Commit | Line | Data |
---|---|---|
5ad3291e KB |
1 | /* |
2 | * Copyright (c) 1988 The Regents of the University of California. | |
3 | * All rights reserved. | |
4 | * | |
1c15e888 C |
5 | * Redistribution and use in source and binary forms are permitted provided |
6 | * that: (1) source distributions retain this entire copyright notice and | |
7 | * comment, and (2) distributions including binaries display the following | |
8 | * acknowledgement: ``This product includes software developed by the | |
9 | * University of California, Berkeley and its contributors'' in the | |
10 | * documentation or other materials provided with the distribution and in | |
11 | * all advertising materials mentioning features or use of this software. | |
12 | * Neither the name of the University nor the names of its contributors may | |
13 | * be used to endorse or promote products derived from this software without | |
14 | * specific prior written permission. | |
15 | * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR IMPLIED | |
16 | * WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED WARRANTIES OF | |
17 | * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE. | |
5ad3291e KB |
18 | */ |
19 | ||
20 | #ifndef lint | |
21 | char copyright[] = | |
22 | "@(#) Copyright (c) 1988 The Regents of the University of California.\n\ | |
23 | All rights reserved.\n"; | |
24 | #endif /* not lint */ | |
25 | ||
26 | #ifndef lint | |
1c15e888 | 27 | static char sccsid[] = "@(#)tr.c 4.7 (Berkeley) 7/23/90"; |
5ad3291e KB |
28 | #endif /* not lint */ |
29 | ||
30 | #include <sys/types.h> | |
14db73f8 | 31 | #include <stdio.h> |
5ad3291e | 32 | #include <ctype.h> |
14db73f8 | 33 | |
5ad3291e KB |
34 | #define NCHARS 256 /* size of u_char */ |
35 | #define OOBCH 257 /* out of band value */ | |
14db73f8 | 36 | |
5ad3291e KB |
37 | typedef struct { |
38 | char *str; | |
39 | int lastch, endrange; | |
40 | enum { NORM, INRANGE, EOS } state; | |
41 | } STR; | |
14db73f8 | 42 | |
5ad3291e KB |
43 | main(argc, argv) |
44 | int argc; | |
45 | char **argv; | |
46 | { | |
47 | extern int optind; | |
48 | STR s1, s2; | |
49 | register int ch, indx, lastch; | |
50 | int cflag, dflag, sflag; | |
51 | u_char *tp, tab[NCHARS], squeeze[NCHARS]; | |
14db73f8 | 52 | |
5ad3291e KB |
53 | cflag = dflag = sflag = 0; |
54 | while ((ch = getopt(argc, argv, "cds")) != EOF) | |
55 | switch((char)ch) { | |
56 | case 'c': | |
57 | cflag = 1; | |
58 | break; | |
59 | case 'd': | |
60 | dflag = 1; | |
61 | break; | |
62 | case 's': | |
63 | sflag = 1; | |
64 | break; | |
65 | case '?': | |
66 | default: | |
67 | fprintf(stderr, | |
68 | "usage: tr [-cds] [string1 [string2]]\n"); | |
69 | exit(1); | |
14db73f8 | 70 | } |
5ad3291e KB |
71 | argc -= optind; |
72 | argv += optind; | |
73 | ||
74 | /* | |
75 | * the original tr was amazingly tolerant of the command line. | |
76 | * Neither -c or -s have any effect unless there are two strings. | |
77 | * Extra arguments are silently ignored. Bag this noise, they | |
78 | * should all be errors. | |
79 | */ | |
80 | if (argc < 2 && !dflag) { | |
81 | while ((ch = getchar()) != EOF) | |
82 | putchar(ch); | |
83 | exit(0); | |
14db73f8 BJ |
84 | } |
85 | ||
5ad3291e KB |
86 | bzero(tab, NCHARS); |
87 | if (sflag) { | |
88 | s1.str = argv[1]; | |
89 | s1.state = NORM; | |
90 | s1.lastch = OOBCH; | |
91 | while (next(&s1)) | |
92 | squeeze[s1.lastch] = 1; | |
93 | } | |
94 | if (dflag) { | |
95 | s1.str = argv[0]; | |
96 | s1.state = NORM; | |
97 | s1.lastch = OOBCH; | |
98 | while (next(&s1)) | |
