| 1 | /* GNU SED, a batch stream editor. |
| 2 | Copyright (C) 1989-1991 Free Software Foundation, Inc. |
| 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; either version 2, or (at your option) |
| 7 | any later version. |
| 8 | |
| 9 | This program is distributed in the hope that it will be useful, |
| 10 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 11 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 12 | GNU General Public License for more details. |
| 13 | |
| 14 | You should have received a copy of the GNU General Public License |
| 15 | along with this program; if not, write to the Free Software |
| 16 | Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. */ |
| 17 | |
| 18 | #ifdef __STDC__ |
| 19 | #define VOID void |
| 20 | #else |
| 21 | #define VOID char |
| 22 | #endif |
| 23 | |
| 24 | #define _GNU_SOURCE |
| 25 | #include <ctype.h> |
| 26 | #ifndef isblank |
| 27 | #define isblank(c) ((c) == ' ' || (c) == '\t') |
| 28 | #endif |
| 29 | #include <stdio.h> |
| 30 | #include <regex.h> |
| 31 | #include <getopt.h> |
| 32 | #if defined(STDC_HEADERS) |
| 33 | #include <stdlib.h> |
| 34 | #endif |
| 35 | #if defined(USG) || defined(STDC_HEADERS) |
| 36 | #include <string.h> |
| 37 | #include <memory.h> |
| 38 | #define bcopy(s, d, n) (memcpy((d), (s), (n))) |
| 39 | #else |
| 40 | #include <strings.h> |
| 41 | VOID *memchr(); |
| 42 | #endif |
| 43 | |
| 44 | char *version_string = "GNU sed version 1.08"; |
| 45 | |
| 46 | /* Struct vector is used to describe a chunk of a sed program. There is one |
| 47 | vector for the main program, and one for each { } pair. */ |
| 48 | struct vector { |
| 49 | struct sed_cmd *v; |
| 50 | int v_length; |
| 51 | int v_allocated; |
| 52 | struct vector *up_one; |
| 53 | struct vector *next_one; |
| 54 | }; |
| 55 | |
| 56 | |
| 57 | /* Goto structure is used to hold both GOTO's and labels. There are two |
| 58 | separate lists, one of goto's, called 'jumps', and one of labels, called |
| 59 | 'labels'. |
| 60 | the V element points to the descriptor for the program-chunk in which the |
| 61 | goto was encountered. |
| 62 | the v_index element counts which element of the vector actually IS the |
| 63 | goto/label. The first element of the vector is zero. |
| 64 | the NAME element is the null-terminated name of the label. |
| 65 | next is the next goto/label in the list. */ |
| 66 | |
| 67 | struct sed_label { |
| 68 | struct vector *v; |
| 69 | int v_index; |
| 70 | char *name; |
| 71 | struct sed_label *next; |
| 72 | }; |
| 73 | |
| 74 | /* ADDR_TYPE is zero for a null address, |
| 75 | one if addr_number is valid, or |
| 76 | two if addr_regex is valid, |
| 77 | three, if the address is '$' |
| 78 | |
| 79 | Other values are undefined. |
| 80 | */ |
| 81 | |
| 82 | #define ADDR_NULL 0 |
| 83 | #define ADDR_NUM 1 |
| 84 | #define ADDR_REGEX 2 |
| 85 | #define ADDR_LAST 3 |
| 86 | |
| 87 | struct addr { |
| 88 | int addr_type; |
| 89 | struct re_pattern_buffer *addr_regex; |
| 90 | int addr_number; |
| 91 | }; |
| 92 | |
| 93 | |
| 94 | /* Aflags: If the low order bit is set, a1 has been |
| 95 | matched; apply this command until a2 matches. |
| 96 | If the next bit is set, apply this command to all |
| 97 | lines that DON'T match the address(es). |
| 98 | */ |
| 99 | |
| 100 | #define A1_MATCHED_BIT 01 |
| 101 | #define ADDR_BANG_BIT 02 |
| 102 | |
| 103 | |
| 104 | struct sed_cmd { |
| 105 | struct addr a1,a2; |
| 106 | int aflags; |
| 107 | |
| 108 | char cmd; |
| 109 | |
| 110 | union { |
| 111 | /* This structure is used for a, i, and c commands */ |
| 112 | struct { |
| 113 | char *text; |
| 114 | int text_len; |
| 115 | } cmd_txt; |
| 116 | |
| 117 | /* This is used for b and t commands */ |
| 118 | struct sed_cmd *label; |
| 119 | |
| 120 | /* This for r and w commands */ |
| 121 | FILE *io_file; |
| 122 | |
| 123 | /* This for the hairy s command */ |
| 124 | /* For the flags var: |
| 125 | low order bit means the 'g' option was given, |
| 126 | next bit means the 'p' option was given, |
| 127 | and the next bit means a 'w' option was given, |
| 128 | and wio_file contains the file to write to. */ |
| 129 | |
| 130 | #define S_GLOBAL_BIT 01 |
| 131 | #define S_PRINT_BIT 02 |
| 132 | #define S_WRITE_BIT 04 |
| 133 | #define S_NUM_BIT 010 |
| 134 | |
| 135 | struct { |
| 136 | struct re_pattern_buffer *regx; |
| 137 | char *replacement; |
| 138 | int replace_length; |
| 139 | int flags; |
| 140 | int numb; |
| 141 | FILE *wio_file; |
| 142 | } cmd_regex; |
| 143 | |
| 144 | /* This for the y command */ |
| 145 | unsigned char *translate; |
| 146 | |
| 147 | /* For { and } */ |
| 148 | struct vector *sub; |
| 149 | struct sed_label *jump; |
| 150 | } x; |
| 151 | }; |
| 152 | |
| 153 | /* Sed operates a line at a time. */ |
| 154 | struct line { |
| 155 | char *text; /* Pointer to line allocated by malloc. */ |
| 156 | int length; /* Length of text. */ |
| 157 | int alloc; /* Allocated space for text. */ |
| 158 | }; |
| 159 | |
| 160 | /* This structure holds information about files opend by the 'r', 'w', |
| 161 | and 's///w' commands. In paticular, it holds the FILE pointer to |
| 162 | use, the file's name, a flag that is non-zero if the file is being |
| 163 | read instead of written. */ |
| 164 | |
| 165 | #define NUM_FPS 32 |
| 166 | struct { |
| 167 | FILE *phile; |
| 168 | char *name; |
| 169 | int readit; |
| 170 | } file_ptrs[NUM_FPS]; |
| 171 | |
| 172 | |
| 173 | #if defined(__STDC__) |
| 174 | # define P_(s) s |
| 175 | #else |
| 176 | # define P_(s) () |
| 177 | #endif |
| 178 | |
| 179 | void panic P_((char *str, ...)); |
| 180 | char *__fp_name P_((FILE *fp)); |
| 181 | FILE *ck_fopen P_((char *name, char *mode)); |
| 182 | void ck_fwrite P_((char *ptr, int size, int nmemb, FILE *stream)); |
| 183 | void ck_fclose P_((FILE *stream)); |
| 184 | VOID *ck_malloc P_((int size)); |
