xref: /freebsd/bin/ed/sub.c (revision a8445737e740901f5f2c8d24c12ef7fc8b00134e)
1 /* sub.c: This file contains the substitution routines for the ed
2    line editor */
3 /*-
4  * Copyright (c) 1993 Andrew Moore, Talke Studio.
5  * All rights reserved.
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions
9  * are met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in the
14  *    documentation and/or other materials provided with the distribution.
15  *
16  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26  * SUCH DAMAGE.
27  */
28 
29 #ifndef lint
30 #if 0
31 static char * const rcsid = "@(#)sub.c,v 1.1 1994/02/01 00:34:44 alm Exp";
32 #else
33 static char * const rcsid =
34 	"$Id: sub.c,v 1.10 1997/08/07 21:33:44 steve Exp $";
35 #endif
36 #endif /* not lint */
37 
38 #include "ed.h"
39 
40 
41 char *rhbuf;			/* rhs substitution buffer */
42 int rhbufsz;			/* rhs substitution buffer size */
43 int rhbufi;			/* rhs substitution buffer index */
44 
45 /* extract_subst_tail: extract substitution tail from the command buffer */
46 int
47 extract_subst_tail(flagp, np)
48 	int *flagp;
49 	long *np;
50 {
51 	char delimiter;
52 
53 	*flagp = *np = 0;
54 	if ((delimiter = *ibufp) == '\n') {
55 		rhbufi = 0;
56 		*flagp = GPR;
57 		return 0;
58 	} else if (extract_subst_template() == NULL)
59 		return  ERR;
60 	else if (*ibufp == '\n') {
61 		*flagp = GPR;
62 		return 0;
63 	} else if (*ibufp == delimiter)
64 		ibufp++;
65 	if ('1' <= *ibufp && *ibufp <= '9') {
66 		STRTOL(*np, ibufp);
67 		return 0;
68 	} else if (*ibufp == 'g') {
69 		ibufp++;
70 		*flagp = GSG;
71 		return 0;
72 	}
73 	return 0;
74 }
75 
76 
77 /* extract_subst_template: return pointer to copy of substitution template
78    in the command buffer */
79 char *
80 extract_subst_template()
81 {
82 	int n = 0;
83 	int i = 0;
84 	char c;
85 	char delimiter = *ibufp++;
86 
87 	if (*ibufp == '%' && *(ibufp + 1) == delimiter) {
88 		ibufp++;
89 		if (!rhbuf) sprintf(errmsg, "no previous substitution");
90 		return rhbuf;
91 	}
92 	while (*ibufp != delimiter) {
93 		REALLOC(rhbuf, rhbufsz, i + 2, NULL);
94 		if ((c = rhbuf[i++] = *ibufp++) == '\n' && *ibufp == '\0') {
95 			i--, ibufp--;
96 			break;
97 		} else if (c != '\\')
98 			;
99 		else if ((rhbuf[i++] = *ibufp++) != '\n')
100 			;
101 		else if (!isglobal) {
102 			while ((n = get_tty_line()) == 0 ||
103 			    (n > 0 && ibuf[n - 1] != '\n'))
104 				clearerr(stdin);
105 			if (n < 0)
106 				return NULL;
107 		}
108 	}
109 	REALLOC(rhbuf, rhbufsz, i + 1, NULL);
110 	rhbuf[rhbufi = i] = '\0';
111 	return  rhbuf;
112 }
113 
114 
115 char *rbuf;			/* substitute_matching_text buffer */
116 int rbufsz;			/* substitute_matching_text buffer size */
117 
118 /* search_and_replace: for each line in a range, change text matching a pattern
119    according to a substitution template; return status  */
120 int
121 search_and_replace(pat, gflag, kth)
122 	pattern_t *pat;
123 	int gflag;
124 	int kth;
125 {
126 	undo_t *up;
127 	char *txt;
128 	char *eot;
129 	long lc;
130 	long xa = current_addr;
131 	int nsubs = 0;
132 	line_t *lp;
133 	int len;
134 
135 	current_addr = first_addr - 1;
136 	for (lc = 0; lc <= second_addr - first_addr; lc++) {
137 		lp = get_addressed_line_node(++current_addr);
138 		if ((len = substitute_matching_text(pat, lp, gflag, kth)) < 0)
139 			return ERR;
140 		else if (len) {
141 			up = NULL;
142 			if (delete_lines(current_addr, current_addr) < 0)
143 				return ERR;
144 			txt = rbuf;
