2 * sed.c - very minimalist version of sed
4 * Copyright (C) 1999,2000 by Lineo, inc.
5 * Written by Mark Whitley <markw@lineo.com>, <markw@enol.com>
7 * This program is free software; you can redistribute it and/or modify
8 * it under the terms of the GNU General Public License as published by
9 * the Free Software Foundation; either version 2 of the License, or
10 * (at your option) any later version.
12 * This program is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * General Public License for more details.
17 * You should have received a copy of the GNU General Public License
18 * along with this program; if not, write to the Free Software
19 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
24 Supported features and commands in this version of sed:
27 - address matching: num|/matchstr/[,num|/matchstr/|$]command
28 - commands: (p)rint, (d)elete, (s)ubstitue (with g & I flags)
29 - edit commands: (a)ppend, (i)nsert, (c)hange
31 (Note: Specifying an address (range) to match is *optional*; commands
32 default to the whole pattern space if no specific address match was
37 - transliteration (y/source-chars/dest-chars/) (use 'tr')
38 - no support for characters other than the '/' character for regex matches
39 - no pattern space hold space storing / swapping (x, etc.)
40 - no labels / branching (: label, b, t, and friends)
41 - and lots, lots more.
46 #include <stdlib.h> /* for realloc() */
47 #include <unistd.h> /* for getopt() */
49 #include <string.h> /* for strdup() */
51 #include <ctype.h> /* for isspace() */
56 extern int optind; /* in unistd.h */
57 extern char *optarg; /* ditto */
60 static int be_quiet = 0;
65 int beg_line; /* 'sed 1p' 0 == no begining line, apply commands to all lines */
66 int end_line; /* 'sed 1,3p' 0 == no end line, use only beginning. -1 == $ */
67 regex_t *beg_match; /* sed -e '/match/cmd' */
68 regex_t *end_match; /* sed -e '/match/,/end_match/cmd' */
71 char cmd; /* p,d,s (add more at your leisure :-) */
73 /* substitution command specific fields */
74 regex_t *sub_match; /* sed -e 's/sub_match/replace/' */
75 char *replace; /* sed -e 's/sub_match/replace/' XXX: who will hold the \1 \2 \3s? */
76 unsigned int sub_g:1; /* sed -e 's/foo/bar/g' (global) */
78 /* edit command (a,i,c) speicific field */
83 static struct sed_cmd *sed_cmds = NULL; /* growable arrary holding a sequence of sed cmds */
84 static int ncmds = 0; /* number of sed commands */
86 /*static char *cur_file = NULL;*/ /* file currently being processed XXX: do I need this? */
88 static const char sed_usage[] =
89 "sed [-Vhnef] pattern [files...]\n"
90 #ifndef BB_FEATURE_TRIVIAL_HELP
92 "-n\tsuppress automatic printing of pattern space\n"
93 "-e script\tadd the script to the commands to be executed\n"
94 "-f scriptfile\tadd the contents of script-file to the commands to be executed\n"
95 "-h\tdisplay this help message\n"
96 "-V\toutput version information and exit\n"
98 "If no -e or -f is given, the first non-option argument is taken as the\n"
99 "sed script to interpret. All remaining arguments are names of input\n"
100 "files; if no input files are specified, then the standard input is read.\n"
105 static void destroy_cmd_strs()
107 if (sed_cmds == NULL)
110 /* destroy all the elements in the array */
111 while (--ncmds >= 0) {
113 if (sed_cmds[ncmds].beg_match) {
114 regfree(sed_cmds[ncmds].beg_match);
115 free(sed_cmds[ncmds].beg_match);
117 if (sed_cmds[ncmds].end_match) {
118 regfree(sed_cmds[ncmds].end_match);
119 free(sed_cmds[ncmds].end_match);
121 if (sed_cmds[ncmds].sub_match) {
122 regfree(sed_cmds[ncmds].sub_match);
123 free(sed_cmds[ncmds].sub_match);
125 if (sed_cmds[ncmds].replace)
126 free(sed_cmds[ncmds].replace);
129 /* destroy the array */
137 * trim_str - trims leading and trailing space from a string
139 * Note: This returns a malloc'ed string so you must store and free it
140 * XXX: This should be in the utility.c file.