| 185 | VOID *ck_realloc P_((VOID *ptr, int size)); |
| 186 | char *ck_strdup P_((char *str)); |
| 187 | VOID *init_buffer P_((void)); |
| 188 | void flush_buffer P_((VOID *bb)); |
| 189 | int size_buffer P_((VOID *b)); |
| 190 | void add_buffer P_((VOID *bb, char *p, int n)); |
| 191 | void add1_buffer P_((VOID *bb, int ch)); |
| 192 | char *get_buffer P_((VOID *bb)); |
| 193 | |
| 194 | void compile_string P_((char *str)); |
| 195 | void compile_file P_((char *str)); |
| 196 | struct vector *compile_program P_((struct vector *vector)); |
| 197 | void bad_prog P_((char *why)); |
| 198 | int inchar P_((void)); |
| 199 | void savchar P_((int ch)); |
| 200 | int compile_address P_((struct addr *addr)); |
| 201 | void compile_regex P_((int slash)); |
| 202 | struct sed_label *setup_jump P_((struct sed_label *list, struct sed_cmd *cmd, struct vector *vec)); |
| 203 | FILE *compile_filename P_((int readit)); |
| 204 | void read_file P_((char *name)); |
| 205 | void execute_program P_((struct vector *vec)); |
| 206 | int match_address P_((struct addr *addr)); |
| 207 | int read_pattern_space P_((void)); |
| 208 | void append_pattern_space P_((void)); |
| 209 | void line_copy P_((struct line *from, struct line *to)); |
| 210 | void line_append P_((struct line *from, struct line *to)); |
| 211 | void str_append P_((struct line *to, char *string, int length)); |
| 212 | void usage P_((void)); |
| 213 | |
| 214 | extern char *myname; |
| 215 | |
| 216 | /* If set, don't write out the line unless explictly told to */ |
| 217 | int no_default_output = 0; |
| 218 | |
| 219 | /* Current input line # */ |
| 220 | int input_line_number = 0; |
| 221 | |
| 222 | /* Are we on the last input file? */ |
| 223 | int last_input_file = 0; |
| 224 | |
| 225 | /* Have we hit EOF on the last input file? This is used to decide if we |
| 226 | have hit the '$' address yet. */ |
| 227 | int input_EOF = 0; |
| 228 | |
| 229 | /* non-zero if a quit command has been executed. */ |
| 230 | int quit_cmd = 0; |
| 231 | |
| 232 | /* Have we done any replacements lately? This is used by the 't' command. */ |
| 233 | int replaced = 0; |
| 234 | |
| 235 | /* How many '{'s are we executing at the moment */ |
| 236 | int program_depth = 0; |
| 237 | |
| 238 | /* The complete compiled SED program that we are going to run */ |
| 239 | struct vector *the_program = 0; |
| 240 | |
| 241 | /* information about labels and jumps-to-labels. This is used to do |
| 242 | the required backpatching after we have compiled all the scripts. */ |
| 243 | struct sed_label *jumps = 0; |
| 244 | struct sed_label *labels = 0; |
| 245 | |
| 246 | /* The 'current' input line. */ |
| 247 | struct line line; |
| 248 | |
| 249 | /* An input line that's been stored by later use by the program */ |
| 250 | struct line hold; |
| 251 | |
| 252 | /* A 'line' to append to the current line when it comes time to write it out */ |
| 253 | struct line append; |
| 254 | |
| 255 | |
| 256 | /* When we're reading a script command from a string, 'prog_start' and |
| 257 | 'prog_end' point to the beginning and end of the string. This |
| 258 | would allow us to compile script strings that contain nulls, except |
| 259 | that script strings are only read from the command line, which is |
| 260 | null-terminated */ |
| 261 | char *prog_start; |
| 262 | char *prog_end; |
| 263 | |
| 264 | /* When we're reading a script command from a string, 'prog_cur' points |
| 265 | to the current character in the string */ |
| 266 | char *prog_cur; |
| 267 | |
| 268 | /* This is the name of the current script file. |
| 269 | It is used for error messages. */ |
| 270 | char *prog_name; |
| 271 | |
| 272 | /* This is the current script file. If it is zero, we are reading |
| 273 | from a string stored in 'prog_start' instead. If both 'prog_file' |
| 274 | and 'prog_start' are zero, we're in trouble! */ |
| 275 | FILE *prog_file; |
| 276 | |
| 277 | /* this is the number of the current script line that we're compiling. It is |
| 278 | used to give out useful and informative error messages. */ |
| 279 | int prog_line = 1; |
| 280 | |
| 281 | /* This is the file pointer that we're currently reading data from. It may |
| 282 | be stdin */ |
| 283 | FILE *input_file; |
| 284 | |
| 285 | /* If this variable is non-zero at exit, one or more of the input |
| 286 | files couldn't be opened. */ |
| 287 | |
| 288 | int bad_input = 0; |
| 289 | |
| 290 | /* 'an empty regular expression is equivalent to the last regular |
| 291 | expression read' so we have to keep track of the last regex used. |
| 292 | Here's where we store a pointer to it (it is only malloc()'d once) */ |
| 293 | struct re_pattern_buffer *last_regex; |
| 294 | |
| 295 | /* Various error messages we may want to print */ |
| 296 | static char ONE_ADDR[] = "Command only uses one address"; |
| 297 | static char NO_ADDR[] = "Command doesn't take any addresses"; |
| 298 | static char LINE_JUNK[] = "Extra characters after command"; |
| 299 | static char BAD_EOF[] = "Unexpected End-of-file"; |
| 300 | static char NO_REGEX[] = "No previous regular expression"; |
| 301 | |
| 302 | static struct option longopts[] = |
| 303 | { |
| 304 | {"expression", 1, NULL, 'e'}, |
| 305 | {"file", 1, NULL, 'f'}, |
| 306 | {"quiet", 0, NULL, 'n'}, |
| 307 | {"silent", 0, NULL, 'n'}, |
| 308 | {"version", 0, NULL, 'V'}, |
| 309 | {NULL, 0, NULL, 0} |
| 310 | }; |
| 311 | |
| 312 | /* Yes, the main program, which parses arguments, and does the right |
| 313 | thing with them; it also inits the temporary storage, etc. */ |
| 314 | void |
| 315 | main(argc,argv) |
| 316 | int argc; |
| 317 | char **argv; |
| 318 | { |
| 319 | int opt; |
| 320 | char *e_strings = NULL; |
| 321 | int compiled = 0; |
| 322 | struct sed_label *go,*lbl; |
| 323 | |
| 324 | myname=argv[0]; |
| 325 | while((opt=getopt_long(argc,argv,"ne:f:V", longopts, (int *) 0)) |
| 326 | !=EOF) { |
