Where Online Learning is simpler!
The C and C++ Include Header Files
/usr/include/gawkapi.h
$ cat -n /usr/include/gawkapi.h 1 /* 2 * gawkapi.h -- Definitions for use by extension functions calling into gawk. 3 */ 4 5 /* 6 * Copyright (C) 2012-2019, 2021, 2022, the Free Software Foundation, Inc. 7 * 8 * This file is part of GAWK, the GNU implementation of the 9 * AWK Programming Language. 10 * 11 * GAWK is free software; you can redistribute it and/or modify 12 * it under the terms of the GNU General Public License as published by 13 * the Free Software Foundation; either version 3 of the License, or 14 * (at your option) any later version. 15 * 16 * GAWK is distributed in the hope that it will be useful, 17 * but WITHOUT ANY WARRANTY; without even the implied warranty of 18 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 19 * GNU General Public License for more details. 20 * 21 * You should have received a copy of the GNU General Public License 22 * along with this program; if not, write to the Free Software 23 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA 24 */ 25 26 /* 27 * The following types and/or macros and/or functions are referenced 28 * in this file. For correct use, you must therefore include the 29 * corresponding standard header file BEFORE including this file. 30 * 31 * FILE - <stdio.h> 32 * NULL - <stddef.h> 33 * memset(), memcpy() - <string.h> 34 * size_t - <sys/types.h> 35 * struct stat - <sys/stat.h> 36 * 37 * Due to portability concerns, especially to systems that are not 38 * fully standards-compliant, it is your responsibility to include 39 * the correct files in the correct way. This requirement is necessary 40 * in order to keep this file clean, instead of becoming a portability 41 * hodge-podge as can be seen in the gawk source code. 42 * 43 * To pass reasonable integer values for ERRNO, you will also need to 44 * include <errno.h>. 45 */ 46 47 #ifndef _GAWK_API_H 48 #define _GAWK_API_H 49 50 /* 51 * General introduction: 52 * 53 * This API purposely restricts itself to ISO C 90 features. In particular, no 54 * bool, no // comments, no use of the restrict keyword, or anything else, 55 * in order to provide maximal portability. 56 * 57 * Exception: the "inline" keyword is used below in the "constructor" 58 * functions. If your compiler doesn't support it, you should either 59 * -Dinline='' on your command line, or use the autotools and include a 60 * config.h in your extensions. 61 * 62 * Additional important information: 63 * 64 * 1. ALL string values in awk_value_t objects need to come from api_malloc(). 65 * Gawk will handle releasing the storage if necessary. This is slightly 66 * awkward, in that you can't take an awk_value_t that you got from gawk 67 * and reuse it directly, even for something that is conceptually pass 68 * by value. 69 * 70 * 2. Due to gawk internals, after using sym_update() to install an array 71 * into gawk, you have to retrieve the array cookie from the value 72 * passed in to sym_update(). Like so: 73 * 74 * new_array = create_array(); 75 * val.val_type = AWK_ARRAY; 76 * val.array_cookie = new_array; 77 * sym_update("array", & val); // install array in the symbol table 78 * 79 * new_array = val.array_cookie; // MUST DO THIS 80 * 81 * // fill in new array with lots of subscripts and values 82 * 83 * Similarly, if installing a new array as a subarray of an existing 84 * array, you must add the new array to its parent before adding any 85 * elements to it. 86 * 87 * You must also retrieve the value of the array_cookie after the call 88 * to set_element(). 89 * 90 * Thus, the correct way to build an array is to work "top down". 91 * Create the array, and immediately install it in gawk's symbol table 92 * using sym_update(), or install it as an element in a previously 93 * existing array using set_element(). 94 * 95 * Thus the new array must ultimately be rooted in a global symbol. This is 96 * necessary before installing any subarrays in it, due to gawk's 97 * internal implementation. Strictly speaking, this is required only 98 * for arrays that will have subarrays as elements; however it is 99 * a good idea to always do this. This restriction may be relaxed 100 * in a subsequent revision of the API. 101 */ 102 103 /* Allow use in C++ code. */ 104 #ifdef __cplusplus 105 extern "C" { 106 #endif 107 108 /* This is used to keep extensions from modifying certain fields in some structs. */ 109 #ifdef GAWK 110 #define awk_const 111 #else 112 #define awk_const const 113 #endif 114 115 typedef enum awk_bool { 116 awk_false = 0, 117 awk_true 118 } awk_bool_t; /* we don't use <stdbool.h> on purpose */ 119 120 /* 121 * If an input parser would like to specify the field positions in the input 122 * record, it may populate an awk_fieldwidth_info_t structure to indicate 123 * the location of each field. The use_chars boolean controls whether the 124 * field lengths are specified in terms of bytes or potentially multi-byte 125 * characters. Performance will be better if the values are supplied in 126 * terms of bytes. The fields[0].skip value indicates how many bytes (or 127 * characters) to skip before $1, and fields[0].len is the length of $1, etc. 128 */ 129 130 typedef struct { 131 awk_bool_t use_chars; /* false ==> use bytes */ 132 size_t nf; 133 struct awk_field_info { 134 size_t skip; /* amount to skip before field starts */ 135 size_t len; /* length of field */ 136 } fields[1]; /* actual dimension should be nf */ 137 } awk_fieldwidth_info_t; 138 139 /* 140 * This macro calculates the total struct size needed. This is useful when 141 * calling malloc or realloc. 