141 * XXX: This is now obsolete. Maybe it belongs nowhere.
143 static char *trim_str(const char *str)
146 char *retstr = strdup(str);
148 /* trim leading whitespace */
149 memmove(retstr, &retstr[strspn(retstr, " \n\t\v")], strlen(retstr));
151 /* trim trailing whitespace */
152 i = strlen(retstr) - 1;
153 while (isspace(retstr[i]))
159 * you know, a strrspn() would really be nice cuz then we could say:
161 * retstr[strrspn(retstr, " \n\t\v") + 1] = 0;
170 * strrspn - works just like strspn() but goes from right to left instead of
173 static size_t strrspn(const char *s, const char *accept)
175 size_t i = strlen(s);
177 while (strchr(accept, s[--i]))
185 * index_of_unescaped_slash - walks left to right through a string beginning
186 * at a specified index and returns the index of the next unescaped slash.
188 static int index_of_next_unescaped_slash(const char *str, int idx)
192 /* test if we've hit the end */
195 } while (str[idx] != '/' && str[idx - 1] != '\\');
201 * returns the index in the string just past where the address ends.
203 static int get_address(const char *str, int *line, regex_t **regex)
205 char *my_str = strdup(str);
208 if (isdigit(my_str[idx])) {
211 } while (isdigit(my_str[idx]));
213 *line = atoi(my_str);
215 else if (my_str[idx] == '$') {
219 else if (my_str[idx] == '/') {
220 idx = index_of_next_unescaped_slash(my_str, idx);
222 fatalError("unterminated match expression\n");
224 *regex = (regex_t *)xmalloc(sizeof(regex_t));
225 xregcomp(*regex, my_str+1, REG_NEWLINE);
226 idx++; /* so it points to the next character after the last '/' */
229 errorMsg("get_address: no address found in string\n"
230 "\t(you probably didn't check the string you passed me)\n");
238 static char *strdup_substr(const char *str, int start, int end)
240 int size = end - start + 1;
241 char *newstr = xmalloc(size);
242 memcpy(newstr, str+start, size-1);
243 newstr[size-1] = '\0';
247 static int parse_subst_cmd(struct sed_cmd *sed_cmd, const char *substr)
249 int oldidx, cflags = REG_NEWLINE;
254 * the string that gets passed to this function should look like this:
259 * (all three of the '/' slashes are mandatory)
262 /* verify that the 's' is followed by a 'slash' */
263 if (substr[++idx] != '/')
264 fatalError("bad format in substitution expression\n");
266 /* save the match string */
268 idx = index_of_next_unescaped_slash(substr, idx);
270 fatalError("bad format in substitution expression\n");
271 match = strdup_substr(substr, oldidx, idx);
273 /* save the replacement string */
275 idx = index_of_next_unescaped_slash(substr, idx);
277 fatalError("bad format in substitution expression\n");
278 sed_cmd->replace = strdup_substr(substr, oldidx, idx);
280 /* process the flags */
281 while (substr[++idx]) {
282 switch (substr[idx]) {
290 /* any whitespace or semicolon trailing after a s/// is ok */
291 if (strchr("; \t\v\n\r", substr[idx]))
294 fatalError("bad option in substitution expression\n");
299 /* compile the regex */
300 sed_cmd->sub_match = (regex_t *)xmalloc(sizeof(regex_t));
301 xregcomp(sed_cmd->sub_match, match, cflags);
307 static int parse_edit_cmd(struct sed_cmd *sed_cmd, const char *editstr)
310 int slashes_eaten = 0;
311 char *ptr; /* shorthand */
314 * the string that gets passed to this function should look like this:
318 * | this backslash (immediately following the edit command) is mandatory
325 * as soon as we hit a TEXT line that has no trailing '\', we're done.