| 327 | switch(opt) { |
| 328 | case 'n': |
| 329 | no_default_output = 1; |
| 330 | break; |
| 331 | case 'e': |
| 332 | if(e_strings == NULL) { |
| 333 | e_strings=ck_malloc(strlen(optarg)+2); |
| 334 | strcpy(e_strings,optarg); |
| 335 | } else { |
| 336 | e_strings=ck_realloc(e_strings,strlen(e_strings)+strlen(optarg)+2); |
| 337 | strcat(e_strings,optarg); |
| 338 | } |
| 339 | strcat(e_strings,"\n"); |
| 340 | compiled = 1; |
| 341 | break; |
| 342 | case 'f': |
| 343 | compile_file(optarg); |
| 344 | compiled = 1; |
| 345 | break; |
| 346 | case 'V': |
| 347 | fprintf(stderr, "%s\n", version_string); |
| 348 | break; |
| 349 | default: |
| 350 | usage(); |
| 351 | } |
| 352 | } |
| 353 | if(e_strings) { |
| 354 | compile_string(e_strings); |
| 355 | free(e_strings); |
| 356 | } |
| 357 | if(!compiled) { |
| 358 | if (optind == argc) |
| 359 | usage(); |
| 360 | compile_string(argv[optind++]); |
| 361 | } |
| 362 | |
| 363 | for(go=jumps;go;go=go->next) { |
| 364 | for(lbl=labels;lbl;lbl=lbl->next) |
| 365 | if(!strcmp(lbl->name,go->name)) |
| 366 | break; |
| 367 | if(*go->name && !lbl) |
| 368 | panic("Can't find label for jump to '%s'",go->name); |
| 369 | go->v->v[go->v_index].x.jump=lbl; |
| 370 | } |
| 371 | |
| 372 | line.length=0; |
| 373 | line.alloc=50; |
| 374 | line.text=ck_malloc(50); |
| 375 | |
| 376 | append.length=0; |
| 377 | append.alloc=50; |
| 378 | append.text=ck_malloc(50); |
| 379 | |
| 380 | hold.length=0; |
| 381 | hold.alloc=50; |
| 382 | hold.text=ck_malloc(50); |
| 383 | |
| 384 | if(argc<=optind) { |
| 385 | last_input_file++; |
| 386 | read_file("-"); |
| 387 | } else while(optind<argc) { |
| 388 | if(optind==argc-1) |
| 389 | last_input_file++; |
| 390 | read_file(argv[optind]); |
| 391 | optind++; |
| 392 | if(quit_cmd) |
| 393 | break; |
| 394 | } |
| 395 | if(bad_input) |
| 396 | exit(2); |
| 397 | exit(0); |
| 398 | } |
| 399 | |
| 400 | /* 'str' is a string (from the command line) that contains a sed command. |
| 401 | Compile the command, and add it to the end of 'the_program' */ |
| 402 | void |
| 403 | compile_string(str) |
| 404 | char *str; |
| 405 | { |
| 406 | prog_file = 0; |
| 407 | prog_line=0; |
| 408 | prog_start=prog_cur=str; |
| 409 | prog_end=str+strlen(str); |
| 410 | the_program=compile_program(the_program); |
| 411 | } |
| 412 | |
| 413 | /* 'str' is the name of a file containing sed commands. Read them in |
| 414 | and add them to the end of 'the_program' */ |
| 415 | void |
| 416 | compile_file(str) |
| 417 | char *str; |
| 418 | { |
| 419 | int ch; |
| 420 | |
| 421 | prog_start=prog_cur=prog_end=0; |
| 422 | prog_name=str; |
| 423 | prog_line=1; |
| 424 | if(str[0]=='-' && str[1]=='\0') |
| 425 | prog_file=stdin; |
| 426 | else |
| 427 | prog_file=ck_fopen(str,"r"); |
| 428 | ch=getc(prog_file); |
| 429 | if(ch=='#') { |
| 430 | ch=getc(prog_file); |
| 431 | if(ch=='n') |
| 432 | no_default_output++; |
| 433 | while(ch!=EOF && ch!='\n') |
| 434 | ch=getc(prog_file); |
| 435 | } else if(ch!=EOF) |
| 436 | ungetc(ch,prog_file); |
| 437 | the_program=compile_program(the_program); |
| 438 | } |
| 439 | |
| 440 | #define MORE_CMDS 40 |
| 441 | |
| 442 | /* Read a program (or a subprogram within '{' '}' pairs) in and store |
| 443 | the compiled form in *'vector' Return a pointer to the new vector. */ |
| 444 | struct vector * |
| 445 | compile_program(vector) |
| 446 | struct vector *vector; |
| 447 | { |
| 448 | struct sed_cmd *cur_cmd; |
| 449 | int ch; |
| 450 | int slash; |
| 451 | VOID *b; |
| 452 | unsigned char *string; |
| 453 | int num; |
| 454 | |
| 455 | if(!vector) { |
| 456 | vector=(struct vector *)ck_malloc(sizeof(struct vector)); |
| 457 | vector->v=(struct sed_cmd *)ck_malloc(MORE_CMDS*sizeof(struct sed_cmd)); |
| 458 | vector->v_allocated=MORE_CMDS; |
| 459 | vector->v_length=0; |
| 460 | vector->up_one = 0; |
| 461 | vector->next_one = 0; |
| 462 | } |
| 463 | for(;;) { |
| 464 | skip_comment: |
| 465 | do ch=inchar(); |
| 466 | while(ch!=EOF && (isblank(ch) || ch=='\n' || ch==';')); |
| 467 | if(ch==EOF) |
| 468 | break; |
| 469 | savchar(ch); |
| 470 | |
| 471 | if(vector->v_length==vector->v_allocated) { |
| 472 | vector->v=(struct sed_cmd *)ck_realloc((VOID *)vector->v,(vector->v_length+MORE_CMDS)*sizeof(struct sed_cmd)); |
| 473 | vector->v_allocated+=MORE_CMDS; |
| 474 | } |
| 475 | cur_cmd=vector->v+vector->v_length; |
| 476 | vector->v_length++; |
| 477 | |
| 478 | cur_cmd->a1.addr_type=0; |
| 479 | cur_cmd->a2.addr_type=0; |
| 480 | cur_cmd->aflags=0; |
| 481 | cur_cmd->cmd=0; |
| 482 | |
| 483 | if(compile_address(&(cur_cmd->a1))) { |
| 484 | ch=inchar(); |
| 485 | if(ch==',') { |
| 486 | do ch=inchar(); |
| 487 | while(ch!=EOF && isblank(ch)); |
| 488 | savchar(ch); |
| 489 | if(compile_address(&(cur_cmd->a2))) |
| 490 | ; |
| 491 | else |
| 492 | bad_prog("Unexpected ','"); |
| 493 | } else |
| 494 | savchar(ch); |
| 495 | } |
| 496 | ch=inchar(); |
| 497 | if(ch==EOF) |
| 498 | break; |
| 499 | new_cmd: |
| 500 | switch(ch) { |
| 501 | case '#': |
| 502 | if(cur_cmd->a1.addr_type!=0) |
| 503 | bad_prog(NO_ADDR); |
| 504 | do ch=inchar(); |
| 505 | while(ch!=EOF && ch!='\n'); |
| 506 | vector->v_length--; |
| 507 | goto skip_comment; |
| 508 | case '!': |
| 509 | if(cur_cmd->aflags & ADDR_BANG_BIT) |
| 510 | bad_prog("Multiple '!'s"); |
| 511 | cur_cmd->aflags|= ADDR_BANG_BIT; |
| 512 | do ch=inchar(); |
| 513 | while(ch!=EOF && isblank(ch)); |
| 514 | if(ch==EOF) |
| 515 | bad_prog(BAD_EOF); |
| 516 | #if 0 |
| 517 | savchar(ch); |
| 518 | #endif |
| 519 | goto new_cmd; |
| 520 | case 'a': |
| 521 | case 'i': |
| 522 | if(cur_cmd->a2.addr_type!=0) |
| 523 | bad_prog(ONE_ADDR); |
| 524 | /* Fall Through */ |
| 525 | case 'c': |
| 526 | cur_cmd->cmd=ch; |
| 527 | if(inchar()!='\\' || inchar()!='\n') |
| 528 | bad_prog(LINE_JUNK); |
| 529 | b=init_buffer(); |
| 530 | while((ch=inchar())!=EOF && ch!='\n') { |
