xref: /freebsd/bin/dd/conv.c (revision cbf6f7d3)
1 /*-
2  * Copyright (c) 1991, 1993, 1994
3  *	The Regents of the University of California.  All rights reserved.
4  *
5  * This code is derived from software contributed to Berkeley by
6  * Keith Muller of the University of California, San Diego and Lance
7  * Visser of Convex Computer Corporation.
8  *
9  * Redistribution and use in source and binary forms, with or without
10  * modification, are permitted provided that the following conditions
11  * are met:
12  * 1. Redistributions of source code must retain the above copyright
13  *    notice, this list of conditions and the following disclaimer.
14  * 2. Redistributions in binary form must reproduce the above copyright
15  *    notice, this list of conditions and the following disclaimer in the
16  *    documentation and/or other materials provided with the distribution.
17  * 3. All advertising materials mentioning features or use of this software
18  *    must display the following acknowledgement:
19  *	This product includes software developed by the University of
20  *	California, Berkeley and its contributors.
21  * 4. Neither the name of the University nor the names of its contributors
22  *    may be used to endorse or promote products derived from this software
23  *    without specific prior written permission.
24  *
25  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
26  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
27  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
28  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
29  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
30  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
31  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
32  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
34  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35  * SUCH DAMAGE.
36  */
37 
38 #ifndef lint
39 #if 0
40 static char const sccsid[] = "@(#)conv.c	8.3 (Berkeley) 4/2/94";
41 #endif
42 static const char rcsid[] =
43 	"$Id$";
44 #endif /* not lint */
45 
46 #include <sys/param.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 	int cnt;
64 	u_char *inp, *t;
65 
66 	if ((t = ctab) != NULL)
67 		for (inp = in.dbp - (cnt = in.dbrcnt); cnt--; ++inp)
68 			*inp = t[*inp];
69 
70 	/* Make the output buffer look right. */
71 	out.dbp = in.dbp;
72 	out.dbcnt = in.dbcnt;
73 
74 	if (in.dbcnt >= out.dbsz) {
75 		/* If the output buffer is full, write it. */
76 		dd_out(0);
77 
78 		/*
79 		 * Ddout copies the leftover output to the beginning of
80 		 * the buffer and resets the output buffer.  Reset the
81 		 * input buffer to match it.
82 	 	 */
83 		in.dbp = out.dbp;
84 		in.dbcnt = out.dbcnt;
85 	}
86 }
87 
88 void
89 def_close()
90 {
91 	/* Just update the count, everything is already in the buffer. */
92 	if (in.dbcnt)
93 		out.dbcnt = in.dbcnt;
94 }
95 
96 /*
97  * Copy variable length newline terminated records with a max size cbsz
98  * bytes to output.  Records less than cbs are padded with spaces.
99  *
100  * max in buffer:  MAX(ibs, cbsz)
101  * max out buffer: obs + cbsz
102  */
103 void
104 block()
105 {
106 	static int intrunc;
107 	int ch, cnt, maxlen;
108 	u_char *inp, *outp, *t;
109 
110 	/*
111 	 * Record truncation can cross block boundaries.  If currently in a
112 	 * truncation state, keep tossing characters until reach a newline.
113 	 * Start at the beginning of the buffer, as the input buffer is always
114 	 * left empty.
115 	 */
116 	if (intrunc) {
117 		for (inp = in.db, cnt = in.dbrcnt;
118 		    cnt && *inp++ != '\n'; --cnt);
119 		if (!cnt) {
120 			in.dbcnt = 0;
121 			in.dbp = in.db;
122 			return;
123 		}
124 		intrunc = 0;
125 		/* Adjust the input buffer numbers. */
126 		in.dbcnt = cnt - 1;
127 		in.dbp = inp + cnt - 1;
128 	}
129 
130 	/*
131 	 * Copy records (max cbsz size chunks) into the output buffer.  The
132 	 * translation is done as we copy into the output buffer.
133 	 */
134 	ch = 0;
135 	for (inp = in.dbp - in.dbcnt, outp = out.dbp; in.dbcnt;) {
136 		maxlen = MIN(cbsz, in.dbcnt);
137 		if ((t = ctab) != NULL)
138 			for (cnt = 0;
139 			    cnt < maxlen && (ch = *inp++) != '\n'; ++cnt)
140 				*outp++ = t[ch];
141 		else
142 			for (cnt = 0;
143 			    cnt < maxlen && (ch = *inp++) != '\n'; ++cnt)
144 				*outp++ = ch;
145 		/*
146 		 * Check for short record without a newline.  Reassemble the
147 		 * input block.
