xref: /openbsd/bin/dd/conv.c (revision 998de4a5)
1 /*	$OpenBSD: conv.c,v 1.13 2016/08/16 16:44:55 krw 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 #include <sys/types.h>
38 #include <sys/time.h>
39 
40 #include <err.h>
41 #include <string.h>
42 
43 #include "dd.h"
44 #include "extern.h"
45 
46 #define MINIMUM(a, b)	(((a) < (b)) ? (a) : (b))
47 
48 /*
49  * def --
50  * Copy input to output.  Input is buffered until reaches obs, and then
51  * output until less than obs remains.  Only a single buffer is used.
52  * Worst case buffer calculation is (ibs + obs - 1).
53  */
54 void
55 def(void)
56 {
57 	size_t cnt;
58 	u_char *inp;
59 	const u_char *t;
60 
61 	if ((t = ctab) != NULL)
62 		for (inp = in.dbp - (cnt = in.dbrcnt); cnt--; ++inp)
63 			*inp = t[*inp];
64 
65 	/* Make the output buffer look right. */
66 	out.dbp = in.dbp;
67 	out.dbcnt = in.dbcnt;
68 
69 	if (in.dbcnt >= out.dbsz) {
70 		/* If the output buffer is full, write it. */
71 		dd_out(0);
72 
73 		/*
74 		 * Ddout copies the leftover output to the beginning of
75 		 * the buffer and resets the output buffer.  Reset the
76 		 * input buffer to match it.
77 		 */
78 		in.dbp = out.dbp;
79 		in.dbcnt = out.dbcnt;
80 	}
81 }
82 
83 void
84 def_close(void)
85 {
86 	/* Just update the count, everything is already in the buffer. */
87 	if (in.dbcnt)
88 		out.dbcnt = in.dbcnt;
89 }
90 
91 #ifdef	NO_CONV
92 /* Build a smaller version (i.e. for a miniroot) */
93 /* These can not be called, but just in case...  */
94 static char no_block[] = "unblock and -DNO_CONV?";
95 void block()       { errx(1, "%s", no_block + 2); }
96 void block_close() { errx(1, "%s", no_block + 2); }
97 void unblock()       { errx(1, "%s", no_block); }
98 void unblock_close() { errx(1, "%s", no_block); }
99 #else	/* NO_CONV */
100 
101 /*
102  * Copy variable length newline terminated records with a max size cbsz
103  * bytes to output.  Records less than cbs are padded with spaces.
104  *
105  * max in buffer:  MAX(ibs, cbsz)
106  * max out buffer: obs + cbsz
107  */
108 void
109 block(void)
110 {
111 	static int intrunc;
112 	int ch = -1;
113 	size_t cnt, maxlen;
114 	u_char *inp, *outp;
115 	const u_char *t;
116 
117 	/*
118 	 * Record truncation can cross block boundaries.  If currently in a
119 	 * truncation state, keep tossing characters until reach a newline.
120 	 * Start at the beginning of the buffer, as the input buffer is always
121 	 * left empty.
122 	 */
123 	if (intrunc) {
124 		for (inp = in.db, cnt = in.dbrcnt;
125 		    cnt && *inp++ != '\n'; --cnt);
126 		if (!cnt) {
127 			in.dbcnt = 0;
128 			in.dbp = in.db;
129 			return;
130 		}
131 		intrunc = 0;
132 		/* Adjust the input buffer numbers. */
133 		in.dbcnt = cnt - 1;
134 		in.dbp = inp + cnt - 1;
135 	}
136 
137 	/*
138 	 * Copy records (max cbsz size chunks) into the output buffer.  The
139 	 * translation is done as we copy into the output buffer.
140 	 */
141 	for (inp = in.dbp - in.dbcnt, outp = out.dbp; in.dbcnt;) {
142 		maxlen = MINIMUM(cbsz, in.dbcnt);
143 		if ((t = ctab) != NULL)
144 			for (cnt = 0;
145 			    cnt < maxlen && (ch = *inp++) != '\n'; ++cnt)
146 				*outp++ = t[ch];
147 		else
148 			for (cnt = 0;
149 			    cnt < maxlen && (ch = *inp++) != '\n'; ++cnt)
150 				*outp++ = ch;
151 		/*
152 		 * Check for short record without a newline.  Reassemble the
153 		 * input block.