| 531 | if(ch=='\\') |
| 532 | ch=inchar(); |
| 533 | add1_buffer(b,ch); |
| 534 | } |
| 535 | if(ch!=EOF) |
| 536 | add1_buffer(b,ch); |
| 537 | num=size_buffer(b); |
| 538 | string=(unsigned char *)ck_malloc(num); |
| 539 | bcopy(get_buffer(b),string,num); |
| 540 | flush_buffer(b); |
| 541 | cur_cmd->x.cmd_txt.text_len=num; |
| 542 | cur_cmd->x.cmd_txt.text=(char *)string; |
| 543 | break; |
| 544 | case '{': |
| 545 | cur_cmd->cmd=ch; |
| 546 | program_depth++; |
| 547 | #if 0 |
| 548 | while((ch=inchar())!=EOF && ch!='\n') |
| 549 | if(!isblank(ch)) |
| 550 | bad_prog(LINE_JUNK); |
| 551 | #endif |
| 552 | cur_cmd->x.sub=compile_program((struct vector *)0); |
| 553 | /* FOO JF is this the right thing to do? */ |
| 554 | break; |
| 555 | case '}': |
| 556 | if(!program_depth) |
| 557 | bad_prog("Unexpected '}'"); |
| 558 | --(vector->v_length); |
| 559 | while((ch=inchar())!=EOF && ch!='\n' && ch!=';') |
| 560 | if(!isblank(ch)) |
| 561 | bad_prog(LINE_JUNK); |
| 562 | return vector; |
| 563 | case ':': |
| 564 | cur_cmd->cmd=ch; |
| 565 | if(cur_cmd->a1.addr_type!=0) |
| 566 | bad_prog(": doesn't want any addresses"); |
| 567 | labels=setup_jump(labels,cur_cmd,vector); |
| 568 | break; |
| 569 | case 'b': |
| 570 | case 't': |
| 571 | cur_cmd->cmd=ch; |
| 572 | jumps=setup_jump(jumps,cur_cmd,vector); |
| 573 | break; |
| 574 | case 'q': |
| 575 | case '=': |
| 576 | if(cur_cmd->a2.addr_type) |
| 577 | bad_prog(ONE_ADDR); |
| 578 | /* Fall Through */ |
| 579 | case 'd': |
| 580 | case 'D': |
| 581 | case 'g': |
| 582 | case 'G': |
| 583 | case 'h': |
| 584 | case 'H': |
| 585 | case 'l': |
| 586 | case 'n': |
| 587 | case 'N': |
| 588 | case 'p': |
| 589 | case 'P': |
| 590 | case 'x': |
| 591 | cur_cmd->cmd=ch; |
| 592 | do ch=inchar(); |
| 593 | while(ch!=EOF && isblank(ch) && ch!='\n' && ch!=';'); |
| 594 | if(ch!='\n' && ch!=';' && ch!=EOF) |
| 595 | bad_prog(LINE_JUNK); |
| 596 | break; |
| 597 | |
| 598 | case 'r': |
| 599 | if(cur_cmd->a2.addr_type!=0) |
| 600 | bad_prog(ONE_ADDR); |
| 601 | /* FALL THROUGH */ |
| 602 | case 'w': |
| 603 | cur_cmd->cmd=ch; |
| 604 | cur_cmd->x.io_file=compile_filename(ch=='r'); |
| 605 | break; |
| 606 | |
| 607 | case 's': |
| 608 | cur_cmd->cmd=ch; |
| 609 | slash=inchar(); |
| 610 | compile_regex(slash); |
| 611 | |
| 612 | cur_cmd->x.cmd_regex.regx=last_regex; |
| 613 | |
| 614 | b=init_buffer(); |
| 615 | while((ch=inchar())!=EOF && ch!=slash) { |
| 616 | if(ch=='\\') { |
| 617 | int ci; |
| 618 | |
| 619 | ci=inchar(); |
| 620 | if(ci!=EOF) { |
| 621 | if(ci!='\n') |
| 622 | add1_buffer(b,ch); |
| 623 | add1_buffer(b,ci); |
| 624 | } |
| 625 | } else |
| 626 | add1_buffer(b,ch); |
| 627 | } |
| 628 | cur_cmd->x.cmd_regex.replace_length=size_buffer(b); |
| 629 | cur_cmd->x.cmd_regex.replacement=ck_malloc(cur_cmd->x.cmd_regex.replace_length); |
| 630 | bcopy(get_buffer(b),cur_cmd->x.cmd_regex.replacement,cur_cmd->x.cmd_regex.replace_length); |
| 631 | flush_buffer(b); |
| 632 | |
| 633 | cur_cmd->x.cmd_regex.flags=0; |
| 634 | cur_cmd->x.cmd_regex.numb=0; |
| 635 | |
| 636 | if(ch==EOF) |
| 637 | break; |
| 638 | do { |
| 639 | ch=inchar(); |
| 640 | switch(ch) { |
| 641 | case 'p': |
| 642 | if(cur_cmd->x.cmd_regex.flags&S_PRINT_BIT) |
| 643 | bad_prog("multiple 'p' options to 's' command"); |
| 644 | cur_cmd->x.cmd_regex.flags|=S_PRINT_BIT; |
| 645 | break; |
| 646 | case 'g': |
| 647 | if(cur_cmd->x.cmd_regex.flags&S_NUM_BIT) |
| 648 | cur_cmd->x.cmd_regex.flags&= ~S_NUM_BIT; |
| 649 | if(cur_cmd->x.cmd_regex.flags&S_GLOBAL_BIT) |
| 650 | bad_prog("multiple 'g' options to 's' command"); |
| 651 | cur_cmd->x.cmd_regex.flags|=S_GLOBAL_BIT; |
| 652 | break; |
| 653 | case 'w': |
| 654 | cur_cmd->x.cmd_regex.flags|=S_WRITE_BIT; |
| 655 | cur_cmd->x.cmd_regex.wio_file=compile_filename(0); |
| 656 | ch='\n'; |
| 657 | break; |
| 658 | case '0': case '1': case '2': case '3': |
| 659 | case '4': case '5': case '6': case '7': |
| 660 | case '8': case '9': |
| 661 | if(cur_cmd->x.cmd_regex.flags&S_NUM_BIT) |
| 662 | bad_prog("multiple number options to 's' command"); |
| 663 | if((cur_cmd->x.cmd_regex.flags&S_GLOBAL_BIT)==0) |
| 664 | cur_cmd->x.cmd_regex.flags|=S_NUM_BIT; |
| 665 | num = 0; |
| 666 | while(isdigit(ch)) { |
| 667 | num=num*10+ch-'0'; |
| 668 | ch=inchar(); |
| 669 | } |
| 670 | savchar(ch); |
| 671 | cur_cmd->x.cmd_regex.numb=num; |
| 672 | break; |
| 673 | case '\n': |
| 674 | case ';': |
| 675 | case EOF: |
| 676 | break; |
| 677 | default: |
| 678 | bad_prog("Unknown option to 's'"); |
| 679 | break; |
| 680 | } |
| 681 | } while(ch!=EOF && ch!='\n' && ch!=';'); |
| 682 | if(ch==EOF) |
| 683 | break; |
| 684 | break; |
| 685 | |
| 686 | case 'y': |
| 687 | cur_cmd->cmd=ch; |
| 688 | string=(unsigned char *)ck_malloc(256); |
| 689 | for(num=0;num<256;num++) |
| 690 | string[num]=num; |
| 691 | b=init_buffer(); |
| 692 | slash=inchar(); |
| 693 | while((ch=inchar())!=EOF && ch!=slash) |
| 694 | add1_buffer(b,ch); |
| 695 | cur_cmd->x.translate=string; |
| 696 | string=(unsigned char *)get_buffer(b); |
| 697 | for(num=size_buffer(b);num;--num) { |
| 698 | ch=inchar(); |
| 699 | if(ch==EOF) |
| 700 | bad_prog(BAD_EOF); |
| 701 | if(ch==slash) |
| 702 | bad_prog("strings for y command are different lengths"); |
| 703 | cur_cmd->x.translate[*string++]=ch; |
| 704 | } |
| 705 | flush_buffer(b); |
| 706 | if(inchar()!=slash || ((ch=inchar())!=EOF && ch!='\n' && ch!=';')) |
| 707 | bad_prog(LINE_JUNK); |
| 708 | break; |
| 709 | |
| 710 | default: |
| 711 | bad_prog("Unknown command"); |
| 712 | } |
| 713 | } |
| 714 | return vector; |
| 715 | } |
| 716 | |
| 717 | /* Complain about a programming error and exit. */ |
| 718 | void |
| 719 | bad_prog(why) |
| 720 | char *why; |
| 721 | { |
| 722 | if(prog_line) |
| 723 | fprintf(stderr,"%s: file %s line %d: %s\n",myname,prog_name,prog_line,why); |
| 724 | else |