326 * this means a command like:
335 if (editstr[1] != '\\' && (editstr[2] != '\n' || editstr[2] != '\r'))
336 fatalError("bad format in edit expression\n");
338 /* store the edit line text */
339 /* make editline big enough to accomodate the extra '\n' we will tack on
341 sed_cmd->editline = xmalloc(strlen(&editstr[3]) + 2);
342 strcpy(sed_cmd->editline, &editstr[3]);
343 ptr = sed_cmd->editline;
345 /* now we need to go through * and: s/\\[\r\n]$/\n/g on the edit line */
347 while (ptr[idx] != '\\' && (ptr[idx+1] != '\n' || ptr[idx+1] != '\r')) {
353 /* move the newline over the '\' before it (effectively eats the '\') */
354 memmove(&ptr[idx], &ptr[idx+1], strlen(&ptr[idx+1]));
355 ptr[strlen(ptr)-1] = 0;
357 /* substitue \r for \n if needed */
358 if (ptr[idx] == '\r')
366 /* this accounts for discrepancies between the modified string and the
367 * original string passed in to this function */
368 idx += slashes_eaten;
370 /* this accounts for the fact that A) we started at index 3, not at index
371 * 0 and B) that we added an extra '\n' at the end (if you think the next
372 * line should read 'idx += 4' remember, arrays are zero-based) */
379 static char *parse_cmd_str(struct sed_cmd *sed_cmd, const char *cmdstr)
384 * format is: [addr][,addr]cmd
390 /* first part (if present) is an address: either a number or a /regex/ */
391 if (isdigit(cmdstr[idx]) || cmdstr[idx] == '/')
392 idx = get_address(cmdstr, &sed_cmd->beg_line, &sed_cmd->beg_match);
394 /* second part (if present) will begin with a comma */
395 if (cmdstr[idx] == ',')
396 idx += get_address(&cmdstr[++idx], &sed_cmd->end_line, &sed_cmd->end_match);
398 /* last part (mandatory) will be a command */
399 if (cmdstr[idx] == '\0')
400 fatalError("missing command\n");
401 if (!strchr("pdsaic", cmdstr[idx])) /* <-- XXX add new commands here */
402 fatalError("invalid command\n");
403 sed_cmd->cmd = cmdstr[idx];
405 /* special-case handling for (s)ubstitution */
406 if (sed_cmd->cmd == 's') {
407 idx += parse_subst_cmd(sed_cmd, &cmdstr[idx]);
409 /* special-case handling for (a)ppend, (i)nsert, and (c)hange */
410 else if (strchr("aic", cmdstr[idx])) {
411 if (sed_cmd->end_line || sed_cmd->end_match)
412 fatalError("only a beginning address can be specified for edit commands\n");
413 idx += parse_edit_cmd(sed_cmd, &cmdstr[idx]);
415 /* if it was a single-letter command (such as 'p' or 'd') we need to
416 * increment the index past that command */
420 /* give back whatever's left over */
421 return (char *)&cmdstr[idx];
424 static void add_cmd_str(const char *cmdstr)
426 char *mystr = (char *)cmdstr;
430 /* trim leading whitespace and semicolons */
431 memmove(mystr, &mystr[strspn(mystr, "; \n\r\t\v")], strlen(mystr));
432 /* if we ate the whole thing, that means there was just trailing
433 * whitespace or a final / no-op semicolon. either way, get out */
434 if (strlen(mystr) == 0)
436 /* if this is a comment, jump past it and keep going */
437 if (mystr[0] == '#') {
438 mystr = strpbrk(mystr, ";\n\r");
442 sed_cmds = realloc(sed_cmds, sizeof(struct sed_cmd) * (++ncmds));
443 /* zero new element */
444 memset(&sed_cmds[ncmds-1], 0, sizeof(struct sed_cmd));
445 /* load command string into new array element, get remainder */
446 mystr = parse_cmd_str(&sed_cmds[ncmds-1], mystr);
448 } while (mystr && strlen(mystr));
452 static void load_cmd_file(char *filename)
458 cmdfile = fopen(filename, "r");
460 fatalError(strerror(errno));
462 while ((line = get_line_from_file(cmdfile)) != NULL) {
463 /* if a line ends with '\' it needs the next line appended to it */
464 while (line[strlen(line)-2] == '\\' &&
465 (nextline = get_line_from_file(cmdfile)) != NULL) {
466 line = realloc(line, strlen(line) + strlen(nextline) + 1);
467 strcat(line, nextline);
470 /* eat trailing newline (if any) --if I don't do this, edit commands
