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 */
136 * trim_str - trims leading and trailing space from a string
138 * Note: This returns a malloc'ed string so you must store and free it
139 * XXX: This should be in the utility.c file.
141 static char *trim_str(const char *str)
144 char *retstr = strdup(str);
146 /* trim leading whitespace */
147 memmove(retstr, &retstr[strspn(retstr, " \n\t\v")], strlen(retstr));
149 /* trim trailing whitespace */
150 i = strlen(retstr) - 1;
151 while (isspace(retstr[i]))
157 * you know, a strrspn() would really be nice cuz then we could say:
159 * retstr[strlen(retstr) - strrspn(retstr, " \n\t\v") + 1] = 0;
166 * index_of_unescaped_slash - walks left to right through a string beginning
167 * at a specified index and returns the index of the next unescaped slash.
169 static int index_of_next_unescaped_slash(const char *str, int idx)
173 /* test if we've hit the end */
176 } while (str[idx] != '/' && str[idx - 1] != '\\');
182 * returns the index in the string just past where the address ends.
184 static int get_address(const char *str, int *line, regex_t **regex)
186 char *my_str = strdup(str);
189 if (isdigit(my_str[idx])) {
192 } while (isdigit(my_str[idx]));
194 *line = atoi(my_str);
196 else if (my_str[idx] == '$') {
200 else if (my_str[idx] == '/') {
201 idx = index_of_next_unescaped_slash(my_str, idx);
203 fatalError("unterminated match expression\n");
205 *regex = (regex_t *)xmalloc(sizeof(regex_t));
206 xregcomp(*regex, my_str+1, REG_NEWLINE);
207 idx++; /* so it points to the next character after the last '/' */
210 errorMsg("get_address: no address found in string\n"
211 "\t(you probably didn't check the string you passed me)\n");
219 static char *strdup_substr(const char *str, int start, int end)
221 int size = end - start + 1;
222 char *newstr = xmalloc(size);
223 memcpy(newstr, str+start, size-1);
224 newstr[size-1] = '\0';
228 static void parse_subst_cmd(struct sed_cmd *sed_cmd, const char *substr)
230 int oldidx, cflags = REG_NEWLINE;
235 * the string that gets passed to this function should look like this:
240 * (all three of the '/' slashes are mandatory)
243 /* verify that the 's' is followed by a 'slash' */
244 if (substr[++idx] != '/')
245 fatalError("bad format in substitution expression\n");
247 /* save the match string */
249 idx = index_of_next_unescaped_slash(substr, idx);
251 fatalError("bad format in substitution expression\n");
252 match = strdup_substr(substr, oldidx, idx);
254 /* save the replacement string */
256 idx = index_of_next_unescaped_slash(substr, idx);
258 fatalError("bad format in substitution expression\n");
259 sed_cmd->replace = strdup_substr(substr, oldidx, idx);
261 /* process the flags */
262 while (substr[++idx]) {
263 switch (substr[idx]) {
271 fatalError("bad option in substitution expression\n");
275 /* compile the regex */
276 sed_cmd->sub_match = (regex_t *)xmalloc(sizeof(regex_t));
277 xregcomp(sed_cmd->sub_match, match, cflags);
281 static void parse_edit_cmd(struct sed_cmd *sed_cmd, const char *editstr)
284 char *ptr; /* shorthand */
287 * the string that gets passed to this function should look like this:
291 * | this backslash (immediately following the edit command) is mandatory
298 * as soon as we hit a TEXT line that has no trailing '\', we're done.
299 * this means a command like:
308 if (editstr[1] != '\\' && (editstr[2] != '\n' || editstr[2] != '\r'))
309 fatalError("bad format in edit expression\n");
311 /* store the edit line text */
312 /* make editline big enough to accomodate the extra '\n' we will tack on
314 sed_cmd->editline = xmalloc(strlen(&editstr[3]) + 2);
315 strcpy(sed_cmd->editline, &editstr[3]);
316 ptr = sed_cmd->editline;
318 /* now we need to go through * and: s/\\[\r\n]$/\n/g on the edit line */
320 while (ptr[idx] != '\\' && (ptr[idx+1] != '\n' || ptr[idx+1] != '\r')) {
328 /* move the newline over the '\' before it (effectively eats the '\') */
329 memmove(&ptr[idx], &ptr[idx+1], strlen(&ptr[idx+1]));
330 ptr[strlen(ptr)-1] = 0;
331 /* substitue \r for \n if needed */
332 if (ptr[idx] == '\r')
337 static void parse_cmd_str(struct sed_cmd *sed_cmd, const char *cmdstr)
342 * format is: [addr][,addr]cmd
347 /* first part (if present) is an address: either a number or a /regex/ */
348 if (isdigit(cmdstr[idx]) || cmdstr[idx] == '/')
349 idx = get_address(cmdstr, &sed_cmd->beg_line, &sed_cmd->beg_match);
351 /* second part (if present) will begin with a comma */
352 if (cmdstr[idx] == ',')
353 idx += get_address(&cmdstr[++idx], &sed_cmd->end_line, &sed_cmd->end_match);
355 /* last part (mandatory) will be a command */
356 if (cmdstr[idx] == '\0')
357 fatalError("missing command\n");
358 if (!strchr("pdsaic", cmdstr[idx])) /* <-- XXX add new commands here */
359 fatalError("invalid command\n");
360 sed_cmd->cmd = cmdstr[idx];
362 /* special-case handling for (s)ubstitution */
363 if (sed_cmd->cmd == 's')
