--- parser3/src/include/pa_string.h 2001/04/28 13:38:12 1.85 +++ parser3/src/include/pa_string.h 2001/07/18 10:06:04 1.98 @@ -5,24 +5,21 @@ Author: Alexander Petrosyan (http://design.ru/paf) - $Id: pa_string.h,v 1.85 2001/04/28 13:38:12 paf Exp $ + $Id: pa_string.h,v 1.98 2001/07/18 10:06:04 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 @@ -85,8 +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 @@ -117,11 +114,14 @@ 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; } + /// 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 char *cstr(Untaint_lang lang=UL_UNSPECIFIED, SQL_Connection *connection=0, const char *charset=0) const { - char *result=(char *)malloc(size()*UNTAINT_TIMES_BIGGER+1); + + char *result=(char *)malloc(cstr_bufsize(lang)); char *eol=store_to(result, lang, connection, charset); *eol=0; return result; @@ -187,26 +187,41 @@ public: int pos(const String& substr, size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; /// @return position of substr in string, -1 means "not found" [const char* version] - int pos(const char *substr, size_t substr_size, + int pos(const char *substr, size_t substr_size=0, size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const; 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); + /** + @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 @@ -253,6 +268,9 @@ private: // used rows in all chunks int fused_rows; + /// origins_mode changes behaviour of store_to[cstr()] + mutable bool forigins_mode; + private: bool chunk_is_full() { @@ -260,6 +278,7 @@ 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,