|
|
| version 1.106, 2001/10/11 10:21:44 | version 1.144.2.6, 2003/01/24 14:36:10 |
|---|---|
| Line 1 | Line 1 |
| /** @file | /** @file |
| Parser: string class decl. | Parser: string class decl. |
| Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com) | Copyright (c) 2001, 2003 ArtLebedev Group (http://www.artlebedev.com) |
| Author: Alexander Petrosyan <paf@design.ru> (http://design.ru/paf) | Author: Alexandr Petrosian <paf@design.ru> (http://paf.design.ru) |
| $Id$ | |
| */ | */ |
| #ifndef PA_STRING_H | #ifndef PA_STRING_H |
| #define PA_STRING_H | #define PA_STRING_H |
| #include "pa_config_includes.h" | static const char* IDENT_STRING_H="$Date$"; |
| #include "pa_pool.h" | #include "pa_pool.h" |
| #include "pa_types.h" | #include "pa_types.h" |
| #include "pa_array.h" | |
| class Table; | // forwards |
| /** | class String; |
| $MAIN:html-typo table elements must enlarge string not more that that. | |
| that's a tradeoff - otherwise we'd have to scan string twice: | typedef object_ptr<String> StringPtr; |
| - first for buffer length | typedef object_ptr<const String> ConstStringPtr; |
| - second for replacements themselves | const ConstStringPtr ConstStringPtrZero(0); |
| */ | |
| #define UNTAINT_TIMES_BIGGER 10 | class Table; |
| typedef object_ptr<Table> TablePtr; | |
| class SQL_Connection; | |
| class Dictionary; | |
| // helpers | |
| #ifndef NO_STRING_ORIGIN | #ifndef NO_STRING_ORIGIN |
| # define STRING_APPEND_PARAMS \ | # define STRING_APPEND_PARAMS \ |
| const char *src, size_t size, \ | const char *src, size_t size, \ |
| String::Untaint_lang lang, \ | uchar lang, \ |
| const char *file, uint line | const char *file, uint line |
| /// appends piece to String @see String::real_append | /// appends piece to String @see String::real_append |
| # define APPEND(src, size, lang, file, line) \ | # define APPEND(src, size, lang, file, line) \ |
| Line 37 class Table; | Line 41 class Table; |
| # define STRING_APPEND_PARAMS \ | # define STRING_APPEND_PARAMS \ |
| const char *src, \ | const char *src, \ |
| size_t size, \ | size_t size, \ |
| String::Untaint_lang lang | uchar lang |
| /// appends piece to String @see String::real_append | /// appends piece to String @see String::real_append |
| # define APPEND(src, size, lang, file, line) \ | # define APPEND(src, size, lang, file, line) \ |
| real_append(src, size, lang) | real_append(src, size, lang) |
| Line 54 class Table; | Line 58 class Table; |
| /// handy: appends const char* piece to String @see String::real_append | /// handy: appends const char* piece to String @see String::real_append |
| #define APPEND_CONST(src) APPEND_AS_IS(src, 0, 0, 0) | #define APPEND_CONST(src) APPEND_AS_IS(src, 0, 0, 0) |
| class Array; | struct String_fragment { |
| class SQL_Connection; | uchar/*Untaint_lang*/ lang; ///< untaint flag, later untaint language |
| class Dictionary; | const char *ptr; ///< pointer to the start |
| size_t size; ///< length | |
| /** | #ifndef NO_STRING_ORIGIN |
| Pooled string. | /// all String pieces hold information of where they come from |
| struct Origin { | |
| const char *file; ///< macros file name | load file name | sql query text | |
| uint line; ///< file line no | record no | |
| } origin; ///< origin | |
| #endif | |
| }; | |
| Internal structure: | /** |
| @verbatim | String which knows the language of all it's fragments. |
| String Chunk0 | |
| ====== ======== | |
| head--------------->[ptr, size, ...] | |
| append_here-------->[ptr, size, ...] | |
| . | |
| . | |
| [ptr, size, ...] | |
| link_row----------->[link to the next chunk] | |
| @endverbatim | |
| All pieces remember | All pieces remember |
| - the file and its line they are from [can be turned off by NO_STRING_ORIGIN] | - the file and its line they are from [can be turned off by NO_STRING_ORIGIN] |
| - whether they are tainted or not, | - whether they are tainted or not, |
| and the language which should be used to detaint them | and the language which should be used to detaint them |
| */ | */ |
