| 1 | /* |
| 2 | Copyright (c) 2000, 2012, Oracle and/or its affiliates. |
| 3 | |
| 4 | This program is free software; you can redistribute it and/or modify |
| 5 | it under the terms of the GNU General Public License as published by |
| 6 | the Free Software Foundation; version 2 of the License. |
| 7 | |
| 8 | This program is distributed in the hope that it will be useful, |
| 9 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 10 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 11 | GNU General Public License for more details. |
| 12 | |
| 13 | You should have received a copy of the GNU General Public License |
| 14 | along with this program; if not, write to the Free Software |
| 15 | Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ |
| 16 | |
| 17 | /** |
| 18 | @file |
| 19 | |
| 20 | @details |
| 21 | @verbatim |
| 22 | The idea of presented algorithm see in |
| 23 | "The Art of Computer Programming" by Donald E. Knuth |
| 24 | Volume 3 "Sorting and searching" |
| 25 | (chapter 6.3 "Digital searching" - name and number of chapter |
| 26 | is back translation from Russian edition :)) |
| 27 | |
| 28 | as illustration of data structures, imagine next table: |
| 29 | |
| 30 | static SYMBOL symbols[] = { |
| 31 | { "ADD", SYM(ADD),0,0}, |
| 32 | { "AND", SYM(AND),0,0}, |
| 33 | { "DAY", SYM(DAY_SYM),0,0}, |
| 34 | }; |
| 35 | |
| 36 | for this structure, presented program generate next searching-structure: |
| 37 | |
| 38 | +-----------+-+-+-+ |
| 39 | | len |1|2|3| |
| 40 | +-----------+-+-+-+ |
| 41 | |first_char |0|0|a| |
| 42 | |last_char |0|0|d| |
| 43 | |link |0|0|+| |
| 44 | | |
| 45 | V |
| 46 | +----------+-+-+-+--+ |
| 47 | | 1 char|a|b|c|d | |
| 48 | +----------+-+-+-+--+ |
| 49 | |first_char|b|0|0|0 | |
| 50 | |last_char |n|0|0|-1| |
| 51 | |link |+|0|0|+ | |
| 52 | | | |
| 53 | | V |
| 54 | | symbols[2] ( "DAY" ) |
| 55 | V |
| 56 | +----------+--+-+-+-+-+-+-+-+-+-+--+ |
| 57 | | 2 char|d |e|f|j|h|i|j|k|l|m|n | |
| 58 | +----------+--+-+-+-+-+-+-+-+-+-+--+ |
| 59 | |first_char|0 |0|0|0|0|0|0|0|0|0|0 | |
| 60 | |last_char |-1|0|0|0|0|0|0|0|0|0|-1| |
| 61 | |link |+ |0|0|0|0|0|0|0|0|0|+ | |
| 62 | | | |
| 63 | V V |
| 64 | symbols[0] ( "ADD" ) symbols[1] ( "AND" ) |
| 65 | |
| 66 | for optimization, link is the 16-bit index in 'symbols' or 'sql_functions' |
| 67 | or search-array.. |
| 68 | |
| 69 | So, we can read full search-structure as 32-bit word |
| 70 | @endverbatim |
| 71 | |
| 72 | @todo |
| 73 | use instead to_upper_lex, special array |
| 74 | (substitute chars) without skip codes.. |
| 75 | @todo |
| 76 | try use reverse order of comparing.. |
| 77 | |
| 78 | */ |
| 79 | |
| 80 | #define NO_YACC_SYMBOLS |
| 81 | #undef CHECK_UNLIKELY |
| 82 | #include "mariadb.h" |
| 83 | #include "mysql_version.h" |
| 84 | #include "lex.h" |
| 85 | #include <string.h> |
| 86 | |
| 87 | #include <welcome_copyright_notice.h> /* ORACLE_WELCOME_COPYRIGHT_NOTICE */ |
| 88 | |
| 89 | struct hash_lex_struct |
| 90 | { |
| 91 | int first_char; |
| 92 | char last_char; |
| 93 | union{ |
| 94 | hash_lex_struct *char_tails; |
| 95 | int iresult; |
| 96 | }; |
| 97 | int ithis; |
