xref: /openbsd/bin/dd/conv.c (revision 29295d1c)
1 /*	$OpenBSD: conv.c,v 1.9 2003/06/02 23:32:07 millert Exp $	*/
2 /*	$NetBSD: conv.c,v 1.6 1996/02/20 19:29:02 jtc Exp $	*/
3 
4 /*-
5  * Copyright (c) 1991, 1993, 1994
6  *	The Regents of the University of California.  All rights reserved.
7  *
8  * This code is derived from software contributed to Berkeley by
9  * Keith Muller of the University of California, San Diego and Lance
10  * Visser of Convex Computer Corporation.
11  *
12  * Redistribution and use in source and binary forms, with or without
13  * modification, are permitted provided that the following conditions
14  * are met:
15  * 1. Redistributions of source code must retain the above copyright
16  *    notice, this list of conditions and the following disclaimer.
17  * 2. Redistributions in binary form must reproduce the above copyright
18  *    notice, this list of conditions and the following disclaimer in the
19  *    documentation and/or other materials provided with the distribution.
20  * 3. Neither the name of the University nor the names of its contributors
21  *    may be used to endorse or promote products derived from this software
22  *    without specific prior written permission.
23  *
24  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34  * SUCH DAMAGE.
35  */
36 
37 #ifndef lint
38 #if 0
39 static char sccsid[] = "@(#)conv.c	8.3 (Berkeley) 4/2/94";
40 #else
41 static char rcsid[] = "$OpenBSD: conv.c,v 1.9 2003/06/02 23:32:07 millert Exp $";
42 #endif
43 #endif /* not lint */
44 
45 #include <sys/param.h>
46 #include <sys/time.h>
47 
48 #include <err.h>
49 #include <string.h>
50 
51 #include "dd.h"
52 #include "extern.h"
53 
54 /*
55  * def --
56  * Copy input to output.  Input is buffered until reaches obs, and then
57  * output until less than obs remains.  Only a single buffer is used.
58  * Worst case buffer calculation is (ibs + obs - 1).
59  */
60 void
61 def()
62 {
63 	size_t cnt;
64 	u_char *inp;
65 	const u_char *t;
66 
67 	if ((t = ctab) != NULL)
68 		for (inp = in.dbp - (cnt = in.dbrcnt); cnt--; ++inp)
69 			*inp = t[*inp];
70 
71 	/* Make the output buffer look right. */
72 	out.dbp = in.dbp;
73 	out.dbcnt = in.dbcnt;
74 
75 	if (in.dbcnt >= out.dbsz) {
76 		/* If the output buffer is full, write it. */
77 		dd_out(0);
78 
79 		/*
80 		 * Ddout copies the leftover output to the beginning of
81 		 * the buffer and resets the output buffer.  Reset the
82 		 * input buffer to match it.
83 	 	 */
84 		in.dbp = out.dbp;
85 		in.dbcnt = out.dbcnt;
86 	}
87 }
88 
89 void
90 def_close()
91 {
92 	/* Just update the count, everything is already in the buffer. */
93 	if (in.dbcnt)
94 		out.dbcnt = in.dbcnt;
95 }
96 
97 #ifdef	NO_CONV
98 /* Build a smaller version (i.e. for a miniroot) */
99 /* These can not be called, but just in case...  */
100 static char no_block[] = "unblock and -DNO_CONV?";
101 void block()       { errx(1, "%s", no_block + 2); }
102 void block_close() { errx(1, "%s", no_block + 2); }
103 void unblock()       { errx(1, "%s", no_block); }
104 void unblock_close() { errx(1, "%s", no_block); }
105 #else	/* NO_CONV */
106 
107 /*
108  * Copy variable length newline terminated records with a max size cbsz
109  * bytes to output.  Records less than cbs are padded with spaces.
110  *
111  * max in buffer:  MAX(ibs, cbsz)
112  * max out buffer: obs + cbsz
113  */
114 void
115 block()
116 {
117 	static int intrunc;
118 	int ch = -1;
119 	size_t cnt, maxlen;
120 	u_char *inp, *outp;
121 	const u_char *t;
122 
123 	/*
124 	 * Record truncation can cross block boundaries.  If currently in a
125 	 * truncation state, keep tossing characters until reach a newline.
126 	 * Start at the beginning of the buffer, as the input buffer is always
127 	 * left empty.
128 	 */
129 	if (intrunc) {
130 		for (inp = in.db, cnt = in.dbrcnt;
131 		    cnt && *inp++ != '\n'; --cnt);
132 		if (!cnt) {
133 			in.dbcnt = 0;
134 			in.dbp = in.db;
135 			return;
136 		}
137 		intrunc = 0;
138 		/* Adjust the input buffer numbers. */
139 		in.dbcnt = cnt - 1;
140 		in.dbp = inp + cnt - 1;
141 	}
142 
143 	/*
144 	 * Copy records (max cbsz size chunks) into the output buffer.  The
145 	 * translation is done as we copy into the output buffer.