148 		 */
149 		if (ch != '\n' && in.dbcnt < cbsz) {
150 			memmove(in.db, in.dbp - in.dbcnt, in.dbcnt);
151 			break;
152 		}
153 
154 		/* Adjust the input buffer numbers. */
155 		in.dbcnt -= cnt;
156 		if (ch == '\n')
157 			--in.dbcnt;
158 
159 		/* Pad short records with spaces. */
160 		if (cnt < cbsz)
161 			(void)memset(outp, ctab ? ctab[' '] : ' ', cbsz - cnt);
162 		else {
163 			/*
164 			 * If the next character wouldn't have ended the
165 			 * block, it's a truncation.
166 			 */
167 			if (!in.dbcnt || *inp != '\n')
168 				++st.trunc;
169 
170 			/* Toss characters to a newline. */
171 			for (; in.dbcnt && *inp++ != '\n'; --in.dbcnt);
172 			if (!in.dbcnt)
173 				intrunc = 1;
174 			else
175 				--in.dbcnt;
176 		}
177 
178 		/* Adjust output buffer numbers. */
179 		out.dbp += cbsz;
180 		if ((out.dbcnt += cbsz) >= out.dbsz)
181 			dd_out(0);
182 		outp = out.dbp;
183 	}
184 	in.dbp = in.db + in.dbcnt;
185 }
186 
187 void
188 block_close()
189 {
190 	/*
191 	 * Copy any remaining data into the output buffer and pad to a record.
192 	 * Don't worry about truncation or translation, the input buffer is
193 	 * always empty when truncating, and no characters have been added for
194 	 * translation.  The bottom line is that anything left in the input
195 	 * buffer is a truncated record.  Anything left in the output buffer
196 	 * just wasn't big enough.
197 	 */
198 	if (in.dbcnt) {
199 		++st.trunc;
200 		memmove(out.dbp, in.dbp - in.dbcnt, in.dbcnt);
201 		(void)memset(out.dbp + in.dbcnt,
202 		    ctab ? ctab[' '] : ' ', cbsz - in.dbcnt);
203 		out.dbcnt += cbsz;
204 	}
205 }
206 
207 /*
208  * Convert fixed length (cbsz) records to variable length.  Deletes any
209  * trailing blanks and appends a newline.
210  *
211  * max in buffer:  MAX(ibs, cbsz) + cbsz
212  * max out buffer: obs + cbsz
213  */
214 void
215 unblock()
216 {
217 	int cnt;
218 	u_char *inp, *t;
219 
220 	/* Translation and case conversion. */
221 	if ((t = ctab) != NULL)
222 		for (cnt = in.dbrcnt, inp = in.dbp; cnt--;)
223 			*--inp = t[*inp];
224 	/*
225 	 * Copy records (max cbsz size chunks) into the output buffer.  The
226 	 * translation has to already be done or we might not recognize the
227 	 * spaces.
228 	 */
229 	for (inp = in.db; in.dbcnt >= cbsz; inp += cbsz, in.dbcnt -= cbsz) {
230 		for (t = inp + cbsz - 1; t >= inp && *t == ' '; --t);
231 		if (t >= inp) {
232 			cnt = t - inp + 1;
233 			memmove(out.dbp, inp, cnt);
234 			out.dbp += cnt;
235 			out.dbcnt += cnt;
236 		}
237 		++out.dbcnt;
238 		*out.dbp++ = '\n';
239 		if (out.dbcnt >= out.dbsz)
240 			dd_out(0);
241 	}
242 	if (in.dbcnt)
243 		memmove(in.db, in.dbp - in.dbcnt, in.dbcnt);
244 	in.dbp = in.db + in.dbcnt;
245 }
246 
247 void
248 unblock_close()
249 {
250 	int cnt;
251 	u_char *t;
252 
253 	if (in.dbcnt) {
254 		warnx("%s: short input record", in.name);
255 		for (t = in.db + in.dbcnt - 1; t >= in.db && *t == ' '; --t);
256 		if (t >= in.db) {
257 			cnt = t - in.db + 1;
258 			memmove(out.dbp, in.db, cnt);
259 			out.dbp += cnt;
260 			out.dbcnt += cnt;
261 		}
262 		++out.dbcnt;
263 		*out.dbp++ = '\n';
264 	}
265 }
266