142 */ 143 #define awk_fieldwidth_info_size(NF) (sizeof(awk_fieldwidth_info_t) + \ 144 (((NF)-1) * sizeof(struct awk_field_info))) 145 146 /* The information about input files that input parsers need to know: */ 147 typedef struct awk_input { 148 const char *name; /* filename */ 149 int fd; /* file descriptor */ 150 #define INVALID_HANDLE (-1) 151 void *opaque; /* private data for input parsers */ 152 153 /* 154 * The get_record function is called to read the next record of data. 155 * 156 * It should return the length of the input record or EOF, and it 157 * should set *out to point to the contents of $0. The rt_start 158 * and rt_len arguments should be used to return RT to gawk. 159 * If EOF is not returned, the parser must set *rt_len (and 160 * *rt_start if *rt_len is non-zero). 161 * 162 * Note that gawk will make a copy of the record in *out, so the 163 * parser is responsible for managing its own memory buffer. 164 * Similarly, gawk will make its own copy of RT, so the parser 165 * is also responsible for managing this memory. 166 * 167 * It is guaranteed that errcode is a valid pointer, so there is 168 * no need to test for a NULL value. Gawk sets *errcode to 0, 169 * so there is no need to set it unless an error occurs. 170 * 171 * If an error does occur, the function should return EOF and set 172 * *errcode to a positive value. In that case, if *errcode is greater 173 * than zero, gawk will automatically update the ERRNO variable based 174 * on the value of *errcode (e.g., setting *errcode = errno should do 175 * the right thing). 176 * 177 * If field_width is non-NULL, then *field_width will be initialized 178 * to NULL, and the function may set it to point to a structure 179 * supplying field width information to override the default 180 * gawk field parsing mechanism. Note that this structure will not 181 * be copied by gawk; it must persist at least until the next call 182 * to get_record or close_func. Note also that field_width will 183 * be NULL when getline is assigning the results to a variable, thus 184 * field parsing is not needed. 185 */ 186 int (*get_record)(char **out, struct awk_input *iobuf, int *errcode, 187 char **rt_start, size_t *rt_len, 188 const awk_fieldwidth_info_t **field_width); 189 190 /* 191 * This replaces the POSIX read() system call. Use it if you want to 192 * manage reading raw bytes yourself, and let gawk parse the record. 193 */ 194 ssize_t (*read_func)(int, void *, size_t); 195 196 /* 197 * The close_func is called to allow the parser to free private data. 198 * Gawk itself will close the fd unless close_func first sets it to 199 * INVALID_HANDLE. 200 */ 201 void (*close_func)(struct awk_input *iobuf); 202 203 /* put last, for alignment. bleah */ 204 struct stat sbuf; /* stat buf */ 205 206 } awk_input_buf_t; 207 208 typedef struct awk_input_parser { 209 const char *name; /* name of parser */ 210 211 /* 212 * The can_take_file function should return true if the parser 213 * would like to parse this file. It should not change any gawk 214 * state! 215 */ 216 awk_bool_t (*can_take_file)(const awk_input_buf_t *iobuf); 217 218 /* 219 * If this parser is selected, then take_control_of will be called. 220 * It can assume that a previous call to can_take_file was successful, 221 * and no gawk state has changed since that call. It should populate 222 * the awk_input_buf_t's get_record, close_func, and opaque values as needed. 223 * It should return true if successful. 224 */ 225 awk_bool_t (*take_control_of)(awk_input_buf_t *iobuf); 226 227 awk_const struct awk_input_parser *awk_const next; /* for use by gawk */ 228 } awk_input_parser_t; 229 230 /* 231 * Similar for output wrapper. 232 */ 233 234 /* First the data structure */ 235 typedef struct awk_output_buf { 236 const char *name; /* name of output file */ 237 const char *mode; /* mode argument to fopen */ 238 FILE *fp; /* stdio file pointer */ 239 awk_bool_t redirected; /* true if a wrapper is active */ 240 void *opaque; /* for use by output wrapper */ 241 242 /* 243 * Replacement functions for I/O. Just like the regular 244 * versions but also take the opaque pointer argument. 245 */ 246 size_t (*gawk_fwrite)(const void *buf, size_t size, size_t count, 247 FILE *fp, void *opaque); 248 int (*gawk_fflush)(FILE *fp, void *opaque); 249 int (*gawk_ferror)(FILE *fp, void *opaque); 250 int (*gawk_fclose)(FILE *fp, void *opaque); 251 } awk_output_buf_t; 252 253 /* Next the output wrapper registered with gawk */ 254 typedef struct awk_output_wrapper { 255 const char *name; /* name of the wrapper */ 256 257 /* 258 * The can_take_file function should return true if the wrapper 259 * would like to process this file. It should not change any gawk 260 * state! 261 */ 262 awk_bool_t (*can_take_file)(const awk_output_buf_t *outbuf); 263 264 /* 265 * If this wrapper is selected, then take_control_of will be called. 266 * It can assume that a previous call to can_take_file was successful, 267 * and no gawk state has changed since that call. It should populate 268 * the awk_output_buf_t function pointers and opaque pointer as needed. 269 * It should return true if successful. 270 */ 271 awk_bool_t (*take_control_of)(awk_output_buf_t *outbuf); 272 273 awk_const struct awk_output_wrapper *awk_const next; /* for use by gawk */ 274 } awk_output_wrapper_t; 275 276 /* A two-way processor combines an input parser and an output wrapper. */ 277 typedef struct awk_two_way_processor { 278 const char *name; /* name of the two-way processor */ 279 280 /* 281 * The can_take_file function should return true if the two-way 282 * processor would like to parse this file. It should not change 283 * any gawk state! 