xref: /openbsd/bin/dd/conv.c (revision f2dfb0a4)
1 /*	$OpenBSD: conv.c,v 1.5 1997/02/14 07:05:19 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. All advertising materials mentioning features or use of this software
21  *    must display the following acknowledgement:
22  *	This product includes software developed by the University of
23  *	California, Berkeley and its contributors.
24  * 4. Neither the name of the University nor the names of its contributors
25  *    may be used to endorse or promote products derived from this software
26  *    without specific prior written permission.
27  *
28  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
29  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
30  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
31  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
32  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
33  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
34  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
35  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
36  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
37  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
38  * SUCH DAMAGE.
39  */
40 
41 #ifndef lint
42 #if 0
43 static char sccsid[] = "@(#)conv.c	8.3 (Berkeley) 4/2/94";
44 #else
45 static char rcsid[] = "$OpenBSD: conv.c,v 1.5 1997/02/14 07:05:19 millert Exp $";
46 #endif
47 #endif /* not lint */
48 
49 #include <sys/param.h>
50 
51 #include <err.h>
52 #include <string.h>
53 
54 #include "dd.h"
55 #include "extern.h"
56 
57 /*
58  * def --
59  * Copy input to output.  Input is buffered until reaches obs, and then
60  * output until less than obs remains.  Only a single buffer is used.
61  * Worst case buffer calculation is (ibs + obs - 1).
62  */
63 void
64 def()
65 {
66 	size_t cnt;
67 	u_char *inp;
68 	const u_char *t;
69 
70 	if ((t = ctab) != NULL)
71 		for (inp = in.dbp - (cnt = in.dbrcnt); cnt--; ++inp)
72 			*inp = t[*inp];
73 
74 	/* Make the output buffer look right. */
75 	out.dbp = in.dbp;
76 	out.dbcnt = in.dbcnt;
77 
78 	if (in.dbcnt >= out.dbsz) {
79 		/* If the output buffer is full, write it. */
80 		dd_out(0);
81 
82 		/*
83 		 * Ddout copies the leftover output to the beginning of
84 		 * the buffer and resets the output buffer.  Reset the
85 		 * input buffer to match it.
86 	 	 */
87 		in.dbp = out.dbp;
88 		in.dbcnt = out.dbcnt;
89 	}
90 }
91 
92 void
93 def_close()
94 {
95 	/* Just update the count, everything is already in the buffer. */
96 	if (in.dbcnt)
97 		out.dbcnt = in.dbcnt;
98 }
99 
100 #ifdef	NO_CONV
101 /* Build a smaller version (i.e. for a miniroot) */
102 /* These can not be called, but just in case...  */
103 static char no_block[] = "unblock and -DNO_CONV?";
104 void block()       { errx(1, no_block + 2); }
105 void block_close() { errx(1, no_block + 2); }
106 void unblock()       { errx(1, no_block); }
107 void unblock_close() { errx(1, no_block); }
108 #else	/* NO_CONV */
109 
110 /*
111  * Copy variable length newline terminated records with a max size cbsz
112  * bytes to output.  Records less than cbs are padded with spaces.
113  *
114  * max in buffer:  MAX(ibs, cbsz)
115  * max out buffer: obs + cbsz
116  */
117 void
118 block()
119 {
120 	static int intrunc;
121 	int ch = -1;
122 	size_t cnt, maxlen;
123 	u_char *inp, *outp;
124 	const u_char *t;
125 
126 	/*
127 	 * Record truncation can cross block boundaries.  If currently in a
128 	 * truncation state, keep tossing characters until reach a newline.
129 	 * Start at the beginning of the buffer, as the input buffer is always
130 	 * left empty.
131 	 */
132 	if (intrunc) {
133 		for (inp = in.db, cnt = in.dbrcnt;
134 		    cnt && *inp++ != '\n'; --cnt);
135 		if (!cnt) {
136 			in.dbcnt = 0;
137 			in.dbp = in.db;
138 			return;
139 		}
140 		intrunc = 0;
141 		/* Adjust the input buffer numbers. */
142 		in.dbcnt = cnt - 1;
143 		in.dbp = inp + cnt - 1;
144 	}
145 
146 	/*
147 	 * Copy records (max cbsz size chunks) into the output buffer.  The
148 	 * translation is done as we copy into the output buffer.
149 	 */
150 	for (inp = in.dbp - in.dbcnt, outp = out.dbp; in.dbcnt;) {
151 		maxlen = MIN(cbsz, in.dbcnt);
152 		if ((t = ctab) != NULL)
153 			for (cnt = 0;
154 			    cnt < maxlen && (ch = *inp++) != '\n'; ++cnt)
155 				*outp++ = t[ch];
156 		else
157 			for (cnt = 0;
158 			    cnt < maxlen && (ch = *inp++) != '\n'; ++cnt)
159 				*outp++ = ch;
160 		/*
161 		 * Check for short record without a newline.  Reassemble the
162 		 * input block.