364 parse_subst_cmd(sed_cmd, &cmdstr[idx]);
366 /* special-case handling for (a)ppend, (i)nsert, and (c)hange */
367 if (strchr("aic", cmdstr[idx])) {
368 if (sed_cmd->end_line || sed_cmd->end_match)
369 fatalError("only a beginning address can be specified for edit commands\n");
370 parse_edit_cmd(sed_cmd, &cmdstr[idx]);
374 static void add_cmd_str(const char *cmdstr)
376 char *my_cmdstr = trim_str(cmdstr);
378 /* if this is a comment, don't even bother */
379 if (my_cmdstr[0] == '#') {
385 sed_cmds = realloc(sed_cmds, sizeof(struct sed_cmd) * (++ncmds));
386 /* zero new element */
387 memset(&sed_cmds[ncmds-1], 0, sizeof(struct sed_cmd));
388 /* load command string into new array element */
389 parse_cmd_str(&sed_cmds[ncmds-1], my_cmdstr);
393 static void load_cmd_file(char *filename)
399 cmdfile = fopen(filename, "r");
401 fatalError(strerror(errno));
403 while ((line = get_line_from_file(cmdfile)) != NULL) {
404 /* if a line ends with '\' it needs the next line appended to it */
405 while (line[strlen(line)-2] == '\\' &&
406 (nextline = get_line_from_file(cmdfile)) != NULL) {
407 line = realloc(line, strlen(line) + strlen(nextline) + 1);
408 strcat(line, nextline);
415 static int do_subst_command(const struct sed_cmd *sed_cmd, const char *line)
419 /* we only substitute if the substitution 'search' expression matches */
420 if (regexec(sed_cmd->sub_match, line, 0, NULL, 0) == 0) {
423 char *ptr = (char *)line;
426 /* if we can match the search string... */
427 if (regexec(sed_cmd->sub_match, ptr, 1, ®match, 0) == 0) {
428 /* print everything before the match, */
429 for (i = 0; i < regmatch.rm_so; i++)
430 fputc(ptr[i], stdout);
431 /* then print the substitution in its place */
432 fputs(sed_cmd->replace, stdout);
433 /* then advance past the match */
434 ptr += regmatch.rm_eo;
435 /* and flag that something has changed */
438 /* if we're not doing this globally... */
442 /* if we COULD NOT match the search string (meaning we've gone past
443 * all previous instances), get out */
448 /* is there anything left to print? */
456 static int do_sed_command(const struct sed_cmd *sed_cmd, const char *line)
460 switch (sed_cmd->cmd) {
471 altered = do_subst_command(sed_cmd, line);
476 fputs(sed_cmd->editline, stdout);
481 fputs(sed_cmd->editline, stdout);
485 fputs(sed_cmd->editline, stdout);
493 static void process_file(FILE *file)
496 static int linenum = 0; /* GNU sed does not restart counting lines at EOF */
497 unsigned int still_in_range = 0;
501 /* go through every line in the file */
502 while ((line = get_line_from_file(file)) != NULL) {
507 /* for every line, go through all the commands */
508 for (i = 0; i < ncmds; i++) {
510 /* are we acting on a range of matched lines? */
511 if (sed_cmds[i].beg_match && sed_cmds[i].end_match) {
512 if (still_in_range || regexec(sed_cmds[i].beg_match, line, 0, NULL, 0) == 0) {
513 line_altered += do_sed_command(&sed_cmds[i], line);
515 if (regexec(sed_cmds[i].end_match, line, 0, NULL, 0) == 0)
520 /* are we trying to match a single line? */
521 else if (sed_cmds[i].beg_match) {
522 if (regexec(sed_cmds[i].beg_match, line, 0, NULL, 0) == 0)
523 line_altered += do_sed_command(&sed_cmds[i], line);
526 /* are we acting on a range of line numbers? */
527 else if (sed_cmds[i].beg_line > 0 && sed_cmds[i].end_line > 0) {
528 if (linenum >= sed_cmds[i].beg_line && linenum <= sed_cmds[i].end_line)
529 line_altered += do_sed_command(&sed_cmds[i], line);
532 /* are we acting on a specified line number */
533 else if (sed_cmds[i].beg_line > 0) {
534 if (linenum == sed_cmds[i].beg_line)
535 line_altered += do_sed_command(&sed_cmds[i], line);
538 /* not acting on matches or line numbers. act on every line */
540 line_altered += do_sed_command(&sed_cmds[i], line);
544 /* we will print the line unless we were told to be quiet or if the
545 * line was altered (via a 'd'elete or 's'ubstitution) */
546 if (!be_quiet && !line_altered)
553 extern int sed_main(int argc, char **argv)
557 /* do special-case option parsing */
558 if (argv[1] && (strcmp(argv[1], "--help") == 0))
562 /* destroy command strings on exit */
563 if (atexit(destroy_cmd_strs) == -1) {
569 /* do normal option parsing */
570 while ((opt = getopt(argc, argv, "Vhne:f:")) > 0) {
573 printf("BusyBox v%s (%s)\n", BB_VER, BB_BT);
586 load_cmd_file(optarg);
591 /* if we didn't get a pattern from a -e and no command file was specified,
592 * argv[optind] should be the pattern. no pattern, no worky */
594 if (argv[optind] == NULL)
597 add_cmd_str(argv[optind]);
603 /* argv[(optind)..(argc-1)] should be names of file to process. If no
604 * files were specified or '-' was specified, take input from stdin.
605 * Otherwise, we process all the files specified. */
606 if (argv[optind] == NULL || (strcmp(argv[optind], "-") == 0)) {
612 for (i = optind; i < argc; i++) {
613 file = fopen(argv[i], "r");
615 errorMsg("%s: %s\n", argv[i], strerror(errno));