284 */ 285 awk_bool_t (*can_take_two_way)(const char *name); 286 287 /* 288 * If this processor is selected, then take_control_of will be called. 289 * It can assume that a previous call to can_take_file was successful, 290 * and no gawk state has changed since that call. It should populate 291 * the awk_input_buf_t and awk_otuput_buf_t structures as needed. 292 * It should return true if successful. 293 */ 294 awk_bool_t (*take_control_of)(const char *name, awk_input_buf_t *inbuf, 295 awk_output_buf_t *outbuf); 296 297 awk_const struct awk_two_way_processor *awk_const next; /* for use by gawk */ 298 } awk_two_way_processor_t; 299 300 #define gawk_api_major_version 3 301 #define gawk_api_minor_version 2 302 303 /* Current version of the API. */ 304 enum { 305 GAWK_API_MAJOR_VERSION = gawk_api_major_version, 306 GAWK_API_MINOR_VERSION = gawk_api_minor_version 307 }; 308 309 /* A number of typedefs related to different types of values. */ 310 311 /* 312 * A mutable string. Gawk owns the memory pointed to if it supplied 313 * the value. Otherwise, it takes ownership of the memory pointed to. 314 * 315 * The API deals exclusively with regular chars; these strings may 316 * be multibyte encoded in the current locale's encoding and character 317 * set. Gawk will convert internally to wide characters if necessary. 318 * 319 * Note that a string provided by gawk will always be terminated 320 * with a '\0' character. 321 */ 322 typedef struct awk_string { 323 char *str; /* data */ 324 size_t len; /* length thereof, in chars */ 325 } awk_string_t; 326 327 enum AWK_NUMBER_TYPE { 328 AWK_NUMBER_TYPE_DOUBLE, 329 AWK_NUMBER_TYPE_MPFR, 330 AWK_NUMBER_TYPE_MPZ 331 }; 332 333 /* 334 * When type is AWK_NUMBER_MPFR or AWK_NUMBER_MPZ, the memory pointed to 335 * by the ptr member belongs to gawk if it came from gawk. Otherwise the 336 * memory belongs to the extension and gawk copies it when its received. 337 * See the manual for further discussion. 338 */ 339 340 typedef struct awk_number { 341 double d; /* always populated in data received from gawk */ 342 enum AWK_NUMBER_TYPE type; 343 void *ptr; /* either NULL or mpfr_ptr or mpz_ptr */ 344 } awk_number_t; 345 346 /* Arrays are represented as an opaque type. */ 347 typedef void *awk_array_t; 348 349 /* Scalars can be represented as an opaque type. */ 350 typedef void *awk_scalar_t; 351 352 /* Any value can be stored as a cookie. */ 353 typedef void *awk_value_cookie_t; 354 355 /* 356 * This tag defines the type of a value. 357 * 358 * Values are associated with regular variables and with array elements. 359 * Since arrays can be multidimensional (as can regular variables) 360 * it's valid to have a "value" that is actually an array. 361 */ 362 typedef enum { 363 AWK_UNDEFINED, 364 AWK_NUMBER, 365 AWK_STRING, 366 AWK_REGEX, 367 AWK_STRNUM, 368 AWK_ARRAY, 369 AWK_SCALAR, /* opaque access to a variable */ 370 AWK_VALUE_COOKIE, /* for updating a previously created value */ 371 AWK_BOOL 372 } awk_valtype_t; 373 374 /* 375 * An awk value. The val_type tag indicates what 376 * is in the union. 377 */ 378 typedef struct awk_value { 379 awk_valtype_t val_type; 380 union { 381 awk_string_t s; 382 awk_number_t n; 383 awk_array_t a; 384 awk_scalar_t scl; 385 awk_value_cookie_t vc; 386 awk_bool_t b; 387 } u; 388 #define str_value u.s 389 #define strnum_value str_value 390 #define regex_value str_value 391 #define num_value u.n.d 392 #define num_type u.n.type 393 #define num_ptr u.n.ptr 394 #define array_cookie u.a 395 #define scalar_cookie u.scl 396 #define value_cookie u.vc 397 #define bool_value u.b 398 } awk_value_t; 399 400 /* 401 * A "flattened" array element. Gawk produces an array of these 402 * inside the awk_flat_array_t. 403 * ALL memory pointed to belongs to gawk. Individual elements may 404 * be marked for deletion. New elements must be added individually, 405 * one at a time, using the separate API for that purpose. 406 */ 407 408 typedef struct awk_element { 409 /* convenience linked list pointer, not used by gawk */ 410 struct awk_element *next; 411 enum { 412 AWK_ELEMENT_DEFAULT = 0, /* set by gawk */ 413 AWK_ELEMENT_DELETE = 1 /* set by extension if 414 should be deleted */ 415 } flags; 416 awk_value_t index; 417 awk_value_t value; 418 } awk_element_t; 419 420 /* 421 * A "flattened" array. See the description above for how 422 * to use the elements contained herein. 423 */ 424 typedef struct awk_flat_array { 425 awk_const void *awk_const opaque1; /* private data for use by gawk */ 426 awk_const void *awk_const opaque2; /* private data for use by gawk */ 427 awk_const size_t count; /* how many elements */ 428 awk_element_t elements[1]; /* will be extended */ 429 } awk_flat_array_t; 430 431 /* 432 * A record describing an extension function. Upon being 433 * loaded, the extension should pass in one of these to gawk for 434 * each C function. 435 * 436 * Each called function must fill in the result with either a scalar 437 * (number, string, or regex). Gawk takes ownership of any string memory. 438 * 439 * The called function must return the value of `result'. 440 * This is for the convenience of the calling code inside gawk. 441 * 442 * Each extension function may decide what to do if the number of 443 * arguments isn't what it expected. Following awk functions, it 444 * is likely OK to ignore extra arguments. 445 * 446 * 'min_required_args' indicates how many arguments MUST be passed. 447 * The API will throw a fatal error if not enough are passed. 