99 | tab[s1.lastch] = 1; | |
100 | if (cflag) | |
101 | for (tp = tab, indx = 0; indx < NCHARS; ++tp, ++indx) | |
102 | *tp = !*tp; | |
103 | if (sflag) | |
104 | for (lastch = OOBCH; (ch = getchar()) != EOF;) { | |
105 | if (tab[ch] || (squeeze[ch] && lastch == ch)) | |
106 | continue; | |
107 | lastch = ch; | |
108 | putchar(ch); | |
109 | } | |
110 | else | |
111 | while ((ch = getchar()) != EOF) | |
112 | if (!tab[ch]) | |
113 | putchar(ch); | |
114 | } else { | |
115 | s1.str = argv[0]; | |
116 | s2.str = argv[1]; | |
117 | s1.state = s2.state = NORM; | |
118 | s1.lastch = s2.lastch = OOBCH; | |
119 | if (cflag) { | |
120 | /* | |
121 | * if cflag is set, tr just pretends it only got one | |
122 | * character in string2. As reasonable as anything | |
123 | * else. Should really be an error. | |
124 | */ | |
125 | while (next(&s2)); | |
126 | lastch = s2.lastch; | |
127 | for (tp = tab, indx = 0; indx < NCHARS; ++tp, ++indx) | |
128 | *tp = lastch; | |
129 | while (next(&s1)) | |
130 | tab[s1.lastch] = s1.lastch; | |
131 | } else { | |
132 | for (tp = tab, indx = 0; indx < NCHARS; ++tp, ++indx) | |
133 | *tp = indx; | |
134 | while (next(&s1)) { | |
135 | (void)next(&s2); | |
136 | tab[s1.lastch] = s2.lastch; | |
137 | } | |
138 | } | |
139 | if (sflag) | |
140 | for (lastch = OOBCH; (ch = getchar()) != EOF;) { | |
141 | ch = tab[ch]; | |
142 | if (squeeze[ch] && lastch == ch) | |
143 | continue; | |
144 | lastch = ch; | |
145 | putchar(ch); | |
bbcc5e9f | 146 | } |
5ad3291e KB |
147 | else |
148 | while ((ch = getchar()) != EOF) | |
149 | putchar((int)tab[ch]); | |
14db73f8 BJ |
150 | } |
151 | exit(0); | |
152 | } | |
153 | ||
154 | next(s) | |
20f87434 | 155 | register STR *s; |
14db73f8 | 156 | { |
5ad3291e | 157 | register int ch; |
14db73f8 | 158 | |
5ad3291e KB |
159 | if (s->state == EOS) |
160 | return(0); | |
161 | if (s->state == INRANGE) { | |
162 | if (++s->lastch == s->endrange) | |
163 | s->state = NORM; | |
164 | return(1); | |
165 | } | |
166 | if (!(ch = *s->str++)) { | |
167 | s->state = EOS; | |
168 | return(0); | |
14db73f8 | 169 | } |
44ec1783 KB |
170 | if (ch == '\\') { /* \### */ |
171 | s->lastch = tran(s); | |
5ad3291e KB |
172 | return(1); |
173 | } | |
174 | if (ch == '-') { /* ranges */ | |
44ec1783 KB |
175 | if (s->lastch == OOBCH) /* "-a" */ |
176 | goto fail2; | |
177 | if (!(ch = *s->str++)) /* "a-" */ | |
178 | goto fail1; | |
179 | if (ch == '\\') /* \### */ | |
180 | ch = tran(s); | |
5ad3291e | 181 | if (s->lastch > ch) { /* "z-a" */ |
44ec1783 KB |
182 | fail1: --s->str; |
183 | fail2: s->lastch = '-'; | |
5ad3291e | 184 | return(1); |
14db73f8 | 185 | } |
5ad3291e KB |
186 | if (s->lastch == ch) /* "a-a" */ |
187 | return(next(s)); | |
188 | s->state = INRANGE; /* "a-z" */ | |
44ec1783 | 189 | s->endrange = ch; |
5ad3291e | 190 | return(1); |
14db73f8 | 191 | } |
5ad3291e KB |
192 | s->lastch = ch; |
193 | return(1); | |
14db73f8 | 194 | } |
44ec1783 | 195 | |
20f87434 CT |
196 | /* |
197 | * Translate \-escapes. Up to 3 octal digits => char; no digits => literal. | |
198 | * Unadorned backslash "\" is like \000. | |
199 | */ | |
44ec1783 KB |
200 | tran(s) |
201 | register STR *s; | |
202 | { | |
20f87434 | 203 | register int ch, cnt = 0, val = 0; |
44ec1783 | 204 | |
20f87434 CT |
205 | for (;;) { |
206 | ch = *s->str++; | |
207 | if (!isascii(ch) || !isdigit(ch) || ++cnt > 3) | |
208 | break; | |
44ec1783 | 209 | val = val * 8 + ch - '0'; |
20f87434 CT |
210 | } |
211 | if (cnt || ch == 0) | |
212 | s->str--; | |
213 | return (cnt ? val : ch); | |
44ec1783 | 214 | } |