Annotation of parser3/src/include/pa_string.h, revision 1.109

1.41      paf         1: /** @file
1.43      paf         2:        Parser: string class decl.
                      3: 
1.29      paf         4:        Copyright (c) 2001 ArtLebedev Group (http://www.artlebedev.com)
1.30      paf         5:        Author: Alexander Petrosyan <paf@design.ru> (http://design.ru/paf)
1.29      paf         6: 
1.109   ! paf         7:        $Id: pa_string.h,v 1.108 2001/10/24 09:03:42 parser Exp $
1.1       paf         8: */
                      9: 
                     10: #ifndef PA_STRING_H
                     11: #define PA_STRING_H
                     12: 
1.49      paf        13: #include "pa_config_includes.h"
1.15      paf        14: #include "pa_pool.h"
1.4       paf        15: #include "pa_types.h"
                     16: 
1.63      paf        17: class Table;
                     18: 
1.50      paf        19: /**
1.92      parser     20:        $MAIN:html-typo table elements must enlarge string not more that that.
                     21: 
1.50      paf        22:        that's a tradeoff - otherwise we'd have to scan string twice:
                     23:        - first for buffer length
                     24:        - second for replacements themselves
                     25: */
1.92      parser     26: #define UNTAINT_TIMES_BIGGER 10
1.31      paf        27: 
1.9       paf        28: #ifndef NO_STRING_ORIGIN
1.33      paf        29: #      define STRING_APPEND_PARAMS \
                     30:                const char *src, size_t size,  \
1.52      paf        31:                String::Untaint_lang lang, \
1.33      paf        32:                const char *file, uint line
1.54      paf        33: /// appends piece to String  @see String::real_append
                     34: #      define APPEND(src, size, lang, file, line) \
                     35:                real_append(src, size, lang, file, line)
1.9       paf        36: #else
1.33      paf        37: #      define STRING_APPEND_PARAMS \
                     38:                const char *src, \
                     39:                size_t size, \
1.52      paf        40:                String::Untaint_lang lang
1.54      paf        41: /// appends piece to String  @see String::real_append
                     42: #      define APPEND(src, size, lang, file, line) \
                     43:                real_append(src, size, lang)
                     44: #endif
1.83      paf        45: /// appends clean piece to String @see String::real_append
1.54      paf        46: #define APPEND_CLEAN(src, size, file, line) \
1.62      paf        47:        APPEND(src, size, String::UL_CLEAN, file, line)
1.83      paf        48: /// appends piece to String as-is @see String::real_append
                     49: #define APPEND_AS_IS(src, size, file, line) \
                     50:        APPEND(src, size, String::UL_AS_IS, file, line)
1.48      paf        51: /// appends tainted piece to String  @see String::real_append
1.54      paf        52: #define APPEND_TAINTED(src, size, file, line) \
1.62      paf        53:        APPEND(src, size, String::UL_TAINTED, file, line)
1.48      paf        54: /// handy: appends const char* piece to String  @see String::real_append
1.83      paf        55: #define        APPEND_CONST(src) APPEND_AS_IS(src, 0, 0, 0)
1.9       paf        56: 
1.62      paf        57: class Array;
1.71      paf        58: class SQL_Connection;
1.101     parser     59: class Dictionary;
1.62      paf        60: 
1.42      paf        61: /** 
1.41      paf        62:        Pooled string.
                     63: 
1.47      paf        64:        Internal structure:
                     65:        @verbatim
                     66:                String                          Chunk0
                     67:                ======                          ========
                     68:                head--------------->[ptr, size, ...]
                     69:                append_here-------->[ptr, size, ...]
                     70:                                                        .
                     71:                                                        .
                     72:                                                        [ptr, size, ...]