448 * 449 * 'max_expected_args' is more benign; if more than that are passed, 450 * the API prints a lint message (IFF lint is enabled, of course). 451 * 452 * In any case, the extension function itself need not compare the 453 * actual number of arguments passed to those two values if it does 454 * not want to. 455 */ 456 typedef struct awk_ext_func { 457 const char *name; 458 awk_value_t *(*const function)(int num_actual_args, 459 awk_value_t *result, 460 struct awk_ext_func *finfo); 461 const size_t max_expected_args; 462 const size_t min_required_args; 463 awk_bool_t suppress_lint; 464 void *data; /* opaque pointer to any extra state */ 465 } awk_ext_func_t; 466 467 typedef void *awk_ext_id_t; /* opaque type for extension id */ 468 469 /* 470 * The API into gawk. Lots of functions here. We hope that they are 471 * logically organized. 472 * 473 * !!! If you make any changes to this structure, please remember to bump !!! 474 * !!! gawk_api_major_version and/or gawk_api_minor_version. !!! 475 */ 476 typedef struct gawk_api { 477 /* First, data fields. */ 478 479 /* These are what gawk thinks the API version is. */ 480 awk_const int major_version; 481 awk_const int minor_version; 482 483 /* GMP/MPFR versions, if extended-precision is available */ 484 awk_const int gmp_major_version; 485 awk_const int gmp_minor_version; 486 awk_const int mpfr_major_version; 487 awk_const int mpfr_minor_version; 488 489 /* 490 * These can change on the fly as things happen within gawk. 491 * Currently only do_lint is prone to change, but we reserve 492 * the right to allow the others to do so also. 493 */ 494 #define DO_FLAGS_SIZE 6 495 awk_const int do_flags[DO_FLAGS_SIZE]; 496 /* Use these as indices into do_flags[] array to check the values */ 497 #define gawk_do_lint 0 498 #define gawk_do_traditional 1 499 #define gawk_do_profile 2 500 #define gawk_do_sandbox 3 501 #define gawk_do_debug 4 502 #define gawk_do_mpfr 5 503 504 /* Next, registration functions: */ 505 506 /* 507 * Add a function to the interpreter, returns true upon success. 508 * Gawk does not modify what func points to, but the extension 509 * function itself receives this pointer and can modify what it 510 * points to, thus it's not const. 511 */ 512 awk_bool_t (*api_add_ext_func)(awk_ext_id_t id, const char *name_space, 513 awk_ext_func_t *func); 514 515 /* Register an input parser; for opening files read-only */ 516 void (*api_register_input_parser)(awk_ext_id_t id, 517 awk_input_parser_t *input_parser); 518 519 /* Register an output wrapper, for writing files */ 520 void (*api_register_output_wrapper)(awk_ext_id_t id, 521 awk_output_wrapper_t *output_wrapper); 522 523 /* Register a processor for two way I/O */ 524 void (*api_register_two_way_processor)(awk_ext_id_t id, 525 awk_two_way_processor_t *two_way_processor); 526 527 /* 528 * Add an exit call back. 529 * 530 * arg0 is a private data pointer for use by the extension; 531 * gawk saves it and passes it into the function pointed 532 * to by funcp at exit. 533 * 534 * Exit callback functions are called in LIFO order. 535 */ 536 void (*api_awk_atexit)(awk_ext_id_t id, 537 void (*funcp)(void *data, int exit_status), 538 void *arg0); 539 540 /* Register a version string for this extension with gawk. */ 541 void (*api_register_ext_version)(awk_ext_id_t id, const char *version); 542 543 /* Functions to print messages */ 544 void (*api_fatal)(awk_ext_id_t id, const char *format, ...); 545 void (*api_warning)(awk_ext_id_t id, const char *format, ...); 546 void (*api_lintwarn)(awk_ext_id_t id, const char *format, ...); 547 void (*api_nonfatal)(awk_ext_id_t id, const char *format, ...); 548 549 /* Functions to update ERRNO */ 550 void (*api_update_ERRNO_int)(awk_ext_id_t id, int errno_val); 551 void (*api_update_ERRNO_string)(awk_ext_id_t id, const char *string); 552 void (*api_unset_ERRNO)(awk_ext_id_t id); 553 554 /* 555 * All of the functions that return a value from inside gawk 556 * (get a parameter, get a global variable, get an array element) 557 * behave in the same way. 558 * 559 * For a function parameter, the return is false if the argument 560 * count is out of range, or if the actual parameter does not match 561 * what is specified in wanted. In that case, result->val_type 562 * will hold the actual type of what was passed. 563 * 564 * Similarly for symbol table access to variables and array elements, 565 * the return is false if the actual variable or array element does 566 * not match what was requested, and result->val_type will hold 567 * the actual type. 568 569 Table entry is type returned: 570 571 572 +----------------------------------------------------------------+ 573 | Type of Actual Value: | 574 +--------+--------+--------+--------+--------+-------+-----------+ 575 | String | Strnum | Number | Regex | Bool | Array | Undefined | 576 +-----------+-----------+--------+--------+--------+--------+--------+-------+-----------+ 577 | | String | String | String | String | String | String | false | false | 578 | +-----------+--------+--------+--------+--------+--------+-------+-----------+ 579 | | Strnum | false | Strnum | Strnum | false | false | false | false | 580 | +-----------+--------+--------+--------+--------+--------+-------+-----------+ 581 | | Number | Number | Number | Number | false | Number | false | false | 582 | +-----------+--------+--------+--------+--------+--------+-------+-----------+ 583 | | Regex | false | false | false | Regex | false | false | false | 584 | +-----------+--------+--------+--------+--------+--------+-------+-----------+ 585 | Type | Bool | false | false | false | false | Bool | false | false | 586 | Requested +-----------+--------+--------+--------+--------+--------+-------+-----------+ 587 | | Array | false | false | false | false | false | Array | false | 588 | +-----------+--------+--------+--------+--------+--------+-------+-----------+ 589 | | Scalar | Scalar | Scalar | Scalar | Scalar | Scalar | false | false | 590 | +-----------+--------+--------+--------+--------+--------+-------+-----------+ 591 | | Undefined | String | Strnum | Number | Regex | Bool | Array | Undefined | 592 | +-----------+--------+--------+--------+--------+--------+-------+-----------+ 593 | | Value | false | false | false | false | false | false | false | 594 | | Cookie | | | | | | | | 595 +-----------+-----------+--------+--------+--------+--------+--------+-------+-----------+ 596 */ 597 598 /* Functions to handle parameters passed to the extension. */ 599 600 /* 601 * Get the count'th parameter, zero-based. 602 * Returns false if count is out of range, or if actual parameter 603 * does not match what is specified in wanted. In that case, 604 * result->val_type is as described above. 605 */ 606 awk_bool_t (*api_get_argument)(awk_ext_id_t id, size_t count, 607 awk_valtype_t wanted, 608 awk_value_t *result); 609 610 /* 611 * Convert a parameter that was undefined into an array 612 * (provide call-by-reference for arrays). Returns false 613 * if count is too big, or if the argument's type is 614 * not undefined. 615 */ 616 awk_bool_t (*api_set_argument)(awk_ext_id_t id, 617 size_t count, 618 awk_array_t array); 619 620 /* 621 * Symbol table access: 622 * - Read-only access to special variables (NF, etc.) 623 * - One special exception: PROCINFO. 624 * - Use sym_update() to change a value, including from UNDEFINED 625 * to scalar or array. 626 */ 627 /* 628 * Lookup a variable, fill in value. No messing with the value 629 * returned. 630 * Returns false if the variable doesn't exist or if the wrong type 631 * was requested. In the latter case, vaule->val_type will have 632 * the real type, as described above. 633 * 634 * awk_value_t val; 635 * if (! api->sym_lookup(id, name, wanted, & val)) 636 * error_code_here(); 637 * else { 638 * // safe to use val 639 * } 640 */ 641 awk_bool_t (*api_sym_lookup)(awk_ext_id_t id, 642 const char *name_space, 643 const char *name, 644 awk_valtype_t wanted, 645 awk_value_t *result); 646 647 /* 648 * Update a value. Adds it to the symbol table if not there. 649 * Changing types (scalar <--> array) is not allowed. 650 * In fact, using this to update an array is not allowed, either. 651 * Such an attempt returns false. 652 */ 653 awk_bool_t (*api_sym_update)(awk_ext_id_t id, 654 const char *name_space, 655 const char *name, 656 awk_value_t *value); 657 658 /* 659 * A ``scalar cookie'' is an opaque handle that provide access 660 * to a global variable or array. It is an optimization that 661 * avoids looking up variables in gawk's symbol table every time 662 * access is needed. 663 * 664 * This function retrieves the current value of a scalar cookie. 665 * Once you have obtained a scalar_cookie using sym_lookup, you can 666 * use this function to get its value more efficiently. 667 * 668 * Return will be false if the value cannot be retrieved. 669 * 670 * Flow is thus 671 * awk_value_t val; 672 * awk_scalar_t cookie; 673 * api->sym_lookup(id, "variable", AWK_SCALAR, & val); // get the cookie 674 * cookie = val.scalar_cookie; 675 * ... 676 * api->sym_lookup_scalar(id, cookie, wanted, & val); // get the value 677 */ 678 awk_bool_t (*api_sym_lookup_scalar)(awk_ext_id_t id, 679 awk_scalar_t cookie, 680 awk_valtype_t wanted, 681 awk_value_t *result); 682 683 /* 684 * Update the value associated with a scalar cookie. 685 * Flow is 686 * sym_lookup with wanted == AWK_SCALAR 687 * if returns false 688 * sym_update with real initial value to install it 689 * sym_lookup again with AWK_SCALAR 690 * else 691 * use the scalar cookie 692 * 693 * Return will be false if the new value is not one of 694 * AWK_STRING, AWK_NUMBER, AWK_REGEX. 695 * 696 * Here too, the built-in variables may not be updated. 697 */ 698 awk_bool_t (*api_sym_update_scalar)(awk_ext_id_t id, 699 awk_scalar_t cookie, awk_value_t *value); 700 701 /* Cached values */ 702 703 /* 704 * Create a cached string,regex, or numeric value for efficient later 705 * assignment. This improves performance when you want to assign 706 * the same value to one or more variables repeatedly. Only 707 * AWK_NUMBER, AWK_STRING, AWK_REGEX and AWK_STRNUM values are allowed. 708 * Any other type is rejected. We disallow AWK_UNDEFINED since that 709 * case would result in inferior performance. 710 */ 711 awk_bool_t (*api_create_value)(awk_ext_id_t id, awk_value_t *value, 712 awk_value_cookie_t *result); 713 714 /* 715 * Release the memory associated with a cookie from api_create_value. 716 * Please call this to free memory when the value is no longer needed. 717 */ 718 awk_bool_t (*api_release_value)(awk_ext_id_t id, awk_value_cookie_t vc); 719 720 /* Array management */ 721 722 /* 723 * Retrieve total number of elements in array. 724 * Returns false if some kind of error. 725 */ 726 awk_bool_t (*api_get_element_count)(awk_ext_id_t id, 727 awk_array_t a_cookie, size_t *count); 728 729 /* 730 * Return the value of an element - read only! 731 * Use set_array_element() to change it. 732 * Behavior for value and return is same as for api_get_argument 733 * and sym_lookup. 