| class String : public Pooled { | |
| public: | |
| enum { | class String: public Array<String_fragment> { |
| CR_PREALLOCATED_COUNT=2, ///< default preallocated item count | size_t fsize; |
| CR_GROW_COUNT=2 ///< each time the String chunk_is_full() string expanded() | int count() const; // hiding fro accidental use instead of size() |
| }; | public: |
| /** piece is tainted or not. the language to use when detaint | /** piece is tainted or not. the language to use when detaint |
| remember to change String_Untaint_lang_name @ untaint.C along | remember to change String_Untaint_lang_name @ untaint.C along |
| */ | */ |
| enum Untaint_lang { | enum Untaint_lang { |
| UL_UNSPECIFIED=0, ///< zero value handy for hash lookup @see untaint_lang_name2enum | UL_UNSPECIFIED=0, ///< zero value handy for hash lookup @see untaint_lang_name2enum |
| // these two must go before others, there are checks for >UL_AS_IS | |
| UL_CLEAN, ///< clean | UL_CLEAN, ///< clean |
| UL_TAINTED, ///< tainted, untaint language as assigned later | UL_AS_IS, ///< leave all characters intact |
| // untaint languages. assigned by ^untaint[lang]{...} | |
| UL_PASS_APPENDED, | UL_PASS_APPENDED, |
| /**< | /**< |
| leave language built into string being appended. | leave language built into string being appended. |
| just a flag, that value not stored | just a flag, that value not stored |
| */ | */ |
| UL_AS_IS, ///< leave all characters intact | UL_TAINTED, ///< tainted, untaint language as assigned later |
| // untaint languages. assigned by ^untaint[lang]{...} | |
| UL_FILE_SPEC, ///< file specification | UL_FILE_SPEC, ///< file specification |
| UL_HTTP_HEADER, ///< text in HTTP response header | UL_HTTP_HEADER, ///< text in HTTP response header |
| UL_MAIL_HEADER, ///< text in mail header | UL_MAIL_HEADER, ///< text in mail header |
| Line 109 public: | Line 111 public: |
| UL_JS, ///< JavaScript code | UL_JS, ///< JavaScript code |
| UL_XML, ///< ^dom:set xml | UL_XML, ///< ^dom:set xml |
| UL_HTML, ///< HTML code (for editing) | UL_HTML, ///< HTML code (for editing) |
| UL_USER_HTML ///< HTML code with USER chars | UL_OPTIMIZE_BIT = 0x80 ///< flag, requiring cstr whitespace optimization |
| }; | }; |
| public: | public: |
| static const ConstStringPtr undefined_origin; | |
| String(Pool& apool, const char *src=0, size_t src_size=0, bool tainted=false); | //static String& OnPool(Pool& apool, const char *local_src=0, size_t src_size=0, bool tainted=false); |
| explicit String(const char *src=0, size_t src_size=0, bool tainted=false); | |
| String(const String& src); | String(const String& src); |
| bool is_empty() const { return !size(); } | |
| size_t size() const { return fsize; } | size_t size() const { return fsize; } |
| /// makes store_to produce origin:text\n output | |
| void set_origins_mode(bool aorigins_mode) const { forigins_mode=aorigins_mode; } | |
| /// convert to C string. if 'lang' known, forcing 'lang' to it | /// convert to C string. if 'lang' known, forcing 'lang' to it |
| char *cstr(Untaint_lang lang=UL_UNSPECIFIED, | CharPtr cstr(Untaint_lang lang=UL_AS_IS, |
| SQL_Connection *connection=0, | SQL_Connection *connection=0, |
| const char *charset=0) const { | Charset *cstr_charset=0, const char *cstr_charset_name=0) const { |
| char *result=(char *)malloc(cstr_bufsize(lang)); | CharPtr result(new char[cstr_bufsize(lang, connection, cstr_charset)]); |
| char *eol=store_to(result, lang, connection, charset); | char *eol=store_to(result.get(), lang, connection, cstr_charset, cstr_charset_name); |
| *eol=0; | *eol=0; |
| return result; | return result; |
| } | } |
| char *cstr_debug_origins() const; | |
| /// puts pieces to buf | |
| void serialize(Pool& pool, size_t prolog_size, char *& buf, size_t& buf_size) const; | |
| /// appends pieces from buf to self | |
| bool deserialize(size_t prolog_size, void *buf, size_t buf_size, const char *file); | |
| /** append fragment | /** append fragment |
| @see APPEND_AS_IS, APPEND_CLEAN, APPEND_TAINTED, APPEND_CONST | @see APPEND_AS_IS, APPEND_CLEAN, APPEND_TAINTED, APPEND_CONST |
| */ | */ |
| Line 157 public: | Line 165 public: |
| */ | */ |
| int cmp(int& partial, const char* src_ptr, size_t src_size=0, | int cmp(int& partial, const char* src_ptr, size_t src_size=0, |
| size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; | size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; |
| /// this starts with src | |
| bool starts_with(const char* src_ptr, size_t src_size=0) const { | |
| int p; cmp(p, src_ptr, src_size); | |
| return p==0 || p==2; | |
| } | |
| bool operator == (const char* src_ptr) const { | bool operator == (const char* src_ptr) const { |
| size_t src_size=src_ptr?strlen(src_ptr):0; | size_t src_size=src_ptr?strlen(src_ptr):0; |
| if(size() != src_size) | if(size() != src_size) |
| Line 169 public: | Line 182 public: |
| return cmp(partial, src_ptr, 0)!=0; | return cmp(partial, src_ptr, 0)!=0; |
| } | } |
| String& append(const String& src, uchar lang, bool forced=false); | |
| /** | |
| appends other String. | |
| marking all tainted pieces of it with @a lang. | |
| or marking ALL pieces of it with a @a lang when @a forced to. | |
| */ | |
| String& append(const String& src, Untaint_lang lang, bool forced=false); | |
| String& operator << (const String& src) { return append(src, UL_PASS_APPENDED); } | String& operator << (const String& src) { return append(src, UL_PASS_APPENDED); } |
| String& operator << (const char *src) { return APPEND_CONST(src); } | String& operator << (const char *src) { return APPEND_CONST(src); } |
| /// simple hash code of string. used by Hash | /// simple hash code of string. used by Hash |
| uint hash_code() const; | uint hash_code() const; |
| /// extracts first char of a string | /// extracts first char of a string, if any |
| char first_char() const; | char first_char() const { |
| return is_empty()?0:*felements[0].ptr; | |
| } | |
| /// extracts [start, finish) piece of string | /// extracts [start, finish) piece of string |
| String& mid(size_t start, size_t finish) const; | StringPtr mid(size_t start, size_t finish) const; |
| /// @return position of substr in string, -1 means "not found" [String version] | /// @return position of substr in string, -1 means "not found" [String version] |
| int pos(const String& substr, | int pos(const String& substr, |
| size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; | int this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; |
| /// @return position of substr in string, -1 means "not found" [const char* version] | /// @return position of substr in string, -1 means "not found" [const char* version] |
| int pos(const char *substr, size_t substr_size=0, | int pos(const char *substr, size_t substr_size=0, |
| size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; | int this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; |
| void split(Array& result, | void split(Array<ConstStringPtr>& result, |
| size_t *pos_after_ref, | size_t *pos_after_ref, |
| const char *delim, size_t delim_size, | const char *delim, size_t delim_size, |
| Untaint_lang lang=UL_UNSPECIFIED, int limit=-1) const; | Untaint_lang lang=UL_UNSPECIFIED, int limit=-1) const; |
| void split(Array& result, | void split(Array<ConstStringPtr>& result, |
| size_t *pos_after_ref, | size_t *pos_after_ref, |
| const String& delim, | const String& delim, |
| Untaint_lang lang=UL_UNSPECIFIED, int limit=-1) const; | Untaint_lang lang=UL_UNSPECIFIED, int limit=-1) const; |
| typedef void (*Row_action)(Table& table, Array *row, int start, int finish, | typedef void (*Row_action)(Table& table, Array<ConstStringPtr> *row, |
| int prestart, int prefinish, | |
| int poststart, int postfinish, | |
| void *info); | void *info); |
| /** | /** |
| @return true if fills table. | @return table of found items, if any. |
| table format is defined and fixed[can be used by others]: | table format is defined and fixed[can be used by others]: |
| @verbatim | @verbatim |
| prematch/match/postmatch/1/2/3/... | prematch/match/postmatch/1/2/3/... |
| @endverbatim | @endverbatim |
| */ | */ |
| bool match(const unsigned char *pcre_tables, | TablePtr match(Charset& source_charset, |
| const String *aorigin, | ConstStringPtr aorigin, |
| const String& regexp, | const String& regexp, |
| const String *options, | ConstStringPtr options, |
| Table **table, | |
| Row_action row_action, void *info, | Row_action row_action, void *info, |
| bool *was_global=0) const; | bool *was_global=0) const; |
| enum Change_case_kind { | enum Change_case_kind { |
| CC_UPPER, | CC_UPPER, |
| CC_LOWER | CC_LOWER |
| }; | }; |
| String& change_case(Pool& pool, const unsigned char *pcre_tables, | StringPtr change_case(Pool& pool, Charset& source_charset, |
| Change_case_kind kind) const; | Change_case_kind kind) const; |
| String& replace(Pool& pool, Dictionary& dict) const; | StringPtr replace(Pool& pool, Dictionary& dict) const; |
| double as_double() const; | double as_double() const; |
| int as_int() const; | int as_int() const; |
| #ifndef NO_STRING_ORIGIN | StringPtr join_chains(Pool& pool) const; |
| /// origin of string. calculated by first row | |
| const Origin& origin() const; | |
| #endif | |
| private: | |
| /// several String fragments | |
| struct Chunk { | |
| size_t count; ///< the number of rows in chunk | |
| /// string fragment or a link to next chunk union | |
| union Row { | |
| /// fragment | |
| struct { | |
| const char *ptr; ///< pointer to the start | |
| size_t size; ///< length | |
| Untaint_lang lang; ///< untaint flag, later untaint language | |
| #ifndef NO_STRING_ORIGIN | #ifndef NO_STRING_ORIGIN |
| Origin origin; ///< origin | /// origin of string. calculated by first row |
| const String_fragment::Origin& origin() const; | |
| #endif | #endif |
| } item; | |
| Chunk *link; ///< link to the next chunk in chain | |
| } rows[CR_PREALLOCATED_COUNT]; | |
| Chunk *preallocated_link; ///< next rows are here | |
| } | |
| head; ///< the head chunk of the chunk chain | |
| /// next append would write to this record | |
| Chunk::Row *append_here; | |
| /** the address of place where lies address | |
| of the link to the next chunk to allocate | |
| */ | |
| Chunk::Row *link_row; | |
| private: | |
| // last chunk | |
| Chunk *last_chunk; | |
| // string size | |
| size_t fsize; | |
| // used rows in all chunks | |
| int fused_rows; | |
| /// origins_mode changes behaviour of store_to[cstr()] | |
| mutable bool forigins_mode; | |
| private: | private: |
| bool chunk_is_full() { | Untaint_lang lang_of(size_t offset) const; |
| return append_here == link_row; | |
| } | |
| void expand(); | |
| size_t cstr_bufsize(Untaint_lang lang) const; | size_t cstr_bufsize(Untaint_lang lang=UL_UNSPECIFIED, |
| SQL_Connection *connection=0, | |
| Charset *buf_charset=0) const; | |
| /// convert to C string, store to 'dest' which must be big enough for proper untaint | /// convert to C string, store to 'dest' which must be big enough for proper untaint |
| char *store_to(char *dest, Untaint_lang lang=UL_UNSPECIFIED, | char *store_to(char *dest, Untaint_lang lang=UL_UNSPECIFIED, |
| SQL_Connection *connection=0, | SQL_Connection *connection=0, |
| const char *charset=0) const; | Charset *store_to_charset=0, |
| const char *store_to_charset_name=0) const; | |
| /* | |
| void join_chain( | |
| const Chunk*& achunk, const Chunk::Row*& arow, uint& acountdown, | |
| uchar& joined_lang, const char *& joined_ptr, size_t& joined_size) const; | |
| */ | |
| private: //disabled | private: //disabled |
| String& operator = (const String&) { return *this; } | String& operator = (const String&) { return *this; } |
| }; | }; |
| // statics | |
| const ConstStringPtr String::undefined_origin=ConstStringPtr(0); | |
| // helpers | |
| #define STRING_PREPARED_FOREACH_FRAGMENT(code) \ | |
| for(; fragment<last; fragment++) { \ | |
| code; \ | |
| } | |
| #define STRING_FOREACH_FRAGMENT(code) \ | |
| { \ | |
| element_type *last = felements+fused; \ | |
| element_type *fragment = felements; \ | |
| STRING_PREPARED_FOREACH_FRAGMENT(code) \ | |
| } | |
| #endif | #endif |