| 725 | fprintf(stderr,"%s: %s\n",myname,why); |
| 726 | exit(1); |
| 727 | } |
| 728 | |
| 729 | /* Read the next character from the program. Return EOF if there isn't |
| 730 | anything to read. Keep prog_line up to date, so error messages can |
| 731 | be meaningful. */ |
| 732 | int |
| 733 | inchar() |
| 734 | { |
| 735 | int ch; |
| 736 | if(prog_file) { |
| 737 | if(feof(prog_file)) |
| 738 | return EOF; |
| 739 | else |
| 740 | ch=getc(prog_file); |
| 741 | } else { |
| 742 | if(!prog_cur) |
| 743 | return EOF; |
| 744 | else if(prog_cur==prog_end) { |
| 745 | ch=EOF; |
| 746 | prog_cur=0; |
| 747 | } else |
| 748 | ch= *prog_cur++; |
| 749 | } |
| 750 | if(ch=='\n' && prog_line) |
| 751 | prog_line++; |
| 752 | return ch; |
| 753 | } |
| 754 | |
| 755 | /* unget 'ch' so the next call to inchar will return it. 'ch' must not be |
| 756 | EOF or anything nasty like that. */ |
| 757 | void |
| 758 | savchar(ch) |
| 759 | int ch; |
| 760 | { |
| 761 | if(ch==EOF) |
| 762 | return; |
| 763 | if(ch=='\n' && prog_line>1) |
| 764 | --prog_line; |
| 765 | if(prog_file) |
| 766 | ungetc(ch,prog_file); |
| 767 | else |
| 768 | *--prog_cur=ch; |
| 769 | } |
| 770 | |
| 771 | |
| 772 | /* Try to read an address for a sed command. If it succeeeds, |
| 773 | return non-zero and store the resulting address in *'addr'. |
| 774 | If the input doesn't look like an address read nothing |
| 775 | and return zero. */ |
| 776 | int |
| 777 | compile_address(addr) |
| 778 | struct addr *addr; |
| 779 | { |
| 780 | int ch; |
| 781 | int num; |
| 782 | |
| 783 | ch=inchar(); |
| 784 | |
| 785 | if(isdigit(ch)) { |
| 786 | num=ch-'0'; |
| 787 | while((ch=inchar())!=EOF && isdigit(ch)) |
| 788 | num=num*10+ch-'0'; |
| 789 | while(ch!=EOF && isblank(ch)) |
| 790 | ch=inchar(); |
| 791 | savchar(ch); |
| 792 | addr->addr_type=ADDR_NUM; |
| 793 | addr->addr_number = num; |
| 794 | return 1; |
| 795 | } else if(ch=='/') { |
| 796 | addr->addr_type=ADDR_REGEX; |
| 797 | compile_regex('/'); |
| 798 | addr->addr_regex=last_regex; |
| 799 | do ch=inchar(); |
| 800 | while(ch!=EOF && isblank(ch)); |
| 801 | savchar(ch); |
| 802 | return 1; |
| 803 | } else if(ch=='$') { |
| 804 | addr->addr_type=ADDR_LAST; |
| 805 | do ch=inchar(); |
| 806 | while(ch!=EOF && isblank(ch)); |
| 807 | savchar(ch); |
| 808 | return 1; |
| 809 | } else |
| 810 | savchar(ch); |
| 811 | return 0; |
| 812 | } |
| 813 | |
| 814 | void |
| 815 | compile_regex (slash) |
| 816 | int slash; |
| 817 | { |
| 818 | VOID *b; |
| 819 | int ch; |
| 820 | int in_char_class = 0; |
| 821 | |
| 822 | b=init_buffer(); |
| 823 | while((ch=inchar())!=EOF && (ch!=slash || in_char_class)) { |
| 824 | if(ch=='^') { |
| 825 | if(size_buffer(b)==0) { |
| 826 | add1_buffer(b,'\\'); |
| 827 | add1_buffer(b,'`'); |
| 828 | } else |
| 829 | add1_buffer(b,ch); |
| 830 | continue; |
| 831 | } else if(ch=='$') { |
| 832 | ch=inchar(); |
| 833 | savchar(ch); |
| 834 | if(ch==slash) { |
| 835 | add1_buffer(b,'\\'); |
| 836 | add1_buffer(b,'\''); |
| 837 | } else |
| 838 | add1_buffer(b,'$'); |
| 839 | continue; |
| 840 | } else if(ch == '[') { |
| 841 | add1_buffer(b,ch); |
| 842 | in_char_class = 1; |
| 843 | continue; |
| 844 | } else if(ch == ']') { |
| 845 | add1_buffer(b,ch); |
| 846 | in_char_class = 0; |
| 847 | continue; |
| 848 | } else if(ch!='\\') { |
| 849 | add1_buffer(b,ch); |
| 850 | continue; |
| 851 | } |
| 852 | ch=inchar(); |
| 853 | switch(ch) { |
| 854 | case 'n': |
| 855 | add1_buffer(b,'\n'); |
| 856 | break; |
| 857 | #if 0 |
| 858 | case 'b': |
| 859 | add1_buffer(b,'\b'); |
| 860 | break; |
| 861 | case 'f': |
| 862 | add1_buffer(b,'\f'); |
| 863 | break; |
| 864 | case 'r': |
| 865 | add1_buffer(b,'\r'); |
| 866 | break; |
| 867 | case 't': |
| 868 | add1_buffer(b,'\t'); |
| 869 | break; |
| 870 | #endif /* 0 */ |
| 871 | case EOF: |
| 872 | break; |
| 873 | default: |
| 874 | add1_buffer(b,'\\'); |
| 875 | add1_buffer(b,ch); |
| 876 | break; |
| 877 | } |
| 878 | } |
| 879 | if(ch==EOF) |
| 880 | bad_prog(BAD_EOF); |
| 881 | if(size_buffer(b)) { |
| 882 | last_regex=(struct re_pattern_buffer *)ck_malloc(sizeof(struct re_pattern_buffer)); |
| 883 | last_regex->allocated=size_buffer(b)+10; |
| 884 | last_regex->buffer=ck_malloc(last_regex->allocated); |
| 885 | last_regex->fastmap=ck_malloc(256); |
| 886 | last_regex->translate=0; |
| 887 | re_compile_pattern(get_buffer(b),size_buffer(b),last_regex); |
| 888 | } else if(!last_regex) |
| 889 | bad_prog(NO_REGEX); |
| 890 | flush_buffer(b); |
| 891 | } |
| 892 | |
| 893 | /* Store a label (or label reference) created by a ':', 'b', or 't' |
| 894 | comand so that the jump to/from the lable can be backpatched after |
| 895 | compilation is complete */ |
| 896 | struct sed_label * |
| 897 | setup_jump(list,cmd,vec) |
| 898 | struct sed_label *list; |
| 899 | struct sed_cmd *cmd; |
| 900 | struct vector *vec; |
| 901 | { |
| 902 | struct sed_label *tmp; |
| 903 | VOID *b; |
| 904 | int ch; |
| 905 | |
| 906 | b=init_buffer(); |
| 907 | while((ch=inchar()) != EOF && isblank(ch)) |
| 908 | ; |
| 909 | while(ch!=EOF && ch!='\n') { |
| 910 | add1_buffer(b,ch); |
| 911 | ch=inchar(); |
| 912 | } |
| 913 | savchar(ch); |
| 914 | add1_buffer(b,'\0'); |
| 915 | tmp=(struct sed_label *)ck_malloc(sizeof(struct sed_label)); |
| 916 | tmp->v=vec; |
| 917 | tmp->v_index=cmd-vec->v; |
| 918 | tmp->name=ck_strdup(get_buffer(b)); |
| 919 | tmp->next=list; |
| 920 | flush_buffer(b); |
| 921 | return tmp; |
| 922 | } |
| 923 | |
| 924 | /* read in a filename for a 'r', 'w', or 's///w' command, and |
| 925 | update the internal structure about files. The file is |
| 926 | opened if it isn't already open. */ |
| 927 | FILE * |
| 928 | compile_filename(readit) |
| 929 | int readit; |
| 930 | { |
| 931 | char *file_name; |
| 932 | int n; |
| 933 | VOID *b; |
| 934 | int ch; |
| 935 | |
| 936 | if(inchar()!=' ') |