734 */ 735 awk_bool_t (*api_get_array_element)(awk_ext_id_t id, 736 awk_array_t a_cookie, 737 const awk_value_t *const index, 738 awk_valtype_t wanted, 739 awk_value_t *result); 740 741 /* 742 * Change (or create) element in existing array with 743 * index and value. 744 * 745 * ARGV and ENVIRON may not be updated. 746 */ 747 awk_bool_t (*api_set_array_element)(awk_ext_id_t id, awk_array_t a_cookie, 748 const awk_value_t *const index, 749 const awk_value_t *const value); 750 751 /* 752 * Remove the element with the given index. 753 * Returns true if removed or false if element did not exist. 754 */ 755 awk_bool_t (*api_del_array_element)(awk_ext_id_t id, 756 awk_array_t a_cookie, const awk_value_t* const index); 757 758 /* Create a new array cookie to which elements may be added. */ 759 awk_array_t (*api_create_array)(awk_ext_id_t id); 760 761 /* Clear out an array. */ 762 awk_bool_t (*api_clear_array)(awk_ext_id_t id, awk_array_t a_cookie); 763 764 /* 765 * Flatten out an array with type conversions as requested. 766 * This supersedes the earlier api_flatten_array function that 767 * did not allow the caller to specify the requested types. 768 * (That API is still available as a macro, defined below.) 769 */ 770 awk_bool_t (*api_flatten_array_typed)(awk_ext_id_t id, 771 awk_array_t a_cookie, 772 awk_flat_array_t **data, 773 awk_valtype_t index_type, awk_valtype_t value_type); 774 775 /* When done, delete any marked elements, release the memory. */ 776 awk_bool_t (*api_release_flattened_array)(awk_ext_id_t id, 777 awk_array_t a_cookie, 778 awk_flat_array_t *data); 779 780 /* 781 * Hooks to provide access to gawk's memory allocation functions. 782 * This ensures that memory passed between gawk and the extension 783 * is allocated and released by the same library. 784 */ 785 void *(*api_malloc)(size_t size); 786 void *(*api_calloc)(size_t nmemb, size_t size); 787 void *(*api_realloc)(void *ptr, size_t size); 788 void (*api_free)(void *ptr); 789 790 /* 791 * Obsolete function, should not be used. It remains only 792 * for binary compatibility. Any value it returns should be 793 * freed via api_free. 794 */ 795 void *(*api_get_mpfr)(awk_ext_id_t id); 796 797 /* 798 * Obsolete function, should not be used. It remains only 799 * for binary compatibility. Any value it returns should be 800 * freed via api_free. 801 */ 802 void *(*api_get_mpz)(awk_ext_id_t id); 803 804 /* 805 * Look up a file. If the name is NULL or name_len is 0, it returns 806 * data for the currently open input file corresponding to FILENAME 807 * (and it will not access the filetype argument, so that may be 808 * undefined). 809 * 810 * If the file is not already open, try to open it. 811 * 812 * The "filetype" argument should be one of: 813 * 814 * ">", ">>", "<", "|>", "|<", and "|&" 815 * 816 * If the file is not already open, and the fd argument is non-negative, 817 * gawk will use that file descriptor instead of opening the file 818 * in the usual way. 819 * 820 * If the fd is non-negative, but the file exists already, gawk 821 * ignores the fd and returns the existing file. It is the caller's 822 * responsibility to notice that the fd in the returned 823 * awk_input_buf_t does not match the requested value. 824 * 825 * Note that supplying a file descriptor is currently NOT supported 826 * for pipes. It should work for input, output, append, and two-way 827 * (coprocess) sockets. If the filetype is two-way, we assume that 828 * it is a socket! 829 * 830 * Note that in the two-way case, the input and output file descriptors 831 * may differ. To check for success, one must check that either of 832 * them matches. 833 * 834 * ibufp and obufp point at gawk's internal copies of the 835 * awk_input_buf_t and awk_output_t associated with the open 836 * file. Treat these data structures as read-only! 837 */ 838 awk_bool_t (*api_get_file)(awk_ext_id_t id, 839 const char *name, 840 size_t name_len, 841 const char *filetype, 842 int fd, 843 /* 844 * Return values (on success, one or both should 845 * be non-NULL): 846 */ 847 const awk_input_buf_t **ibufp, 848 const awk_output_buf_t **obufp); 849 850 /* Destroy an array. */ 851 awk_bool_t (*api_destroy_array)(awk_ext_id_t id, awk_array_t a_cookie); 852 } gawk_api_t; 853 854 #ifndef GAWK /* these are not for the gawk code itself! */ 855 /* 856 * Use these if you want to define "global" variables named api 857 * and ext_id to make the code a little easier to read. 858 * See the sample boilerplate code, below. 859 */ 860 #define do_lint (api->do_flags[gawk_do_lint]) 861 #define do_traditional (api->do_flags[gawk_do_traditional]) 862 #define do_profile (api->do_flags[gawk_do_profile]) 863 #define do_sandbox (api->do_flags[gawk_do_sandbox]) 864 #define do_debug (api->do_flags[gawk_do_debug]) 865 #define do_mpfr (api->do_flags[gawk_do_mpfr]) 866 867 #define get_argument(count, wanted, result) \ 868 (api->api_get_argument(ext_id, count, wanted, result)) 869 #define set_argument(count, new_array) \ 870 (api->api_set_argument(ext_id, count, new_array)) 871 872 #define fatal api->api_fatal 873 #define nonfatal api->api_nonfatal 874 #define warning api->api_warning 875 #define lintwarn api->api_lintwarn 876 877 #define register_input_parser(parser) (api->api_register_input_parser(ext_id, parser)) 878 #define register_output_wrapper(wrapper) (api->api_register_output_wrapper(ext_id, wrapper)) 879 #define register_two_way_processor(processor) \ 880 (api->api_register_two_way_processor(ext_id, processor)) 881 882 #define update_ERRNO_int(e) (api->api_update_ERRNO_int(ext_id, e)) 883 #define update_ERRNO_string(str) \ 884 (api->api_update_ERRNO_string(ext_id, str)) 885 #define