| 98 | }; |
| 99 | |
| 100 | hash_lex_struct *get_hash_struct_by_len(hash_lex_struct **root_by_len, |
| 101 | int len, int *max_len) |
| 102 | { |
| 103 | if (*max_len<len){ |
| 104 | *root_by_len= (hash_lex_struct *)realloc((char*)*root_by_len, |
| 105 | sizeof(hash_lex_struct)*len); |
| 106 | hash_lex_struct *cur, *end= *root_by_len + len; |
| 107 | for (cur= *root_by_len + *max_len; cur<end; cur++) |
| 108 | cur->first_char= 0; |
| 109 | *max_len= len; |
| 110 | } |
| 111 | return (*root_by_len)+(len-1); |
| 112 | } |
| 113 | |
| 114 | void insert_into_hash(hash_lex_struct *root, const char *name, |
| 115 | int len_from_begin, int index, int function) |
| 116 | { |
| 117 | hash_lex_struct *end, *cur, *tails; |
| 118 | |
| 119 | if (!root->first_char) |
| 120 | { |
| 121 | root->first_char= -1; |
| 122 | root->iresult= index; |
| 123 | return; |
| 124 | } |
| 125 | |
| 126 | if (root->first_char == -1) |
| 127 | { |
| 128 | int index2= root->iresult; |
| 129 | const char *name2= (index2 < 0 ? sql_functions[-index2-1] : |
| 130 | symbols[index2]).name + len_from_begin; |
| 131 | root->first_char= (int) (uchar) name2[0]; |
| 132 | root->last_char= (char) root->first_char; |
| 133 | tails= (hash_lex_struct*)malloc(sizeof(hash_lex_struct)); |
| 134 | root->char_tails= tails; |
| 135 | tails->first_char= -1; |
| 136 | tails->iresult= index2; |
| 137 | } |
| 138 | |
| 139 | size_t real_size= (root->last_char-root->first_char+1); |
| 140 | |
| 141 | if (root->first_char>(*name)) |
| 142 | { |
| 143 | size_t new_size= root->last_char-(*name)+1; |
| 144 | if (unlikely(new_size<real_size)) |
| 145 | printf("error!!!!\n" ); |
| 146 | tails= root->char_tails; |
| 147 | tails= (hash_lex_struct*)realloc((char*)tails, |
| 148 | sizeof(hash_lex_struct)*new_size); |
| 149 | root->char_tails= tails; |
| 150 | memmove(tails+(new_size-real_size),tails,real_size*sizeof(hash_lex_struct)); |
| 151 | end= tails + new_size - real_size; |
| 152 | for (cur= tails; cur<end; cur++) |
| 153 | cur->first_char= 0; |
| 154 | root->first_char= (int) (uchar) *name; |
| 155 | } |
| 156 | |
| 157 | if (root->last_char<(*name)) |
| 158 | { |
| 159 | size_t new_size= (*name)-root->first_char+1; |
| 160 | if (unlikely(new_size<real_size)) |
| 161 | printf("error!!!!\n" ); |
| 162 | tails= root->char_tails; |
| 163 | tails= (hash_lex_struct*)realloc((char*)tails, |
| 164 | sizeof(hash_lex_struct)*new_size); |
| 165 | root->char_tails= tails; |
| 166 | end= tails + new_size; |
| 167 | for (cur= tails+real_size; cur<end; cur++) |
| 168 | cur->first_char= 0; |
| 169 | root->last_char= (*name); |
| 170 | } |
| 171 | |
| 172 | insert_into_hash(root->char_tails+(*name)-root->first_char, |
| 173 | name+1,len_from_begin+1,index,function); |
| 174 | } |
| 175 | |
| 176 | |
| 177 | hash_lex_struct *root_by_len= 0; |
| 178 | int max_len=0; |
| 179 | |
| 180 | hash_lex_struct *root_by_len2= 0; |
| 181 | int max_len2=0; |
| 182 | |
| 183 | void insert_symbols() |
| 184 | { |
| 185 | size_t i= 0; |
| 186 | SYMBOL *cur; |
| 187 | for (cur= symbols; i<array_elements(symbols); cur++, i++){ |
| 188 | hash_lex_struct *root= |