163 		 */
164 		if (ch != '\n' && in.dbcnt < cbsz) {
165 			(void)memmove(in.db, in.dbp - in.dbcnt, in.dbcnt);
166 			break;
167 		}
168 
169 		/* Adjust the input buffer numbers. */
170 		in.dbcnt -= cnt;
171 		if (ch == '\n')
172 			--in.dbcnt;
173 
174 		/* Pad short records with spaces. */
175 		if (cnt < cbsz)
176 			(void)memset(outp, ctab ? ctab[' '] : ' ', cbsz - cnt);
177 		else {
178 			/*
179 			 * If the next character wouldn't have ended the
180 			 * block, it's a truncation.
181 			 */
182 			if (!in.dbcnt || *inp != '\n')
183 				++st.trunc;
184 
185 			/* Toss characters to a newline. */
186 			for (; in.dbcnt && *inp++ != '\n'; --in.dbcnt);
187 			if (!in.dbcnt)
188 				intrunc = 1;
189 			else
190 				--in.dbcnt;
191 		}
192 
193 		/* Adjust output buffer numbers. */
194 		out.dbp += cbsz;
195 		if ((out.dbcnt += cbsz) >= out.dbsz)
196 			dd_out(0);
197 		outp = out.dbp;
198 	}
199 	in.dbp = in.db + in.dbcnt;
200 }
201 
202 void
203 block_close()
204 {
205 	/*
206 	 * Copy any remaining data into the output buffer and pad to a record.
207 	 * Don't worry about truncation or translation, the input buffer is
208 	 * always empty when truncating, and no characters have been added for
209 	 * translation.  The bottom line is that anything left in the input
210 	 * buffer is a truncated record.  Anything left in the output buffer
211 	 * just wasn't big enough.
212 	 */
213 	if (in.dbcnt) {
214 		++st.trunc;
215 		(void)memmove(out.dbp, in.dbp - in.dbcnt, in.dbcnt);
216 		(void)memset(out.dbp + in.dbcnt,
217 		    ctab ? ctab[' '] : ' ', cbsz - in.dbcnt);
218 		out.dbcnt += cbsz;
219 	}
220 }
221 
222 /*
223  * Convert fixed length (cbsz) records to variable length.  Deletes any
224  * trailing blanks and appends a newline.
225  *
226  * max in buffer:  MAX(ibs, cbsz) + cbsz
227  * max out buffer: obs + cbsz
228  */
229 void
230 unblock()
231 {
232 	size_t cnt;
233 	u_char *inp;
234 	const u_char *t;
235 
236 	/* Translation and case conversion. */
237 	if ((t = ctab) != NULL)
238 		for (cnt = in.dbrcnt, inp = in.dbp; cnt--;)
239 			*--inp = t[*inp];
240 	/*
241 	 * Copy records (max cbsz size chunks) into the output buffer.  The
242 	 * translation has to already be done or we might not recognize the
243 	 * spaces.
244 	 */
245 	for (inp = in.db; in.dbcnt >= cbsz; inp += cbsz, in.dbcnt -= cbsz) {
246 		for (t = inp + cbsz - 1; t >= inp && *t == ' '; --t);
247 		if (t >= inp) {
248 			cnt = t - inp + 1;
249 			(void)memmove(out.dbp, inp, cnt);
250 			out.dbp += cnt;
251 			out.dbcnt += cnt;
252 		}
253 		++out.dbcnt;
254 		*out.dbp++ = '\n';
255 		if (out.dbcnt >= out.dbsz)
256 			dd_out(0);
257 	}
258 	if (in.dbcnt)
259 		(void)memmove(in.db, in.dbp - in.dbcnt, in.dbcnt);
260 	in.dbp = in.db + in.dbcnt;
261 }
262 
263 void
264 unblock_close()
265 {
266 	size_t cnt;
267 	u_char *t;
268 
269 	if (in.dbcnt) {
270 		warnx("%s: short input record", in.name);
271 		for (t = in.db + in.dbcnt - 1; t >= in.db && *t == ' '; --t);
272 		if (t >= in.db) {
273 			cnt = t - in.db + 1;
274 			(void)memmove(out.dbp, in.db, cnt);
275 			out.dbp += cnt;
276 			out.dbcnt += cnt;
277 		}
278 		++out.dbcnt;
279 		*out.dbp++ = '\n';
280 	}
281 }
282 
283 #endif	/* NO_CONV */
284