146 	 */
147 	for (inp = in.dbp - in.dbcnt, outp = out.dbp; in.dbcnt;) {
148 		maxlen = MIN(cbsz, in.dbcnt);
149 		if ((t = ctab) != NULL)
150 			for (cnt = 0;
151 			    cnt < maxlen && (ch = *inp++) != '\n'; ++cnt)
152 				*outp++ = t[ch];
153 		else
154 			for (cnt = 0;
155 			    cnt < maxlen && (ch = *inp++) != '\n'; ++cnt)
156 				*outp++ = ch;
157 		/*
158 		 * Check for short record without a newline.  Reassemble the
159 		 * input block.
160 		 */
161 		if (ch != '\n' && in.dbcnt < cbsz) {
162 			(void)memmove(in.db, in.dbp - in.dbcnt, in.dbcnt);
163 			break;
164 		}
165 
166 		/* Adjust the input buffer numbers. */
167 		in.dbcnt -= cnt;
168 		if (ch == '\n')
169 			--in.dbcnt;
170 
171 		/* Pad short records with spaces. */
172 		if (cnt < cbsz)
173 			(void)memset(outp, ctab ? ctab[' '] : ' ', cbsz - cnt);
174 		else {
175 			/*
176 			 * If the next character wouldn't have ended the
177 			 * block, it's a truncation.
178 			 */
179 			if (!in.dbcnt || *inp != '\n')
180 				++st.trunc;
181 
182 			/* Toss characters to a newline. */
183 			for (; in.dbcnt && *inp++ != '\n'; --in.dbcnt);
184 			if (!in.dbcnt)
185 				intrunc = 1;
186 			else
187 				--in.dbcnt;
188 		}
189 
190 		/* Adjust output buffer numbers. */
191 		out.dbp += cbsz;
192 		if ((out.dbcnt += cbsz) >= out.dbsz)
193 			dd_out(0);
194 		outp = out.dbp;
195 	}
196 	in.dbp = in.db + in.dbcnt;
197 }
198 
199 void
200 block_close()
201 {
202 	/*
203 	 * Copy any remaining data into the output buffer and pad to a record.
204 	 * Don't worry about truncation or translation, the input buffer is
205 	 * always empty when truncating, and no characters have been added for
206 	 * translation.  The bottom line is that anything left in the input
207 	 * buffer is a truncated record.  Anything left in the output buffer
208 	 * just wasn't big enough.
209 	 */
210 	if (in.dbcnt) {
211 		++st.trunc;
212 		(void)memmove(out.dbp, in.dbp - in.dbcnt, in.dbcnt);
213 		(void)memset(out.dbp + in.dbcnt,
214 		    ctab ? ctab[' '] : ' ', cbsz - in.dbcnt);
215 		out.dbcnt += cbsz;
216 	}
217 }
218 
219 /*
220  * Convert fixed length (cbsz) records to variable length.  Deletes any
221  * trailing blanks and appends a newline.
222  *
223  * max in buffer:  MAX(ibs, cbsz) + cbsz
224  * max out buffer: obs + cbsz
225  */
226 void
227 unblock()
228 {
229 	size_t cnt;
230 	u_char *inp;
231 	const u_char *t;
232 
233 	/* Translation and case conversion. */
234 	if ((t = ctab) != NULL)
235 		for (cnt = in.dbrcnt, inp = in.dbp - 1; cnt--; inp--)
236 			*inp = t[*inp];
237 	/*
238 	 * Copy records (max cbsz size chunks) into the output buffer.  The
239 	 * translation has to already be done or we might not recognize the
240 	 * spaces.
241 	 */
242 	for (inp = in.db; in.dbcnt >= cbsz; inp += cbsz, in.dbcnt -= cbsz) {
243 		for (t = inp + cbsz - 1; t >= inp && *t == ' '; --t);
244 		if (t >= inp) {
245 			cnt = t - inp + 1;
246 			(void)memmove(out.dbp, inp, cnt);
247 			out.dbp += cnt;
248 			out.dbcnt += cnt;
249 		}
250 		++out.dbcnt;
251 		*out.dbp++ = '\n';
252 		if (out.dbcnt >= out.dbsz)
253 			dd_out(0);
254 	}
255 	if (in.dbcnt)
256 		(void)memmove(in.db, in.dbp - in.dbcnt, in.dbcnt);
257 	in.dbp = in.db + in.dbcnt;
258 }
259 
260 void
261 unblock_close()
262 {
263 	size_t cnt;
264 	u_char *t;
265 
266 	if (in.dbcnt) {
267 		warnx("%s: short input record", in.name);
268 		for (t = in.db + in.dbcnt - 1; t >= in.db && *t == ' '; --t);
269 		if (t >= in.db) {
270 			cnt = t - in.db + 1;
271 			(void)memmove(out.dbp, in.db, cnt);
272 			out.dbp += cnt;
273 			out.dbcnt += cnt;
274 		}
275 		++out.dbcnt;
276 		*out.dbp++ = '\n';
277 	}
278 }
279 
280 #endif	/* NO_CONV */
281