| 189 | get_hash_struct_by_len(&root_by_len,cur->length,&max_len); |
| 190 | insert_into_hash(root,cur->name,0,(uint) i,0); |
| 191 | } |
| 192 | } |
| 193 | |
| 194 | void insert_sql_functions() |
| 195 | { |
| 196 | int i= 0; |
| 197 | SYMBOL *cur; |
| 198 | for (cur= sql_functions; i < (int) array_elements(sql_functions); cur++, i++) |
| 199 | { |
| 200 | hash_lex_struct *root= |
| 201 | get_hash_struct_by_len(&root_by_len,cur->length,&max_len); |
| 202 | insert_into_hash(root,cur->name,0,-i-1,1); |
| 203 | } |
| 204 | } |
| 205 | |
| 206 | void calc_length() |
| 207 | { |
| 208 | SYMBOL *cur, *end= symbols + array_elements(symbols); |
| 209 | for (cur= symbols; cur < end; cur++) |
| 210 | cur->length=(uchar) strlen(cur->name); |
| 211 | end= sql_functions + array_elements(sql_functions); |
| 212 | for (cur= sql_functions; cur<end; cur++) |
| 213 | cur->length=(uchar) strlen(cur->name); |
| 214 | } |
| 215 | |
| 216 | void generate_find_structs() |
| 217 | { |
| 218 | root_by_len= 0; |
| 219 | max_len=0; |
| 220 | |
| 221 | insert_symbols(); |
| 222 | |
| 223 | root_by_len2= root_by_len; |
| 224 | max_len2= max_len; |
| 225 | |
| 226 | root_by_len= 0; |
| 227 | max_len= 0; |
| 228 | |
| 229 | insert_symbols(); |
| 230 | insert_sql_functions(); |
| 231 | } |
| 232 | |
| 233 | char *hash_map= 0; |
| 234 | int size_hash_map= 0; |
| 235 | |
| 236 | void add_struct_to_map(hash_lex_struct *st) |
| 237 | { |
| 238 | st->ithis= size_hash_map/4; |
| 239 | size_hash_map+= 4; |
| 240 | hash_map= (char*)realloc((char*)hash_map,size_hash_map); |
| 241 | hash_map[size_hash_map-4]= (char) (st->first_char == -1 ? 0 : |
| 242 | st->first_char); |
| 243 | hash_map[size_hash_map-3]= (char) (st->first_char == -1 || |
| 244 | st->first_char == 0 ? 0 : st->last_char); |
| 245 | if (st->first_char == -1) |
| 246 | { |
| 247 | hash_map[size_hash_map-2]= ((unsigned int)(int16)st->iresult)&255; |
| 248 | hash_map[size_hash_map-1]= ((unsigned int)(int16)st->iresult)>>8; |
| 249 | } |
| 250 | else if (st->first_char == 0) |
| 251 | { |
| 252 | hash_map[size_hash_map-2]= ((unsigned int)(int16)array_elements(symbols))&255; |
| 253 | hash_map[size_hash_map-1]= ((unsigned int)(int16)array_elements(symbols))>>8; |
| 254 | } |
| 255 | } |
| 256 | |
| 257 | |
| 258 | void add_structs_to_map(hash_lex_struct *st, int len) |
| 259 | { |
| 260 | hash_lex_struct *cur, *end= st+len; |
| 261 | for (cur= st; cur<end; cur++) |
| 262 | add_struct_to_map(cur); |
| 263 | for (cur= st; cur<end; cur++) |
| 264 | { |
| 265 | if (cur->first_char && cur->first_char != -1) |
| 266 | add_structs_to_map(cur->char_tails,cur->last_char-cur->first_char+1); |
| 267 | } |
| 268 | } |
| 269 | |
| 270 | void set_links(hash_lex_struct *st, int len) |
| 271 | { |
| 272 | hash_lex_struct *cur, *end= st+len; |
| 273 | for (cur= st; cur<end; cur++) |
| 274 | { |
| 275 | if (cur->first_char != 0 && cur->first_char != -1) |
| 276 | { |
| 277 | int ilink= cur->char_tails->ithis; |
| 278 | hash_map[cur->ithis*4+2]= ilink%256; |
| 279 | hash_map[cur->ithis*4+3]= ilink/256; |
| 280 | set_links(cur->char_tails,cur->last_char-cur->first_char+1); |
| 281 | } |
| 282 | } |
| 283 | } |