                     73:                link_row----------->[link to the next chunk]
1.41      paf        74:        @endverbatim
                     75: 
                     76:        All pieces remember 
                     77:        - the file and its line they are from [can be turned off by NO_STRING_ORIGIN]
                     78:        - whether they are tainted or not, 
                     79:          and the language which should be used to detaint them
                     80: */
1.15      paf        81: class String : public Pooled {
1.1       paf        82: public:
1.48      paf        83: 
1.1       paf        84:        enum {
1.89      parser     85:                CR_PREALLOCATED_COUNT=2, ///< default preallocated item count
1.91      parser     86:                CR_GROW_COUNT=2 ///< each time the String chunk_is_full() string expanded()
1.1       paf        87:        };
                     88: 
1.106     parser     89:        /** piece is tainted or not. the language to use when detaint
                     90:                remember to change String_Untaint_lang_name @ untaint.C along
                     91:        */
1.27      paf        92:        enum Untaint_lang {
1.72      paf        93:                UL_UNSPECIFIED=0, ///< zero value handy for hash lookup @see untaint_lang_name2enum
1.62      paf        94:                UL_CLEAN, ///< clean
                     95:                UL_TAINTED,  ///< tainted, untaint language as assigned later 
1.27      paf        96:                // untaint languages. assigned by ^untaint[lang]{...}
1.48      paf        97:                UL_PASS_APPENDED,
1.41      paf        98:                        /**<
                     99:                                leave language built into string being appended.
                    100:                                just a flag, that value not stored
                    101:                        */
1.48      paf       102:                UL_AS_IS,     ///< leave all characters intact
1.103     parser    103:                UL_FILE_SPEC, ///< file specification
1.74      paf       104:                UL_HTTP_HEADER,    ///< text in HTTP response header
                    105:                UL_MAIL_HEADER,    ///< text in mail header
1.48      paf       106:                UL_URI,       ///< text in uri
                    107:                UL_TABLE,     ///< ^table:set body
                    108:                UL_SQL,       ///< ^table:sql body
                    109:                UL_JS,        ///< JavaScript code
1.102     parser    110:                UL_XML,         ///< ^dom:set xml
1.48      paf       111:                UL_HTML,      ///< HTML code (for editing)
1.85      paf       112:                UL_USER_HTML  ///< HTML code with USER chars
1.27      paf       113:        };
                    114: 
1.8       paf       115: public:
                    116: 
1.77      paf       117:        String(Pool& apool, const char *src=0, size_t src_size=0, bool tainted=false);
1.14      paf       118:        String(const String& src);
                    119:        size_t size() const { return fsize; }
1.98      parser    120:        /// makes store_to produce origin:text\n output
                    121:        void set_origins_mode(bool aorigins_mode) const { forigins_mode=aorigins_mode; }
1.69      paf       122:        /// convert to C string. if 'lang' known, forcing 'lang' to it
1.109   ! paf       123:        char *cstr(Untaint_lang lang=UL_AS_IS, 
1.82      paf       124:                SQL_Connection *connection=0,
                    125:                const char *charset=0) const {
1.88      parser    126: 
1.94      parser    127:                char *result=(char *)malloc(cstr_bufsize(lang));
1.82      paf       128:                char *eol=store_to(result, lang, connection, charset);
1.50      paf       129:                *eol=0;
                    130:                return result;
                    131:        }
1.108     parser    132:        /// puts pieces to buf
                    133:        void serialize(size_t prolog_size,  void *& buf, size_t& buf_size) const;
                    134:        /// appends pieces from buf to self
                    135:        void deserialize(size_t prolog_size, void *buf, size_t buf_size, const char *file);
1.46      paf       136:        /** append fragment
1.83      paf       137:                @see APPEND_AS_IS, APPEND_CLEAN, APPEND_TAINTED, APPEND_CONST
1.46      paf       138:        */
1.9       paf       139:        String& real_append(STRING_APPEND_PARAMS);
1.44      paf       140:        /// @return <0 ==0 or >0 depending on comparison result
1.62      paf       141:        int cmp (int& partial, const String& src, 
1.72      paf       142:                size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const;
1.56      paf       143:        bool operator < (const String& src) const {     int p; return cmp(p, src)<0; }
                    144:        bool operator > (const String& src) const {     int p; return cmp(p, src)>0; }
                    145:        bool operator <= (const String& src) const { int p; return cmp(p, src)<=0; }
                    146:        bool operator >= (const String& src) const { int p; return cmp(p, src)>=0; }
1.26      paf       147:        bool operator == (const String& src) const {
                    148:                if(size()!=src.size()) // can speed up in trivial case
                    149:                        return false;
1.56      paf       150:                int p; return cmp(p, src)==0;
1.26      paf       151:        }
1.56      paf       152:        bool operator != (const String& src) const { int p; return cmp(p, src)!=0; }
1.26      paf       153: 
1.50      paf       154:        /**
                    155:                 @param partial 
                    156:                        returns partial match status. 