145 			eot = rbuf + len;
146 			SPL1();
147 			do {
148 				if ((txt = put_sbuf_line(txt)) == NULL) {
149 					SPL0();
150 					return ERR;
151 				} else if (up)
152 					up->t = get_addressed_line_node(current_addr);
153 				else if ((up = push_undo_stack(UADD,
154 				    current_addr, current_addr)) == NULL) {
155 					SPL0();
156 					return ERR;
157 				}
158 			} while (txt != eot);
159 			SPL0();
160 			nsubs++;
161 			xa = current_addr;
162 		}
163 	}
164 	current_addr = xa;
165 	if  (nsubs == 0 && !(gflag & GLB)) {
166 		sprintf(errmsg, "no match");
167 		return ERR;
168 	} else if ((gflag & (GPR | GLS | GNP)) &&
169 	    display_lines(current_addr, current_addr, gflag) < 0)
170 		return ERR;
171 	return 0;
172 }
173 
174 
175 /* substitute_matching_text: replace text matched by a pattern according to
176    a substitution template; return pointer to the modified text */
177 int
178 substitute_matching_text(pat, lp, gflag, kth)
179 	pattern_t *pat;
180 	line_t *lp;
181 	int gflag;
182 	int kth;
183 {
184 	int off = 0;
185 	int changed = 0;
186 	int matchno = 0;
187 	int i = 0;
188 	regmatch_t rm[SE_MAX];
189 	char *txt;
190 	char *eot;
191 
192 	if ((txt = get_sbuf_line(lp)) == NULL)
193 		return ERR;
194 	if (isbinary)
195 		NUL_TO_NEWLINE(txt, lp->len);
196 	eot = txt + lp->len;
197 	if (!regexec(pat, txt, SE_MAX, rm, 0)) {
198 		do {
199 			if (!kth || kth == ++matchno) {
200 				changed++;
201 				i = rm[0].rm_so;
202 				REALLOC(rbuf, rbufsz, off + i, ERR);
203 				if (isbinary)
204 					NEWLINE_TO_NUL(txt, rm[0].rm_eo);
205 				memcpy(rbuf + off, txt, i);
206 				off += i;
207 				if ((off = apply_subst_template(txt, rm, off,
208 				    pat->re_nsub)) < 0)
209 					return ERR;
210 			} else {
211 				i = rm[0].rm_eo;
212 				REALLOC(rbuf, rbufsz, off + i, ERR);
213 				if (isbinary)
214 					NEWLINE_TO_NUL(txt, i);
215 				memcpy(rbuf + off, txt, i);
216 				off += i;
217 			}
218 			txt += rm[0].rm_eo;
219 		} while (*txt &&
220                         (!changed || ((gflag & GSG) && rm[0].rm_eo)) &&
221 		        !regexec(pat, txt, SE_MAX, rm, REG_NOTBOL));
222 		i = eot - txt;
223 		REALLOC(rbuf, rbufsz, off + i + 2, ERR);
224 		if (i > 0 && !rm[0].rm_eo && (gflag & GSG)) {
225 			sprintf(errmsg, "infinite substitution loop");
226 			return  ERR;
227 		}
228 		if (isbinary)
229 			NEWLINE_TO_NUL(txt, i);
230 		memcpy(rbuf + off, txt, i);
231 		memcpy(rbuf + off + i, "\n", 2);
232 	}
233 	return changed ? off + i + 1 : 0;
234 }
235 
236 
237 /* apply_subst_template: modify text according to a substitution template;
238    return offset to end of modified text */
239 int
240 apply_subst_template(boln, rm, off, re_nsub)
241 	char *boln;
242 	regmatch_t *rm;
243 	int off;
244 	int re_nsub;
245 {
246 	int j = 0;
247 	int k = 0;
248 	int n;
249 	char *sub = rhbuf;
250 
251 	for (; sub - rhbuf < rhbufi; sub++)
252 		if (*sub == '&') {
253 			j = rm[0].rm_so;
254 			k = rm[0].rm_eo;
255 			REALLOC(rbuf, rbufsz, off + k - j, ERR);
256 			while (j < k)
257 				rbuf[off++] = boln[j++];
258 		} else if (*sub == '\\' && '1' <= *++sub && *sub <= '9' &&
259 		    (n = *sub - '0') <= re_nsub) {
260 			j = rm[n].rm_so;
261 			k = rm[n].rm_eo;
262 			REALLOC(rbuf, rbufsz, off + k - j, ERR);
263 			while (j < k)
264 				rbuf[off++] = boln[j++];
265 		} else {
266 			REALLOC(rbuf, rbufsz, off + 1, ERR);
267 			rbuf[off++] = *sub;
268 		}
269 	REALLOC(rbuf, rbufsz, off + 1, ERR);
270 	rbuf[off] = '\0';
271 	return off;
272 }
273