| 937 | bad_prog("missing ' ' before filename"); |
| 938 | b=init_buffer(); |
| 939 | while((ch=inchar())!=EOF && ch!='\n') |
| 940 | add1_buffer(b,ch); |
| 941 | add1_buffer(b,'\0'); |
| 942 | file_name=get_buffer(b); |
| 943 | for(n=0;n<NUM_FPS;n++) { |
| 944 | if(!file_ptrs[n].name) |
| 945 | break; |
| 946 | if(!strcmp(file_ptrs[n].name,file_name)) { |
| 947 | if(file_ptrs[n].readit!=readit) |
| 948 | bad_prog("Can't open file for both reading and writing"); |
| 949 | flush_buffer(b); |
| 950 | return file_ptrs[n].phile; |
| 951 | } |
| 952 | } |
| 953 | if(n<NUM_FPS) { |
| 954 | file_ptrs[n].name=ck_strdup(file_name); |
| 955 | file_ptrs[n].readit=readit; |
| 956 | if (!readit) |
| 957 | file_ptrs[n].phile=ck_fopen(file_name,"a"); |
| 958 | else if (access(file_name, 4) == 0) |
| 959 | file_ptrs[n].phile=ck_fopen(file_name,"r"); |
| 960 | else |
| 961 | file_ptrs[n].phile=ck_fopen("/dev/null", "r"); |
| 962 | flush_buffer(b); |
| 963 | return file_ptrs[n].phile; |
| 964 | } else { |
| 965 | bad_prog("Hopelessely evil compiled in limit on number of open files. re-compile sed"); |
| 966 | return 0; |
| 967 | } |
| 968 | } |
| 969 | |
| 970 | /* Parse a filename given by a 'r' 'w' or 's///w' command. */ |
| 971 | void |
| 972 | read_file(name) |
| 973 | char *name; |
| 974 | { |
| 975 | if(*name=='-' && name[1]=='\0') |
| 976 | input_file=stdin; |
| 977 | else { |
| 978 | input_file=fopen(name,"r"); |
| 979 | if(input_file==0) { |
| 980 | extern int errno; |
| 981 | extern char *sys_errlist[]; |
| 982 | extern int sys_nerr; |
| 983 | |
| 984 | char *ptr; |
| 985 | |
| 986 | ptr=(errno>=0 && errno<sys_nerr) ? sys_errlist[errno] : "Unknown error code"; |
| 987 | bad_input++; |
| 988 | fprintf(stderr,"%s: can't read %s: %s\n",myname,name,ptr); |
| 989 | |
| 990 | return; |
| 991 | } |
| 992 | } |
| 993 | while(read_pattern_space()) { |
| 994 | execute_program(the_program); |
| 995 | if(!no_default_output) |
| 996 | ck_fwrite(line.text,1,line.length,stdout); |
| 997 | if(append.length) { |
| 998 | ck_fwrite(append.text,1,append.length,stdout); |
| 999 | append.length=0; |
| 1000 | } |
| 1001 | if(quit_cmd) |
| 1002 | break; |
| 1003 | } |
| 1004 | ck_fclose(input_file); |
| 1005 | } |
| 1006 | |
| 1007 | /* Execute the program 'vec' on the current input line. */ |
| 1008 | void |
| 1009 | execute_program(vec) |
| 1010 | struct vector *vec; |
| 1011 | { |
| 1012 | struct sed_cmd *cur_cmd; |
| 1013 | int n; |
| 1014 | int addr_matched; |
| 1015 | static int end_cycle; |
| 1016 | |
| 1017 | int start; |
| 1018 | int remain; |
| 1019 | int offset; |
| 1020 | |
| 1021 | static struct line tmp; |
| 1022 | struct line t; |
| 1023 | char *rep,*rep_end,*rep_next,*rep_cur; |
| 1024 | |
| 1025 | struct re_registers regs; |
| 1026 | int count = 0; |
| 1027 | |
| 1028 | end_cycle = 0; |
| 1029 | |
| 1030 | for(cur_cmd=vec->v,n=vec->v_length;n;cur_cmd++,n--) { |
| 1031 | |
| 1032 | exe_loop: |
| 1033 | addr_matched=0; |
| 1034 | if(cur_cmd->aflags&A1_MATCHED_BIT) { |
| 1035 | addr_matched=1; |
| 1036 | if(match_address(&(cur_cmd->a2))) |
| 1037 | cur_cmd->aflags&=~A1_MATCHED_BIT; |
| 1038 | } else if(match_address(&(cur_cmd->a1))) { |
| 1039 | addr_matched=1; |
| 1040 | if(cur_cmd->a2.addr_type!=ADDR_NULL) |
| 1041 | cur_cmd->aflags|=A1_MATCHED_BIT; |
| 1042 | } |
| 1043 | if(cur_cmd->aflags&ADDR_BANG_BIT) |
| 1044 | addr_matched= !addr_matched; |
| 1045 | if(!addr_matched) |
| 1046 | continue; |
| 1047 | switch(cur_cmd->cmd) { |
| 1048 | case '{': /* Execute sub-program */ |
| 1049 | execute_program(cur_cmd->x.sub); |
| 1050 | break; |
| 1051 | |
| 1052 | case ':': /* Executing labels is easy. */ |
| 1053 | break; |
| 1054 | |
| 1055 | case '=': |
| 1056 | printf("%d\n",input_line_number); |
| 1057 | break; |
| 1058 | |
| 1059 | case 'a': |
| 1060 | while(append.alloc-append.length<cur_cmd->x.cmd_txt.text_len) { |
| 1061 | append.alloc *= 2; |
| 1062 | append.text=ck_realloc(append.text,append.alloc); |
| 1063 | } |
| 1064 | bcopy(cur_cmd->x.cmd_txt.text,append.text+append.length,cur_cmd->x.cmd_txt.text_len); |
| 1065 | append.length+=cur_cmd->x.cmd_txt.text_len; |
| 1066 | break; |
| 1067 | |
| 1068 | case 'b': |
| 1069 | if(!cur_cmd->x.jump) |
| 1070 | end_cycle++; |
| 1071 | else { |
| 1072 | struct sed_label *j = cur_cmd->x.jump; |
| 1073 | |
| 1074 | n= j->v->v_length - j->v_index; |
| 1075 | cur_cmd= j->v->v + j->v_index; |
| 1076 | goto exe_loop; |
| 1077 | } |
| 1078 | break; |
| 1079 | |
| 1080 | case 'c': |
| 1081 | line.length=0; |
| 1082 | if(!(cur_cmd->aflags&A1_MATCHED_BIT)) |
| 1083 | ck_fwrite(cur_cmd->x.cmd_txt.text,1,cur_cmd->x.cmd_txt.text_len,stdout); |
| 1084 | end_cycle++; |
| 1085 | break; |
| 1086 | |
| 1087 | case 'd': |
| 1088 | line.length=0; |
| 1089 | end_cycle++; |
| 1090 | break; |
| 1091 | |
| 1092 | case 'D': |
| 1093 | { |
| 1094 | char *tmp; |
| 1095 | int newlength; |
| 1096 | |
| 1097 | tmp=memchr(line.text,'\n',line.length); |
| 1098 | newlength=line.length-(tmp-line.text); |
| 1099 | if(newlength) |
| 1100 | memmove(line.text,tmp,newlength); |
| 1101 | line.length=newlength; |
| 1102 | } |
| 1103 | end_cycle++; |
| 1104 | break; |
| 1105 | |
| 1106 | case 'g': |
| 1107 | line_copy(&hold,&line); |
| 1108 | break; |
| 1109 | |
| 1110 | case 'G': |
| 1111 | line_append(&hold,&line); |
| 1112 | break; |
| 1113 | |
| 1114 | case 'h': |
| 1115 | line_copy(&line,&hold); |
| 1116 | break; |
| 1117 | |
| 1118 | case 'H': |
| 1119 | line_append(&line,&hold); |
| 1120 | break; |
| 1121 | |
| 1122 | case 'i': |
| 1123 | ck_fwrite(cur_cmd->x.cmd_txt.text,1,cur_cmd->x.cmd_txt.text_len,stdout); |
| 1124 | break; |
| 1125 | |
| 1126 | case 'l': |
| 1127 | { |
| 1128 | char *tmp; |
| 1129 | int n; |
| 1130 | int width = 0; |
| 1131 | |
| 1132 | n=line.length; |
| 1133 | tmp=line.text; |
| 1134 | /* Use --n so this'll skip the trailing newline */ |
| 1135 | while(--n) { |
| 1136 | if(width>77) { |