unset_ERRNO() (api->api_unset_ERRNO(ext_id)) 886 887 #define add_ext_func(ns, func) (api->api_add_ext_func(ext_id, ns, func)) 888 #define awk_atexit(funcp, arg0) (api->api_awk_atexit(ext_id, funcp, arg0)) 889 890 #define sym_lookup(name, wanted, result) \ 891 sym_lookup_ns("", name, wanted, result) 892 #define sym_update(name, value) \ 893 sym_update_ns("", name, value) 894 895 #define sym_lookup_ns(name_space, name, wanted, result) \ 896 (api->api_sym_lookup(ext_id, name_space, name, wanted, result)) 897 #define sym_update_ns(name_space, name, value) \ 898 (api->api_sym_update(ext_id, name_space, name, value)) 899 900 #define sym_lookup_scalar(scalar_cookie, wanted, result) \ 901 (api->api_sym_lookup_scalar(ext_id, scalar_cookie, wanted, result)) 902 #define sym_update_scalar(scalar_cookie, value) \ 903 (api->api_sym_update_scalar)(ext_id, scalar_cookie, value) 904 905 #define get_array_element(array, index, wanted, result) \ 906 (api->api_get_array_element(ext_id, array, index, wanted, result)) 907 908 #define set_array_element(array, index, value) \ 909 (api->api_set_array_element(ext_id, array, index, value)) 910 911 #define set_array_element_by_elem(array, elem) \ 912 (api->api_set_array_element(ext_id, array, & (elem)->index, & (elem)->value)) 913 914 #define del_array_element(array, index) \ 915 (api->api_del_array_element(ext_id, array, index)) 916 917 #define get_element_count(array, count_p) \ 918 (api->api_get_element_count(ext_id, array, count_p)) 919 920 #define create_array() (api->api_create_array(ext_id)) 921 922 #define destroy_array(array) (api->api_destroy_array(ext_id, array)) 923 924 #define clear_array(array) (api->api_clear_array(ext_id, array)) 925 926 #define flatten_array_typed(array, data, index_type, value_type) \ 927 (api->api_flatten_array_typed(ext_id, array, data, index_type, value_type)) 928 929 #define flatten_array(array, data) \ 930 flatten_array_typed(array, data, AWK_STRING, AWK_UNDEFINED) 931 932 #define release_flattened_array(array, data) \ 933 (api->api_release_flattened_array(ext_id, array, data)) 934 935 #define gawk_malloc(size) (api->api_malloc(size)) 936 #define gawk_calloc(nmemb, size) (api->api_calloc(nmemb, size)) 937 #define gawk_realloc(ptr, size) (api->api_realloc(ptr, size)) 938 #define gawk_free(ptr) (api->api_free(ptr)) 939 940 #define create_value(value, result) \ 941 (api->api_create_value(ext_id, value,result)) 942 943 #define release_value(value) \ 944 (api->api_release_value(ext_id, value)) 945 946 #define get_file(name, namelen, filetype, fd, ibuf, obuf) \ 947 (api->api_get_file(ext_id, name, namelen, filetype, fd, ibuf, obuf)) 948 949 /* These two are obsolete and should not be used. */ 950 #define get_mpfr_ptr() (api->api_get_mpfr(ext_id)) 951 #define get_mpz_ptr() (api->api_get_mpz(ext_id)) 952 953 #define register_ext_version(version) \ 954 (api->api_register_ext_version(ext_id, version)) 955 956 #define emalloc(pointer, type, size, message) \ 957 do { \ 958 if ((pointer = (type) gawk_malloc(size)) == 0) \ 959 fatal(ext_id, "%s: malloc of %d bytes failed", message, size); \ 960 } while(0) 961 962 #define ezalloc(pointer, type, size, message) \ 963 do { \ 964 if ((pointer = (type) gawk_calloc(1, size)) == 0) \ 965 fatal(ext_id, "%s: calloc of %d bytes failed", message, size); \ 966 } while(0) 967 968 #define erealloc(pointer, type, size, message) \ 969 do { \ 970 if ((pointer = (type) gawk_realloc(pointer, size)) == 0) \ 971 fatal(ext_id, "%s: realloc of %d bytes failed", message, size); \ 972 } while(0) 973 974 /* Constructor functions */ 975 976 /* r_make_string_type --- make a string or strnum or regexp value in result from the passed-in string */ 977 978 static inline awk_value_t * 979 r_make_string_type(const gawk_api_t *api, /* needed for emalloc */ 980 awk_ext_id_t ext_id, /* ditto */ 981 const char *string, 982 size_t length, 983 awk_bool_t duplicate, 984 awk_value_t *result, 985 awk_valtype_t val_type) 986 { 987 char *cp = NULL; 988 989 memset(result, 0, sizeof(*result)); 990 991 result->val_type = val_type; 992 result->str_value.len = length; 993 994 if (duplicate) { 995 emalloc(cp, char *, length + 1, "r_make_string"); 996 memcpy(cp, string, length); 997 cp[length] = '\0'; 998 result->str_value.str = cp; 999 } else { 1000 result->str_value.str = (char *) string; 1001 } 1002 1003 return result; 1004 } 1005 1006 /* r_make_string --- make a string value in result from the passed-in string */ 1007 1008 static inline awk_value_t * 1009 r_make_string(const gawk_api_t *api, /* needed for emalloc */ 1010 awk_ext_id_t ext_id, /* ditto */ 1011 const char *string, 1012 size_t length, 1013 awk_bool_t duplicate, 1014 awk_value_t *result) 1015 { 1016 return r_make_string_type(api, ext_id, string, length, duplicate, result, AWK_STRING); 1017 } 1018 1019 #define make_const_string(str, len, result) r_make_string(api, ext_id, str, len, awk_true, result) 1020 #define make_malloced_string(str, len, result) r_make_string(api, ext_id, str, len, awk_false, result) 1021 1022 #define make_const_regex(str, len, result) r_make_string_type(api, ext_id, str, len, awk_true, result, AWK_REGEX) 1023 #define make_malloced_regex(str, len, result) r_make_string_type(api, ext_id, str, len, awk_false, result, AWK_REGEX) 1024 1025 /* 1026 * Note: The caller may not create a STRNUM, but it can create a string that is 1027 * flagged as user input that MAY be a STRNUM. Gawk will decide whether it's a 1028 * STRNUM or a string by checking whether the string is numeric. 