471 * (aic) will print an extra newline */
472 if (line[strlen(line)-1] == '\n')
473 line[strlen(line)-1] = 0;
479 static int do_subst_command(const struct sed_cmd *sed_cmd, const char *line)
483 /* we only substitute if the substitution 'search' expression matches */
484 if (regexec(sed_cmd->sub_match, line, 0, NULL, 0) == 0) {
487 char *ptr = (char *)line;
490 /* if we can match the search string... */
491 if (regexec(sed_cmd->sub_match, ptr, 1, ®match, 0) == 0) {
492 /* print everything before the match, */
493 for (i = 0; i < regmatch.rm_so; i++)
494 fputc(ptr[i], stdout);
495 /* then print the substitution in its place */
496 fputs(sed_cmd->replace, stdout);
497 /* then advance past the match */
498 ptr += regmatch.rm_eo;
499 /* and flag that something has changed */
502 /* if we're not doing this globally... */
506 /* if we COULD NOT match the search string (meaning we've gone past
507 * all previous instances), get out */
512 /* is there anything left to print? */
520 static int do_sed_command(const struct sed_cmd *sed_cmd, const char *line)
524 switch (sed_cmd->cmd) {
535 altered = do_subst_command(sed_cmd, line);
540 fputs(sed_cmd->editline, stdout);
545 fputs(sed_cmd->editline, stdout);
549 fputs(sed_cmd->editline, stdout);
557 static void process_file(FILE *file)
560 static int linenum = 0; /* GNU sed does not restart counting lines at EOF */
561 unsigned int still_in_range = 0;
565 /* go through every line in the file */
566 while ((line = get_line_from_file(file)) != NULL) {
571 /* for every line, go through all the commands */
572 for (i = 0; i < ncmds; i++) {
574 /* are we acting on a range of matched lines? */
575 if (sed_cmds[i].beg_match && sed_cmds[i].end_match) {
576 if (still_in_range || regexec(sed_cmds[i].beg_match, line, 0, NULL, 0) == 0) {
577 line_altered += do_sed_command(&sed_cmds[i], line);
579 if (regexec(sed_cmds[i].end_match, line, 0, NULL, 0) == 0)
584 /* are we trying to match a single line? */
585 else if (sed_cmds[i].beg_match) {
586 if (regexec(sed_cmds[i].beg_match, line, 0, NULL, 0) == 0)
587 line_altered += do_sed_command(&sed_cmds[i], line);
590 /* are we acting on a range of line numbers? */
591 else if (sed_cmds[i].beg_line > 0 && sed_cmds[i].end_line > 0) {
592 if (linenum >= sed_cmds[i].beg_line && linenum <= sed_cmds[i].end_line)
593 line_altered += do_sed_command(&sed_cmds[i], line);
596 /* are we acting on a specified line number */
597 else if (sed_cmds[i].beg_line > 0) {
598 if (linenum == sed_cmds[i].beg_line)
599 line_altered += do_sed_command(&sed_cmds[i], line);
602 /* not acting on matches or line numbers. act on every line */
604 line_altered += do_sed_command(&sed_cmds[i], line);
608 /* we will print the line unless we were told to be quiet or if the
609 * line was altered (via a 'd'elete or 's'ubstitution) */
610 if (!be_quiet && !line_altered)
617 extern int sed_main(int argc, char **argv)
621 /* do special-case option parsing */
622 if (argv[1] && (strcmp(argv[1], "--help") == 0))
626 /* destroy command strings on exit */
627 if (atexit(destroy_cmd_strs) == -1) {
633 /* do normal option parsing */
634 while ((opt = getopt(argc, argv, "Vhne:f:")) > 0) {
637 printf("BusyBox v%s (%s)\n", BB_VER, BB_BT);
650 load_cmd_file(optarg);
655 /* if we didn't get a pattern from a -e and no command file was specified,
656 * argv[optind] should be the pattern. no pattern, no worky */
658 if (argv[optind] == NULL)
661 add_cmd_str(argv[optind]);
667 /* argv[(optind)..(argc-1)] should be names of file to process. If no
668 * files were specified or '-' was specified, take input from stdin.
669 * Otherwise, we process all the files specified. */
670 if (argv[optind] == NULL || (strcmp(argv[optind], "-") == 0)) {
676 for (i = optind; i < argc; i++) {
677 file = fopen(argv[i], "r");
679 errorMsg("%s: %s\n", argv[i], strerror(errno));