1.51      paf       157:                        - -1: strings too different
                    158:                        -  0: full match
                    159:                        -  1: means @c this starts @c src
                    160:                        -  2: means @src starts @this
1.50      paf       161:        */
1.61      paf       162:        int cmp(int& partial, const char* src_ptr, size_t src_size=0, 
1.72      paf       163:                size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const;
1.50      paf       164:        bool operator == (const char* src_ptr) const { 
                    165:                size_t src_size=src_ptr?strlen(src_ptr):0;
                    166:                if(size() != src_size)
                    167:                        return false;
                    168:                int partial; // unused
1.56      paf       169:                return cmp(partial, src_ptr, src_size)==0; 
1.50      paf       170:        }
1.80      paf       171:        bool operator != (const char* src_ptr) const { 
                    172:                int partial; // unused
                    173:                return cmp(partial, src_ptr, 0)!=0; 
                    174:        }
                    175: 
1.50      paf       176: 
1.42      paf       177:        /** 
                    178:                appends other String.
1.41      paf       179: 
1.47      paf       180:                marking all tainted pieces of it with @a lang.
                    181:                or marking ALL pieces of it with a @a lang when @a forced to.
1.41      paf       182:        */
1.39      paf       183:        String& append(const String& src, Untaint_lang lang, bool forced=false);
1.76      paf       184:        String& operator << (const String& src) { return append(src, UL_PASS_APPENDED); }
                    185:        String& operator << (const char *src) { return APPEND_CONST(src); }
1.8       paf       186: 
1.41      paf       187:        /// simple hash code of string. used by Hash
1.14      paf       188:        uint hash_code() const;
1.100     parser    189: 
                    190:        /// extracts first char of a string
                    191:        char first_char() const;
1.54      paf       192: 
                    193:        /// extracts [start, finish) piece of string
1.70      paf       194:        String& mid(size_t start, size_t finish) const;
1.55      paf       195: 
1.59      paf       196:        /// @return position of substr in string, -1 means "not found" [String version]
1.62      paf       197:        int pos(const String& substr, 
1.72      paf       198:                size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const;
1.59      paf       199:        /// @return position of substr in string, -1 means "not found" [const char* version]
1.97      parser    200:        int pos(const char *substr, size_t substr_size=0, 
1.72      paf       201:                size_t this_offset=0, Untaint_lang lang=UL_UNSPECIFIED) const;
1.62      paf       202: 
                    203:        void split(Array& result, 
                    204:                size_t *pos_after_ref, 
                    205:                const char *delim, size_t delim_size, 
1.86      paf       206:                Untaint_lang lang=UL_UNSPECIFIED, int limit=-1) const;
1.62      paf       207:        void split(Array& result, 
                    208:                size_t *pos_after_ref, 
                    209:                const String& delim, 
1.86      paf       210:                Untaint_lang lang=UL_UNSPECIFIED, int limit=-1) const;
1.62      paf       211: 
1.68      paf       212:        typedef void (*Row_action)(Table& table, Array *row, int start, int finish, 
                    213:                void *info);
1.87      parser    214:        /**
                    215:                @return true if fills table.
                    216:                table format is defined and fixed[can be used by others]: 
                    217:                @verbatim
                    218:                        prematch/match/postmatch/1/2/3/...