1029 */ 1030 #define make_const_user_input(str, len, result) r_make_string_type(api, ext_id, str, len, 1, result, AWK_STRNUM) 1031 #define make_malloced_user_input(str, len, result) r_make_string_type(api, ext_id, str, len, 0, result, AWK_STRNUM) 1032 1033 /* make_null_string --- make a null string value */ 1034 1035 static inline awk_value_t * 1036 make_null_string(awk_value_t *result) 1037 { 1038 memset(result, 0, sizeof(*result)); 1039 result->val_type = AWK_UNDEFINED; 1040 1041 return result; 1042 } 1043 1044 /* make_number --- make a number value in result */ 1045 1046 static inline awk_value_t * 1047 make_number(double num, awk_value_t *result) 1048 { 1049 result->val_type = AWK_NUMBER; 1050 result->num_value = num; 1051 result->num_type = AWK_NUMBER_TYPE_DOUBLE; 1052 return result; 1053 } 1054 1055 /* 1056 * make_number_mpz --- make an mpz number value in result. 1057 * The mpz_ptr must be from a call to get_mpz_ptr. 1058 */ 1059 1060 static inline awk_value_t * 1061 make_number_mpz(void *mpz_ptr, awk_value_t *result) 1062 { 1063 result->val_type = AWK_NUMBER; 1064 result->num_type = AWK_NUMBER_TYPE_MPZ; 1065 result->num_ptr = mpz_ptr; 1066 return result; 1067 } 1068 1069 /* 1070 * make_number_mpfr --- make an mpfr number value in result. 1071 * The mpfr_ptr must be from a call to get_mpfr_ptr. 1072 */ 1073 1074 static inline awk_value_t * 1075 make_number_mpfr(void *mpfr_ptr, awk_value_t *result) 1076 { 1077 result->val_type = AWK_NUMBER; 1078 result->num_type = AWK_NUMBER_TYPE_MPFR; 1079 result->num_ptr = mpfr_ptr; 1080 return result; 1081 } 1082 1083 /* make_bool --- make a bool value in result */ 1084 1085 static inline awk_value_t * 1086 make_bool(awk_bool_t boolval, awk_value_t *result) 1087 { 1088 result->val_type = AWK_BOOL; 1089 result->bool_value = boolval; 1090 return result; 1091 } 1092 1093 1094 /* 1095 * Each extension must define a function with this prototype: 1096 * 1097 * int dl_load(gawk_api_t *api_p, awk_ext_id_t id) 1098 * 1099 * The return value should be zero on failure and non-zero on success. 1100 * 1101 * For the macros to work, the function should save api_p in a global 1102 * variable named 'api' and save id in a global variable named 'ext_id'. 1103 * In addition, a global function pointer named 'init_func' should be 1104 * defined and set to either NULL or an initialization function that 1105 * returns non-zero on success and zero upon failure. 1106 */ 1107 1108 extern int dl_load(const gawk_api_t *const api_p, awk_ext_id_t id); 1109 1110 #if 0 1111 /* Boilerplate code: */ 1112 int plugin_is_GPL_compatible; 1113 1114 static gawk_api_t *const api; 1115 static awk_ext_id_t ext_id; 1116 static const char *ext_version = NULL; /* or ... = "some string" */ 1117 1118 static awk_ext_func_t func_table[] = { 1119 { "name", do_name, 1 }, 1120 /* ... */ 1121 }; 1122 1123 /* EITHER: */ 1124 1125 static awk_bool_t (*init_func)(void) = NULL; 1126 1127 /* OR: */ 1128 1129 static awk_bool_t 1130 init_my_extension(void) 1131 { 1132 ... 1133 } 1134 1135 static awk_bool_t (*init_func)(void) = init_my_extension; 1136 1137 dl_load_func(func_table, some_name, "name_space_in_quotes") 1138 #endif 1139 1140 #define dl_load_func(func_table, extension, name_space) \ 1141 int dl_load(const gawk_api_t *const api_p, awk_ext_id_t id) \ 1142 { \ 1143 size_t i, j; \ 1144 int errors = 0; \ 1145 \ 1146 api = api_p; \ 1147 ext_id = (void **) id; \ 1148 \ 1149 if (api->major_version != GAWK_API_MAJOR_VERSION \ 1150 || api->minor_version < GAWK_API_MINOR_VERSION) { \ 1151 fprintf(stderr, #extension ": version mismatch with gawk!\n"); \ 1152 fprintf(stderr, "\tmy version (API %d.%d), gawk version (API %d.%d)\n", \ 1153 GAWK_API_MAJOR_VERSION, GAWK_API_MINOR_VERSION, \ 1154 api->major_version, api->minor_version); \ 1155 exit(1); \ 1156 } \ 1157 \ 1158 check_mpfr_version(extension); \ 1159 \ 1160 /* load functions */ \ 1161 for (i = 0, j = sizeof(func_table) / sizeof(func_table[0]); i < j; i++) { \ 1162 if (func_table[i].name == NULL) \ 1163 break; \ 1164 if (! add_ext_func(name_space, & func_table[i])) { \ 1165 warning(ext_id, #extension ": could not add %s", \ 1166 func_table[i].name); \ 1167 errors++; \ 1168 } \ 1169 } \ 1170 \ 1171 if (init_func != NULL) { \ 1172 if (! init_func()) { \ 1173 warning(ext_id, #extension ": initialization function failed"); \ 1174 errors++; \ 1175 } \ 1176 } \ 1177 \ 1178 if (ext_version != NULL) \ 1179 register_ext_version(ext_version); \ 1180 \ 1181 return (errors == 0); \ 1182 } 1183 1184 #if defined __GNU_MP_VERSION && defined MPFR_VERSION_MAJOR 1185 #define check_mpfr_version(extension) do { \ 1186 if (api->gmp_major_version != __GNU_MP_VERSION \ 1187 || api->gmp_minor_version < __GNU_MP_VERSION_MINOR) { \ 1188 fprintf(stderr, #extension ": GMP version mismatch with gawk!\n"); \ 1189 fprintf(stderr, "\tmy version (%d, %d), gawk version (%d, %d)\n", \ 1190 __GNU_MP_VERSION, __GNU_MP_VERSION_MINOR, \ 1191 api->gmp_major_version, api->gmp_minor_version); \ 1192 exit(1); \ 1193 } \ 1194 if (api->mpfr_major_version != MPFR_VERSION_MAJOR \ 1195 || api->mpfr_minor_version < MPFR_VERSION_MINOR) { \ 1196 fprintf(stderr, #extension ": MPFR version mismatch with gawk!\n"); \ 1197 fprintf(stderr, "\tmy version (%d, %d), gawk version (%d, %d)\n", \ 1198 MPFR_VERSION_MAJOR, MPFR_VERSION_MINOR, \ 1199 api->mpfr_major_version, api->mpfr_minor_version); \ 1200 exit(1); \ 1201 } \ 1202 } while (0) 1203 #else 1204 #define check_mpfr_version(extension) /* nothing */ 1205 #endif 1206 1207 #endif /* GAWK */ 1208 1209 #ifdef __cplusplus 1210 } 1211 #endif /* C++ */ 1212 1213 #endif /* _GAWK_API_H */
Contact us
|
About us
|
Term of use
|
Copyright © 2000-2025 MyWebUniversity.com ™