154 		 */
155 		if (ch != '\n' && in.dbcnt < cbsz) {
156 			(void)memmove(in.db, in.dbp - in.dbcnt, in.dbcnt);
157 			break;
158 		}
159 
160 		/* Adjust the input buffer numbers. */
161 		in.dbcnt -= cnt;
162 		if (ch == '\n')
163 			--in.dbcnt;
164 
165 		/* Pad short records with spaces. */
166 		if (cnt < cbsz)
167 			(void)memset(outp, ctab ? ctab[' '] : ' ', cbsz - cnt);
168 		else {
169 			/*
170 			 * If the next character wouldn't have ended the
171 			 * block, it's a truncation.
172 			 */
173 			if (!in.dbcnt || *inp != '\n')
174 				++st.trunc;
175 
176 			/* Toss characters to a newline. */
177 			for (; in.dbcnt && *inp++ != '\n'; --in.dbcnt);
178 			if (!in.dbcnt)
179 				intrunc = 1;
180 			else
181 				--in.dbcnt;
182 		}
183 
184 		/* Adjust output buffer numbers. */
185 		out.dbp += cbsz;
186 		if ((out.dbcnt += cbsz) >= out.dbsz)
187 			dd_out(0);
188 		outp = out.dbp;
189 	}
190 	in.dbp = in.db + in.dbcnt;
191 }
192 
193 void
194 block_close(void)
195 {
196 	/*
197 	 * Copy any remaining data into the output buffer and pad to a record.
198 	 * Don't worry about truncation or translation, the input buffer is
199 	 * always empty when truncating, and no characters have been added for
200 	 * translation.  The bottom line is that anything left in the input
201 	 * buffer is a truncated record.  Anything left in the output buffer
202 	 * just wasn't big enough.
203 	 */
204 	if (in.dbcnt) {
205 		++st.trunc;
206 		(void)memmove(out.dbp, in.dbp - in.dbcnt, in.dbcnt);
207 		(void)memset(out.dbp + in.dbcnt,
208 		    ctab ? ctab[' '] : ' ', cbsz - in.dbcnt);
209 		out.dbcnt += cbsz;
210 	}
211 }
212 
213 /*
214  * Convert fixed length (cbsz) records to variable length.  Deletes any
215  * trailing blanks and appends a newline.
216  *
217  * max in buffer:  MAX(ibs, cbsz) + cbsz
218  * max out buffer: obs + cbsz
219  */
220 void
221 unblock(void)
222 {
223 	size_t cnt;
224 	u_char *inp;
225 	const u_char *t;
226 
227 	/* Translation and case conversion. */
228 	if ((t = ctab) != NULL)
229 		for (cnt = in.dbrcnt, inp = in.dbp - 1; cnt--; inp--)
230 			*inp = t[*inp];
231 	/*
232 	 * Copy records (max cbsz size chunks) into the output buffer.  The
233 	 * translation has to already be done or we might not recognize the
234 	 * spaces.
235 	 */
236 	for (inp = in.db; in.dbcnt >= cbsz; inp += cbsz, in.dbcnt -= cbsz) {
237 		for (t = inp + cbsz - 1; t >= inp && *t == ' '; --t);
238 		if (t >= inp) {
239 			cnt = t - inp + 1;
240 			(void)memmove(out.dbp, inp, cnt);
241 			out.dbp += cnt;
242 			out.dbcnt += cnt;
243 		}
244 		++out.dbcnt;
245 		*out.dbp++ = '\n';
246 		if (out.dbcnt >= out.dbsz)
247 			dd_out(0);
248 	}
249 	if (in.dbcnt)
250 		(void)memmove(in.db, in.dbp - in.dbcnt, in.dbcnt);
251 	in.dbp = in.db + in.dbcnt;
252 }
253 
254 void
255 unblock_close(void)
256 {
257 	size_t cnt;
258 	u_char *t;
259 
260 	if (in.dbcnt) {
261 		warnx("%s: short input record", in.name);
262 		for (t = in.db + in.dbcnt - 1; t >= in.db && *t == ' '; --t);
263 		if (t >= in.db) {
264 			cnt = t - in.db + 1;
265 			(void)memmove(out.dbp, in.db, cnt);
266 			out.dbp += cnt;
267 			out.dbcnt += cnt;
268 		}
269 		++out.dbcnt;
270 		*out.dbp++ = '\n';
271 	}
272 }
273 
274 #endif	/* NO_CONV */
275