                    219:                @endverbatim
                    220:        */
1.81      paf       221:        bool match(const unsigned char *pcre_tables,
                    222:                const String *aorigin,          
1.64      paf       223:                const String& regexp, 
1.65      paf       224:                const String *options,
1.66      paf       225:                Table **table,
1.99      parser    226:                Row_action row_action, void *info,
                    227:                bool *was_global=0) const;
1.87      parser    228:        enum Change_case_kind {
                    229:                CC_UPPER,
                    230:                CC_LOWER
                    231:        };
                    232:        String& change_case(Pool& pool, const unsigned char *pcre_tables, 
                    233:                Change_case_kind kind) const;
1.101     parser    234:        String& replace(Pool& pool, Dictionary& dict) const;
1.96      parser    235:        double as_double() const;
                    236:        int as_int() const;
1.13      paf       237: 
1.41      paf       238: #ifndef NO_STRING_ORIGIN
                    239:        /// origin of string. calculated by first row
1.43      paf       240:        const Origin& origin() const;
1.41      paf       241: #endif
1.8       paf       242: 
1.10      paf       243: private:
                    244: 
1.84      paf       245:        /// several String fragments
1.1       paf       246:        struct Chunk {
1.84      paf       247:                size_t count; ///< the number of rows in chunk
                    248:                /// string fragment or a link to next chunk union
1.1       paf       249:                union Row {
1.84      paf       250:                        /// fragment
1.27      paf       251:                        struct { 
1.84      paf       252:                                const char *ptr;  ///< pointer to the start
                    253:                                size_t size;  ///< length
                    254:                                Untaint_lang lang; ///< untaint flag, later untaint language
1.27      paf       255: #ifndef NO_STRING_ORIGIN
1.84      paf       256:                                Origin origin;  ///< origin
1.27      paf       257: #endif
1.1       paf       258:                        } item;
1.84      paf       259:                        Chunk *link;  ///< link to the next chunk in chain
1.2       paf       260:                } rows[CR_PREALLOCATED_COUNT];
1.84      paf       261:                Chunk *preallocated_link; ///< next rows are here
1.1       paf       262:        }
1.84      paf       263:                head;  ///< the head chunk of the chunk chain
1.1       paf       264: 
1.84      paf       265:        /// next append would write to this record
1.1       paf       266:        Chunk::Row *append_here;
                    267:        
1.84      paf       268:        /** the address of place where lies address 
                    269:                of the link to the next chunk to allocate
                    270:        */
1.1       paf       271:        Chunk::Row *link_row;
                    272: 
1.5       paf       273: private:
1.25      paf       274:        // last chunk
                    275:        Chunk *last_chunk;
1.5       paf       276: 
                    277:        // string size
                    278:        size_t fsize;
                    279: 
                    280:        // used rows in all chunks
                    281:        int fused_rows;
1.98      parser    282: 
                    283:        /// origins_mode changes behaviour of store_to[cstr()]
                    284:        mutable bool forigins_mode;
1.5       paf       285: 
                    286: private:
1.1       paf       287: 
                    288:        bool chunk_is_full() {
                    289:                return append_here == link_row;
                    290:        }
                    291:        void expand();
1.79      paf       292: 
1.94      parser    293:        size_t cstr_bufsize(Untaint_lang lang) const;
1.79      paf       294:        /// convert to C string, store to 'dest' which must be big enough for proper untaint
1.82      paf       295:        char *store_to(char *dest, Untaint_lang lang=UL_UNSPECIFIED, 
                    296:                SQL_Connection *connection=0,
                    297:                const char *charset=0) const;
1.107     parser    298: 
                    299:        String& reconstruct(Pool& pool) const;
                    300:        void join_chain(Pool& pool, 
                    301:                                           size_t& ai, const Chunk*& achunk, const Chunk::Row*& arow,
                    302:                                           Untaint_lang& joined_lang, const char *& joined_ptr, size_t& joined_size) const;
                    303:        String& replace_in_reconstructed(Pool& pool, Dictionary& dict) const;
1.7       paf       304: 
                    305: private: //disabled
                    306: 
1.12      paf       307:        String& operator = (const String&) { return *this; }
1.7       paf       308: 
1.1       paf       309: };
                    310: 
                    311: #endif

E-mail: