--- parser3/src/include/pa_string.h 2001/04/09 14:31:40 1.78 +++ parser3/src/include/pa_string.h 2001/05/21 16:39:32 1.96 @@ -5,24 +5,21 @@ Author: Alexander Petrosyan (http://design.ru/paf) - $Id: pa_string.h,v 1.78 2001/04/09 14:31:40 paf Exp $ + $Id: pa_string.h,v 1.96 2001/05/21 16:39:32 parser Exp $ */ #ifndef PA_STRING_H #define PA_STRING_H #include "pa_config_includes.h" - -#include -#include - #include "pa_pool.h" #include "pa_types.h" class Table; /** - $MAIN:html-typo table elements must enlarge string not more that that + $MAIN:html-typo table elements must enlarge string not more that that. + that's a tradeoff - otherwise we'd have to scan string twice: - first for buffer length - second for replacements themselves @@ -46,14 +43,17 @@ class Table; # define APPEND(src, size, lang, file, line) \ real_append(src, size, lang) #endif -/// appends clean piece to String @see String::real_append +/// appends clean piece to String @see String::real_append #define APPEND_CLEAN(src, size, file, line) \ APPEND(src, size, String::UL_CLEAN, file, line) +/// appends piece to String as-is @see String::real_append +#define APPEND_AS_IS(src, size, file, line) \ + APPEND(src, size, String::UL_AS_IS, file, line) /// appends tainted piece to String @see String::real_append #define APPEND_TAINTED(src, size, file, line) \ APPEND(src, size, String::UL_TAINTED, file, line) /// handy: appends const char* piece to String @see String::real_append -#define APPEND_CONST(src) APPEND_CLEAN(src, 0, 0, 0) +#define APPEND_CONST(src) APPEND_AS_IS(src, 0, 0, 0) class Array; class SQL_Connection; @@ -82,8 +82,8 @@ class String : public Pooled { public: enum { - CR_PREALLOCATED_COUNT=5, ///< default preallocated item count - CR_GROW_PERCENT=60 ///< each time the Array chunk_is_full() array expanded() + CR_PREALLOCATED_COUNT=2, ///< default preallocated item count + CR_GROW_COUNT=2 ///< each time the String chunk_is_full() string expanded() }; /// piece is tainted or not. the language to use when detaint @@ -106,7 +106,7 @@ public: UL_SQL, ///< ^table:sql body UL_JS, ///< JavaScript code UL_HTML, ///< HTML code (for editing) - UL_HTML_TYPO ///< HTML code with TYPOgraphic replacements (for showing) + UL_USER_HTML ///< HTML code with USER chars }; public: @@ -114,18 +114,18 @@ public: String(Pool& apool, const char *src=0, size_t src_size=0, bool tainted=false); String(const String& src); size_t size() const { return fsize; } - /// 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, SQL_Connection *connection=0) const; /// convert to C string. if 'lang' known, forcing 'lang' to it - char *cstr(Untaint_lang lang=UL_UNSPECIFIED, SQL_Connection *connection=0) const { - char *result=(char *)malloc(size()*UNTAINT_TIMES_BIGGER+1); - char *eol=store_to(result, lang, connection); + char *cstr(Untaint_lang lang=UL_UNSPECIFIED, + SQL_Connection *connection=0, + const char *charset=0) const { + + char *result=(char *)malloc(cstr_bufsize(lang)); + char *eol=store_to(result, lang, connection, charset); *eol=0; return result; } /** append fragment - @see APPEND_CLEAN, APPEND_TAINTED, APPEND_CONST + @see APPEND_AS_IS, APPEND_CLEAN, APPEND_TAINTED, APPEND_CONST */ String& real_append(STRING_APPEND_PARAMS); /// @return <0 ==0 or >0 depending on comparison result @@ -159,6 +159,11 @@ public: int partial; // unused return cmp(partial, src_ptr, src_size)==0; } + bool operator != (const char* src_ptr) const { + int partial; // unused + return cmp(partial, src_ptr, 0)!=0; + } + /** appends other String. @@ -186,19 +191,35 @@ public: void split(Array& result, size_t *pos_after_ref, const char *delim, size_t delim_size, - Untaint_lang lang, int limit=-1) const; + Untaint_lang lang=UL_UNSPECIFIED, int limit=-1) const; void split(Array& result, size_t *pos_after_ref, const String& delim, - Untaint_lang lang, 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, void *info); - bool match(const String *aorigin, + /** + @return true if fills table. + table format is defined and fixed[can be used by others]: + @verbatim + prematch/match/postmatch/1/2/3/... + @endverbatim + */ + bool match(const unsigned char *pcre_tables, + const String *aorigin, const String& regexp, const String *options, Table **table, Row_action row_action, void *info) const; + enum Change_case_kind { + CC_UPPER, + CC_LOWER + }; + String& change_case(Pool& pool, const unsigned char *pcre_tables, + Change_case_kind kind) const; + double as_double() const; + int as_int() const; #ifndef NO_STRING_ORIGIN /// origin of string. calculated by first row @@ -207,31 +228,32 @@ public: private: + /// several String fragments struct Chunk { - // the number of rows in chunk - size_t count; + size_t count; ///< the number of rows in chunk + /// string fragment or a link to next chunk union union Row { - // fragment + /// fragment struct { - const char *ptr; // pointer to the start - size_t size; // length - Untaint_lang lang; // untaint flag, later untaint language + const char *ptr; ///< pointer to the start + size_t size; ///< length + Untaint_lang lang; ///< untaint flag, later untaint language #ifndef NO_STRING_ORIGIN - Origin origin; // origin + Origin origin; ///< origin #endif } item; - Chunk *link; // link to the next chunk in chain + Chunk *link; ///< link to the next chunk in chain } rows[CR_PREALLOCATED_COUNT]; - // next rows are here - Chunk *preallocated_link; + Chunk *preallocated_link; ///< next rows are here } - head; // the head chunk of the chunk chain + head; ///< the head chunk of the chunk chain - // next append would write to this record + /// 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 + /** the address of place where lies address + of the link to the next chunk to allocate + */ Chunk::Row *link_row; private: @@ -251,6 +273,12 @@ private: } void expand(); + size_t cstr_bufsize(Untaint_lang lang) const; + /// 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, + SQL_Connection *connection=0, + const char *charset=0) const; + private: //disabled String& operator = (const String&) { return *this; }