| 1137 | width=0; |
| 1138 | putchar('\n'); |
| 1139 | } |
| 1140 | if(*tmp == '\\') { |
| 1141 | printf("\\\\"); |
| 1142 | width+=2; |
| 1143 | } else if(isprint(*tmp)) { |
| 1144 | putchar(*tmp); |
| 1145 | width++; |
| 1146 | } else switch(*tmp) { |
| 1147 | #if 0 |
| 1148 | /* Should print \00 instead of \0 because (a) POSIX requires it, and |
| 1149 | (b) this way \01 is unambiguous. */ |
| 1150 | case '\0': |
| 1151 | printf("\\0"); |
| 1152 | width+=2; |
| 1153 | break; |
| 1154 | #endif |
| 1155 | case 007: |
| 1156 | printf("\\a"); |
| 1157 | width+=2; |
| 1158 | break; |
| 1159 | case '\b': |
| 1160 | printf("\\b"); |
| 1161 | width+=2; |
| 1162 | break; |
| 1163 | case '\f': |
| 1164 | printf("\\f"); |
| 1165 | width+=2; |
| 1166 | break; |
| 1167 | case '\n': |
| 1168 | printf("\\n"); |
| 1169 | width+=2; |
| 1170 | break; |
| 1171 | case '\r': |
| 1172 | printf("\\r"); |
| 1173 | width+=2; |
| 1174 | break; |
| 1175 | case '\t': |
| 1176 | printf("\\t"); |
| 1177 | width+=2; |
| 1178 | break; |
| 1179 | case '\v': |
| 1180 | printf("\\v"); |
| 1181 | width+=2; |
| 1182 | break; |
| 1183 | default: |
| 1184 | printf("\\%02x",(*tmp)&0xFF); |
| 1185 | width+=2; |
| 1186 | break; |
| 1187 | } |
| 1188 | tmp++; |
| 1189 | } |
| 1190 | putchar('\n'); |
| 1191 | } |
| 1192 | break; |
| 1193 | |
| 1194 | case 'n': |
| 1195 | if (feof(input_file)) goto quit; |
| 1196 | ck_fwrite(line.text,1,line.length,stdout); |
| 1197 | read_pattern_space(); |
| 1198 | break; |
| 1199 | |
| 1200 | case 'N': |
| 1201 | if (feof(input_file)) goto quit; |
| 1202 | append_pattern_space(); |
| 1203 | break; |
| 1204 | |
| 1205 | case 'p': |
| 1206 | ck_fwrite(line.text,1,line.length,stdout); |
| 1207 | break; |
| 1208 | |
| 1209 | case 'P': |
| 1210 | { |
| 1211 | char *tmp; |
| 1212 | |
| 1213 | tmp=memchr(line.text,'\n',line.length); |
| 1214 | ck_fwrite(line.text, 1, |
| 1215 | tmp ? tmp - line.text + 1 |
| 1216 | : line.length, stdout); |
| 1217 | } |
| 1218 | break; |
| 1219 | |
| 1220 | case 'q': quit: |
| 1221 | quit_cmd++; |
| 1222 | end_cycle++; |
| 1223 | break; |
| 1224 | |
| 1225 | case 'r': |
| 1226 | { |
| 1227 | int n = 0; |
| 1228 | |
| 1229 | rewind(cur_cmd->x.io_file); |
| 1230 | do { |
| 1231 | append.length += n; |
| 1232 | if(append.length==append.alloc) { |
| 1233 | append.alloc *= 2; |
| 1234 | append.text = ck_realloc(append.text, append.alloc); |
| 1235 | } |
| 1236 | } while((n=fread(append.text+append.length,sizeof(char),append.alloc-append.length,cur_cmd->x.io_file))>0); |
| 1237 | if(ferror(cur_cmd->x.io_file)) |
| 1238 | panic("Read error on input file to 'r' command"); |
| 1239 | } |
| 1240 | break; |
| 1241 | |
| 1242 | case 's': |
| 1243 | if(!tmp.alloc) { |
| 1244 | tmp.alloc=50; |
| 1245 | tmp.text=ck_malloc(50); |
| 1246 | } |
| 1247 | count=0; |
| 1248 | start = 0; |
| 1249 | remain=line.length-1; |
| 1250 | tmp.length=0; |
| 1251 | rep = cur_cmd->x.cmd_regex.replacement; |
| 1252 | rep_end=rep+cur_cmd->x.cmd_regex.replace_length; |
| 1253 | |
| 1254 | while((offset = re_search(cur_cmd->x.cmd_regex.regx, |
| 1255 | line.text, |
| 1256 | line.length-1, |
| 1257 | start, |
| 1258 | remain, |
| 1259 | ®s))>=0) { |
| 1260 | count++; |
| 1261 | if(offset-start) |
| 1262 | str_append(&tmp,line.text+start,offset-start); |
| 1263 | |
| 1264 | if(cur_cmd->x.cmd_regex.flags&S_NUM_BIT) { |
| 1265 | if(count!=cur_cmd->x.cmd_regex.numb) { |
| 1266 | str_append(&tmp,line.text+regs.start[0],regs.end[0]-regs.start[0]); |
| 1267 | start = (offset == regs.end[0] ? offset + 1 : regs.end[0]); |
| 1268 | remain = (line.length-1) - start; |
| 1269 | continue; |
| 1270 | } |
| 1271 | } |
| 1272 | |
| 1273 | for(rep_next=rep_cur=rep;rep_next<rep_end;rep_next++) { |
| 1274 | if(*rep_next=='&') { |
| 1275 | if(rep_next-rep_cur) |
| 1276 | str_append(&tmp,rep_cur,rep_next-rep_cur); |
| 1277 | str_append(&tmp,line.text+regs.start[0],regs.end[0]-regs.start[0]); |
| 1278 | rep_cur=rep_next+1; |
| 1279 | } else if(*rep_next=='\\') { |
| 1280 | if(rep_next-rep_cur) |
| 1281 | str_append(&tmp,rep_cur,rep_next-rep_cur); |
| 1282 | rep_next++; |
| 1283 | if(rep_next!=rep_end) { |
| 1284 | int n; |
| 1285 | |
| 1286 | if(*rep_next>='0' && *rep_next<='9') { |
| 1287 | n= *rep_next -'0'; |
| 1288 | str_append(&tmp,line.text+regs.start[n],regs.end[n]-regs.start[n]); |
| 1289 | } else |
| 1290 | str_append(&tmp,rep_next,1); |
| 1291 | } |
| 1292 | rep_cur=rep_next+1; |
| 1293 | } |
| 1294 | } |
| 1295 | if(rep_next-rep_cur) |
| 1296 | str_append(&tmp,rep_cur,rep_next-rep_cur); |
| 1297 | if (offset == regs.end[0]) { |
| 1298 | str_append(&tmp, line.text + offset, 1); |
| 1299 | ++regs.end[0]; |
| 1300 | } |
| 1301 | start = regs.end[0]; |
| 1302 | |
| 1303 | remain = (line.length-1) - start; |
| 1304 | if(remain<0) |
| 1305 | break; |
| 1306 | if(!(cur_cmd->x.cmd_regex.flags&S_GLOBAL_BIT)) |
| 1307 | break; |
| 1308 | } |
| 1309 | if(!count) |
| 1310 | break; |
| 1311 | replaced=1; |
| 1312 | str_append(&tmp,line.text+start,remain+1); |
| 1313 | t.text=line.text; |
| 1314 | t.length=line.length; |
| 1315 | t.alloc=line.alloc; |
| 1316 | line.text=tmp.text; |
| 1317 | line.length=tmp.length; |
| 1318 | line.alloc=tmp.alloc; |
| 1319 | tmp.text=t.text; |
| 1320 | tmp.length=t.length; |
| 1321 | tmp.alloc=t.alloc; |
| 1322 | if(cur_cmd->x.cmd_regex.flags&S_WRITE_BIT) |
| 1323 | ck_fwrite(line.text,1,line.length,cur_cmd->x.cmd_regex.wio_file); |
| 1324 | if(cur_cmd->x.cmd_regex.flags&S_PRINT_BIT) |
| 1325 | ck_fwrite(line.text,1,line.length,stdout); |
| 1326 | break; |
| 1327 | |
| 1328 | case 't': |
| 1329 | if(replaced) { |
| 1330 | replaced = 0; |
| 1331 | if(!cur_cmd->x.jump) |
| 1332 | end_cycle++; |
| 1333 | else { |
| 1334 | struct sed_label *j = cur_cmd->x.jump; |
| 1335 | |
| 1336 | n= j->v->v_length - j->v_index; |
| 1337 | cur_cmd= j->v->v + j->v_index; |
| 1338 | goto exe_loop; |
| 1339 | } |
