey_slash; - while((after_monkey_slash=lastposafter(*dirname, 0, "/", 1, true))>0) { + size_t last_slash=file_spec->strrpbrk("/\\"); + const String& dirname=file_spec->mid(0, last_slash!=STRING_NOT_FOUND?last_slash:0); + const String& basename=file_spec->mid(last_slash!=STRING_NOT_FOUND?last_slash+1:0, file_spec->length()); + + size_t rpos=dirname.is_empty()?0:dirname.length()-1; + while((rpos=dirname.rskipchars("/\\", 0, rpos))!=STRING_NOT_FOUND){ + size_t slash=dirname.strrpbrk("/\\", 0, rpos); + if(slash==STRING_NOT_FOUND) + break; String test_name; - test_name<<*(dirname=&dirname->mid(0, after_monkey_slash)); - test_name< . + // / > / + // /a > / + // /a/ > / // /a/some.tar.gz > /a - // /a/b/ > /a - int afterslash=lastposafter(file_spec, 0, "/", 1, true); - if(afterslash>0) - r.write_assign_lang(file_spec.mid(0, afterslash==1?1:afterslash-1)); - else + // /a/b/ > /a + // /a///b/ > /a + // /a/b/// > /a + // file > . + + if(file_spec.is_empty()) { + r.write_assign_lang(String(".")); + return; + } + + size_t p; + size_t slash; + if((p=file_spec.rskipchars("/\\"))==STRING_NOT_FOUND) + r.write_assign_lang(String("/")); + else { + if((slash=file_spec.strrpbrk("/\\", 0, p))!=STRING_NOT_FOUND) { + if((p=file_spec.rskipchars("/\\", 0, slash))==STRING_NOT_FOUND) + p=slash; + r.write_assign_lang(file_spec.mid(0, p+1)); + return; + } r.write_assign_lang(String(".")); + } } static void _basename(Request& r, MethodParams& params) { const String& file_spec=params.as_string(0, FILE_NAME_MUST_BE_STRING); + // works as *nix basename + + // empty > . + // / > / + // /a > a + // /a/ > a // /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())); + // /a/b/ > b + // /a///b/ > b + // /a/b/// > b + // file > file + + if(file_spec.is_empty()) { + r.write_assign_lang(String(".")); + return; + } + + size_t p=file_spec.rskipchars("/\\"); + if(p==STRING_NOT_FOUND) + r.write_assign_lang(String("/")); + else + r.write_assign_lang(file_spec.mid(afterlastslash(file_spec, p), p+1)); } 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 - 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())); + // /a/b.c/ > empty + // /a/b.c > b + size_t pos=afterlastslash(file_spec); + size_t dotpos=file_spec.strrpbrk(".", pos); + r.write_assign_lang(file_spec.mid(pos, dotpos!=STRING_NOT_FOUND?dotpos: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 - int afterdot=lastposafter(file_spec, 0, ".", 1); - if(afterdot>0) - r.write_assign_lang(file_spec.mid(afterdot, file_spec.length())); + // /a/b.c/ > empty + size_t pos=afterlastslash(file_spec); + size_t dotpos=file_spec.strrpbrk(".", pos); + if(dotpos!=STRING_NOT_FOUND) + r.write_assign_lang(file_spec.mid(dotpos+1, file_spec.length())); } static void _fullpath(Request& r, MethodParams& params) { @@ -873,7 +987,7 @@ static void _sql(Request& r, MethodParam ulong offset=0; if(params.count()>1) - if(HashStringValue* options=params.as_hash(1)){ + if(HashStringValue* options=params.as_hash(1, "sql options")) { int valid_options=0; if(Value* vfilename=options->get(NAME_NAME)) { valid_options++; @@ -903,17 +1017,14 @@ static void _sql(Request& r, MethodParam handlers, statement_string); - if(!handlers.value) - throw Exception(PARSER_RUNTIME, - 0, - "produced no result"); + if(!handlers.value.str) + throw Exception(PARSER_RUNTIME, 0, "produced no result"); VFile& self=GET_SELF(r, VFile); - self.set(true/*tainted*/, handlers.value.str, handlers.value.length, handlers.user_file_name + self.set_binary(true/*tainted*/, handlers.value.str, handlers.value.length, handlers.user_file_name , handlers.user_content_type ? new VString(*handlers.user_content_type) : 0 , &r); - self.set_mode(false/*binary*/); } static void _base64(Request& r, MethodParams& params) { @@ -923,8 +1034,9 @@ static void _base64(Request& r, MethodPa if(params.count()) { // decode: // ^file::base64[encoded] // backward - // ^file::base64[mode;user-file-name;encoded[;$.content-type[...]]] + // ^file::base64[mode;user-file-name;encoded[;$.content-type[...] $.strict(true|false)]] bool is_text=false; + bool strict=false; VString* vcontent_type=0; const String* user_file_name=0; size_t param_index=0; @@ -946,6 +1058,10 @@ static void _base64(Request& r, MethodPa vcontent_type=new VString(value->as_string()); valid_options++; } + if(Value* vstrict=options->get(BASE64_STRICT_OPTION_NAME)) { + strict=r.process_to_value(*vstrict).as_bool(); + valid_options++; + } if(valid_options!=options->count()) throw Exception(PARSER_RUNTIME, 0, CALLED_WITH_INVALID_OPTION); } @@ -957,15 +1073,9 @@ static void _base64(Request& r, MethodPa char* decoded=0; size_t length=0; - pa_base64_decode(encoded, strlen(encoded), decoded, length); - - if(length && is_text) - fix_line_breaks(decoded, length); - - self.set(true/*tainted*/, decoded, length, user_file_name, vcontent_type, &r); + pa_base64_decode(encoded, strlen(encoded), decoded, length, strict); - if(params.count() > 1) - self.set_mode(is_text); + self.set(true/*tainted*/, is_text, decoded, length, user_file_name, vcontent_type, &r); } else { // encode: ^f.base64[] const char* encoded=pa_base64_encode(self.value_ptr(), self.value_size()); @@ -998,11 +1108,7 @@ static void _crc32(Request& r, MethodPar } -static void file_md5_file_action( - struct stat& finfo, - int f, - const String& , const char* /*fname*/, bool, - void *context) +static void file_md5_file_action(struct stat& finfo, int f, const String&, void *context) { PA_MD5_CTX& md5context=*static_cast(context); if(finfo.st_size) { @@ -1070,10 +1176,12 @@ MFile::MFile(): Methoded("file") { add_native_method("save", Method::CT_DYNAMIC, _save, 2, 3); // ^file:delete[file-name] - add_native_method("delete", Method::CT_STATIC, _delete, 1, 1); + // ^file:delete[file-name;$.keep-empty-dir(true)$.exception(false)] + add_native_method("delete", Method::CT_STATIC, _delete, 1, 2); // ^file:move[from-file-name;to-file-name] - add_native_method("move", Method::CT_STATIC, _move, 2, 2); + // ^file:move[from-file-name;to-file-name;$.keep-empty-dir(true)] + add_native_method("move", Method::CT_STATIC, _move, 2, 3); // ^file::load[mode;disk-name] // ^file::load[mode;disk-name;user-name] @@ -1096,6 +1204,7 @@ MFile::MFile(): Methoded("file") { // ^file:list[path] // ^file:list[path][regexp] + // ^file:list[path][$.filter[regexp] $.stat(true)] add_native_method("list", Method::CT_STATIC, _list, 1, 2); // ^file:lock[path]{code}