| 284 | |
| 285 | |
| 286 | void print_hash_map(const char *name) |
| 287 | { |
| 288 | char *cur; |
| 289 | int i; |
| 290 | |
| 291 | printf("static uchar %s[%d]= {\n" ,name,size_hash_map); |
| 292 | for (i=0, cur= hash_map; i<size_hash_map; i++, cur++) |
| 293 | { |
| 294 | switch(i%4){ |
| 295 | case 0: case 1: |
| 296 | if (!*cur) |
| 297 | printf("0, " ); |
| 298 | else |
| 299 | printf("\'%c\', " ,*cur); |
| 300 | break; |
| 301 | case 2: printf("%u, " ,(uint)(uchar)*cur); break; |
| 302 | case 3: printf("%u,\n" ,(uint)(uchar)*cur); break; |
| 303 | } |
| 304 | } |
| 305 | printf("};\n" ); |
| 306 | } |
| 307 | |
| 308 | |
| 309 | void print_find_structs() |
| 310 | { |
| 311 | add_structs_to_map(root_by_len,max_len); |
| 312 | set_links(root_by_len,max_len); |
| 313 | print_hash_map("sql_functions_map" ); |
| 314 | free(hash_map); |
| 315 | |
| 316 | hash_map= 0; |
| 317 | size_hash_map= 0; |
| 318 | |
| 319 | printf("\n" ); |
| 320 | |
| 321 | add_structs_to_map(root_by_len2,max_len2); |
| 322 | set_links(root_by_len2,max_len2); |
| 323 | print_hash_map("symbols_map" ); |
| 324 | free(hash_map); |
| 325 | } |
| 326 | |
| 327 | |
| 328 | int check_dup_symbols(SYMBOL *s1, SYMBOL *s2) |
| 329 | { |
| 330 | if (s1->length!=s2->length || strncmp(s1->name,s2->name,s1->length)) |
| 331 | return 0; |
| 332 | |
| 333 | const char *err_tmpl= "\ngen_lex_hash fatal error : \ |
| 334 | Unfortunately gen_lex_hash can not generate a hash,\n since \ |
| 335 | your lex.h has duplicate definition for a symbol \"%s\"\n\n" ; |
| 336 | printf (err_tmpl,s1->name); |
| 337 | fprintf (stderr,err_tmpl,s1->name); |
| 338 | |
| 339 | return 1; |
| 340 | } |
| 341 | |
| 342 | |
| 343 | int check_duplicates() |
| 344 | { |
| 345 | SYMBOL *cur1, *cur2, *s_end, *f_end; |
| 346 | |
| 347 | s_end= symbols + array_elements(symbols); |
| 348 | f_end= sql_functions + array_elements(sql_functions); |
| 349 | |
| 350 | for (cur1= symbols; cur1<s_end; cur1++) |
| 351 | { |
| 352 | for (cur2= cur1+1; cur2<s_end; cur2++) |
| 353 | { |
| 354 | if (check_dup_symbols(cur1,cur2)) |
| 355 | return 1; |
| 356 | } |
| 357 | for (cur2= sql_functions; cur2<f_end; cur2++) |
| 358 | { |
| 359 | if (check_dup_symbols(cur1,cur2)) |
| 360 | return 1; |
| 361 | } |
| 362 | } |
| 363 | |
| 364 | for (cur1= sql_functions; cur1<f_end; cur1++) |
| 365 | { |
| 366 | for (cur2= cur1+1; cur2< f_end; cur2++) |
| 367 | { |
| 368 | if (check_dup_symbols(cur1,cur2)) |
| 369 | return 1; |
| 370 | } |
| 371 | } |
| 372 | return 0; |
| 373 | } |
| 374 | |
| 375 | |
| 376 | int main(int argc,char **argv) |
| 377 | { |
| 378 | |
| 379 | |
| 380 | /* Broken up to indicate that it's not advice to you, gentle reader. */ |
| 381 | printf("/*\n\n Do " "not " "edit " "this " "file " "directly!\n\n*/\n" ); |
| 382 | |
| 383 | puts("/*" ); |
| 384 | puts(ORACLE_WELCOME_COPYRIGHT_NOTICE("2000" )); |
| 385 | puts("*/" ); |
| 386 | |
| 387 | /* Broken up to indicate that it's not advice to you, gentle reader. */ |
| 388 | printf("/* Do " "not " "edit " "this " "file! This is generated by " |
| 389 | "gen_lex_hash.cc\nthat seeks for a perfect hash function */\n\n" ); |