| 1340 | } |
| 1341 | break; |
| 1342 | |
| 1343 | case 'w': |
| 1344 | ck_fwrite(line.text,1,line.length,cur_cmd->x.io_file); |
| 1345 | break; |
| 1346 | |
| 1347 | case 'x': |
| 1348 | { |
| 1349 | struct line tmp; |
| 1350 | |
| 1351 | tmp=line; |
| 1352 | line=hold; |
| 1353 | hold=tmp; |
| 1354 | } |
| 1355 | break; |
| 1356 | |
| 1357 | case 'y': |
| 1358 | { |
| 1359 | unsigned char *p,*e; |
| 1360 | |
| 1361 | for(p=(unsigned char *)(line.text),e=p+line.length;p<e;p++) |
| 1362 | *p=cur_cmd->x.translate[*p]; |
| 1363 | } |
| 1364 | break; |
| 1365 | |
| 1366 | default: |
| 1367 | panic("INTERNAL ERROR: Bad cmd %c",cur_cmd->cmd); |
| 1368 | } |
| 1369 | if(end_cycle) |
| 1370 | break; |
| 1371 | } |
| 1372 | } |
| 1373 | |
| 1374 | |
| 1375 | /* Return non-zero if the current line matches the address |
| 1376 | pointed to by 'addr'. */ |
| 1377 | int |
| 1378 | match_address(addr) |
| 1379 | struct addr *addr; |
| 1380 | { |
| 1381 | switch(addr->addr_type) { |
| 1382 | case ADDR_NULL: |
| 1383 | return 1; |
| 1384 | case ADDR_NUM: |
| 1385 | return (input_line_number==addr->addr_number); |
| 1386 | |
| 1387 | case ADDR_REGEX: |
| 1388 | return (re_search(addr->addr_regex, |
| 1389 | line.text, |
| 1390 | line.length-1, |
| 1391 | 0, |
| 1392 | line.length-1, |
| 1393 | (struct re_registers *)0)>=0) ? 1 : 0; |
| 1394 | |
| 1395 | case ADDR_LAST: |
| 1396 | return (input_EOF) ? 1 : 0; |
| 1397 | |
| 1398 | default: |
| 1399 | panic("INTERNAL ERROR: bad address type"); |
| 1400 | break; |
| 1401 | } |
| 1402 | return -1; |
| 1403 | } |
| 1404 | |
| 1405 | /* Read in the next line of input, and store it in the |
| 1406 | pattern space. Return non-zero if this is the last line of input */ |
| 1407 | |
| 1408 | int |
| 1409 | read_pattern_space() |
| 1410 | { |
| 1411 | int n; |
| 1412 | char *p; |
| 1413 | int ch; |
| 1414 | |
| 1415 | p=line.text; |
| 1416 | n=line.alloc; |
| 1417 | |
| 1418 | if(feof(input_file)) |
| 1419 | return 0; |
| 1420 | input_line_number++; |
| 1421 | replaced=0; |
| 1422 | for(;;) { |
| 1423 | if(n==0) { |
| 1424 | line.text=ck_realloc(line.text,line.alloc*2); |
| 1425 | p=line.text+line.alloc; |
| 1426 | n=line.alloc; |
| 1427 | line.alloc*=2; |
| 1428 | } |
| 1429 | ch=getc(input_file); |
| 1430 | if(ch==EOF) { |
| 1431 | if(n==line.alloc) |
| 1432 | return 0; |
| 1433 | *p++='\n'; |
| 1434 | --n; |
| 1435 | line.length=line.alloc-n; |
| 1436 | if(last_input_file) |
| 1437 | input_EOF++; |
| 1438 | return 1; |
| 1439 | } |
| 1440 | *p++=ch; |
| 1441 | --n; |
| 1442 | if(ch=='\n') { |
| 1443 | line.length=line.alloc-n; |
| 1444 | break; |
| 1445 | } |
| 1446 | } |
| 1447 | ch=getc(input_file); |
| 1448 | if(ch!=EOF) |
| 1449 | ungetc(ch,input_file); |
| 1450 | else if(last_input_file) |
| 1451 | input_EOF++; |
| 1452 | return 1; |
| 1453 | } |
| 1454 | |
| 1455 | /* Inplement the 'N' command, which appends the next line of input to |
| 1456 | the pattern space. */ |
| 1457 | void |
| 1458 | append_pattern_space() |
| 1459 | { |
| 1460 | char *p; |
| 1461 | int n; |
| 1462 | int ch; |
| 1463 | |
| 1464 | p=line.text+line.length; |
| 1465 | n=line.alloc-line.length; |
| 1466 | |
| 1467 | input_line_number++; |
| 1468 | replaced=0; |
| 1469 | for(;;) { |
| 1470 | ch=getc(input_file); |
| 1471 | if(ch==EOF) { |
| 1472 | if(n==line.alloc) |
| 1473 | return; |
| 1474 | *p++='\n'; |
| 1475 | --n; |
| 1476 | line.length=line.alloc-n; |
| 1477 | if(last_input_file) |
| 1478 | input_EOF++; |
| 1479 | return; |
| 1480 | } |
| 1481 | *p++=ch; |
| 1482 | --n; |
| 1483 | if(ch=='\n') { |
| 1484 | line.length=line.alloc-n; |
| 1485 | break; |
| 1486 | } |
| 1487 | if(n==0) { |
| 1488 | line.text=ck_realloc(line.text,line.alloc*2); |
| 1489 | p=line.text+line.alloc; |
| 1490 | n=line.alloc; |
| 1491 | line.alloc*=2; |
| 1492 | } |
| 1493 | } |
| 1494 | ch=getc(input_file); |
| 1495 | if(ch!=EOF) |
| 1496 | ungetc(ch,input_file); |
| 1497 | else if(last_input_file) |
| 1498 | input_EOF++; |
| 1499 | } |
| 1500 | |
| 1501 | /* Copy the contents of the line 'from' into the line 'to'. |
| 1502 | This destroys the old contents of 'to'. It will still work |
| 1503 | if the line 'from' contains nulls. */ |
| 1504 | void |
| 1505 | line_copy(from,to) |
| 1506 | struct line *from,*to; |
| 1507 | { |
| 1508 | if(from->length>to->alloc) { |
| 1509 | to->alloc=from->length; |
| 1510 | to->text=ck_realloc(to->text,to->alloc); |
| 1511 | } |
| 1512 | bcopy(from->text,to->text,from->length); |
| 1513 | to->length=from->length; |
| 1514 | } |
| 1515 | |
| 1516 | /* Append the contents of the line 'from' to the line 'to'. |
| 1517 | This routine will work even if the line 'from' contains nulls */ |
| 1518 | void |
| 1519 | line_append(from,to) |
| 1520 | struct line *from,*to; |
| 1521 | { |
| 1522 | if(from->length>(to->alloc-to->length)) { |
| 1523 | to->alloc+=from->length; |
| 1524 | to->text=ck_realloc(to->text,to->alloc); |
| 1525 | } |
| 1526 | bcopy(from->text,to->text+to->length,from->length); |
| 1527 | to->length+=from->length; |
| 1528 | } |
| 1529 | |
| 1530 | /* Append 'length' bytes from 'string' to the line 'to' |
| 1531 | This routine *will* append bytes with nulls in them, without |
| 1532 | failing. */ |
| 1533 | void |
| 1534 | str_append(to,string,length) |
| 1535 | struct line *to; |
| 1536 | char *string; |
| 1537 | int length; |
| 1538 | { |
| 1539 | if(length>to->alloc-to->length) { |
| 1540 | to->alloc+=length; |
| 1541 | to->text=ck_realloc(to->text,to->alloc); |
| 1542 | } |
| 1543 | bcopy(string,to->text+to->length,length); |
| 1544 | to->length+=length; |
| 1545 | } |
| 1546 | |
| 1547 | void |
| 1548 | usage() |
| 1549 | { |
| 1550 | fprintf(stderr, "\ |
| 1551 | Usage: %s [-nV] [+quiet] [+silent] [+version] [-e script] [-f script-file]\n\ |
| 1552 | [+expression=script] [+file=script-file] [file...]\n", myname); |
| 1553 | exit(4); |
| 1554 | } |