| 390 | printf("#include \"lex.h\"\n\n" ); |
| 391 | |
| 392 | calc_length(); |
| 393 | |
| 394 | if (check_duplicates()) |
| 395 | exit(1); |
| 396 | |
| 397 | generate_find_structs(); |
| 398 | print_find_structs(); |
| 399 | |
| 400 | printf("\nstatic unsigned int sql_functions_max_len=%d;\n" , max_len); |
| 401 | printf("\nstatic unsigned int symbols_max_len=%d;\n\n" , max_len2); |
| 402 | |
| 403 | printf("\ |
| 404 | static SYMBOL *get_hash_symbol(const char *s,\n\ |
| 405 | unsigned int len,bool function)\n\ |
| 406 | {\n\ |
| 407 | uchar *hash_map;\n\ |
| 408 | const char *cur_str= s;\n\ |
| 409 | \n\ |
| 410 | if (len == 0) {\n\ |
| 411 | DBUG_PRINT(\"warning\", (\"get_hash_symbol() received a request for a zero-length symbol, which is probably a mistake.\"));\ |
| 412 | return(NULL);\n\ |
| 413 | }\n" |
| 414 | ); |
| 415 | |
| 416 | printf("\ |
| 417 | if (function){\n\ |
| 418 | if (len>sql_functions_max_len) return 0;\n\ |
| 419 | hash_map= sql_functions_map;\n\ |
| 420 | uint32 cur_struct= uint4korr(hash_map+((len-1)*4));\n\ |
| 421 | \n\ |
| 422 | for (;;){\n\ |
| 423 | uchar first_char= (uchar)cur_struct;\n\ |
| 424 | \n\ |
| 425 | if (first_char == 0)\n\ |
| 426 | {\n\ |
| 427 | int16 ires= (int16)(cur_struct>>16);\n\ |
| 428 | if (ires==array_elements(symbols)) return 0;\n\ |
| 429 | SYMBOL *res;\n\ |
| 430 | if (ires>=0) \n\ |
| 431 | res= symbols+ires;\n\ |
| 432 | else\n\ |
| 433 | res= sql_functions-ires-1;\n\ |
| 434 | uint count= (uint) (cur_str - s);\n\ |
| 435 | return lex_casecmp(cur_str,res->name+count,len-count) ? 0 : res;\n\ |
| 436 | }\n\ |
| 437 | \n\ |
| 438 | uchar cur_char= (uchar)to_upper_lex[(uchar)*cur_str];\n\ |
| 439 | if (cur_char<first_char) return 0;\n\ |
| 440 | cur_struct>>=8;\n\ |
| 441 | if (cur_char>(uchar)cur_struct) return 0;\n\ |
| 442 | \n\ |
| 443 | cur_struct>>=8;\n\ |
| 444 | cur_struct= uint4korr(hash_map+\n\ |
| 445 | (((uint16)cur_struct + cur_char - first_char)*4));\n\ |
| 446 | cur_str++;\n\ |
| 447 | }\n" |
| 448 | ); |
| 449 | |
| 450 | printf("\ |
| 451 | }else{\n\ |
| 452 | if (len>symbols_max_len) return 0;\n\ |
| 453 | hash_map= symbols_map;\n\ |
| 454 | uint32 cur_struct= uint4korr(hash_map+((len-1)*4));\n\ |
| 455 | \n\ |
| 456 | for (;;){\n\ |
| 457 | uchar first_char= (uchar)cur_struct;\n\ |
| 458 | \n\ |
| 459 | if (first_char==0) {\n\ |
| 460 | int16 ires= (int16)(cur_struct>>16);\n\ |
| 461 | if (ires==array_elements(symbols)) return 0;\n\ |
| 462 | SYMBOL *res= symbols+ires;\n\ |
| 463 | uint count= (uint) (cur_str - s);\n\ |
| 464 | return lex_casecmp(cur_str,res->name+count,len-count)!=0 ? 0 : res;\n\ |
| 465 | }\n\ |
| 466 | \n\ |
| 467 | uchar cur_char= (uchar)to_upper_lex[(uchar)*cur_str];\n\ |
| 468 | if (cur_char<first_char) return 0;\n\ |
| 469 | cur_struct>>=8;\n\ |
| 470 | if (cur_char>(uchar)cur_struct) return 0;\n\ |
| 471 | \n\ |
| 472 | cur_struct>>=8;\n\ |
| 473 | cur_struct= uint4korr(hash_map+\n\ |
| 474 | (((uint16)cur_struct + cur_char - first_char)*4));\n\ |
| 475 | cur_str++;\n\ |
| 476 | }\n\ |
| 477 | }\n\ |
| 478 | }\n" |
| 479 | ); |
| 480 | exit(0); |
| 481 | } |
| 482 | |
| 483 | |