1 /* tr -- a filter to translate characters
2 Copyright (C) 91, 1995-2003 Free Software Foundation, Inc.
4 This program is free software; you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation; either version 2, or (at your option)
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program; if not, write to the Free Software Foundation,
16 Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */
18 /* Written by Jim Meyering */
24 #include <sys/types.h>
29 #include "safe-read.h"
32 /* The official name of this program (e.g., no `g' prefix). */
33 #define PROGRAM_NAME "tr"
35 #define WRITTEN_BY _("Written by Jim Meyering.")
37 #define N_CHARS (UCHAR_MAX + 1)
39 /* A pointer to a filtering function. */
40 typedef size_t (*Filter) (/* unsigned char *, size_t, Filter */);
42 /* Convert from character C to its index in the collating
43 sequence array. Just cast to an unsigned int to avoid
44 problems with sign-extension. */
45 #define ORD(c) (unsigned int)(c)
47 /* The inverse of ORD. */
48 #define CHR(i) (unsigned char)(i)
50 /* The value for Spec_list->state that indicates to
51 get_next that it should initialize the tail pointer.
52 Its value should be as large as possible to avoid conflict
53 a valid value for the state field -- and that may be as
54 large as any valid repeat_count. */
55 #define BEGIN_STATE (INT_MAX - 1)
57 /* The value for Spec_list->state that indicates to
58 get_next that the element pointed to by Spec_list->tail is
59 being considered for the first time on this pass through the
60 list -- it indicates that get_next should make any necessary
62 #define NEW_ELEMENT (BEGIN_STATE + 1)
64 /* A value distinct from any character that may have been stored in a
65 buffer as the result of a block-read in the function squeeze_filter. */
66 #define NOT_A_CHAR (unsigned int)(-1)
68 /* The following (but not CC_NO_CLASS) are indices into the array of
69 valid character class strings. */
72 CC_ALNUM = 0, CC_ALPHA = 1, CC_BLANK = 2, CC_CNTRL = 3,
73 CC_DIGIT = 4, CC_GRAPH = 5, CC_LOWER = 6, CC_PRINT = 7,
74 CC_PUNCT = 8, CC_SPACE = 9, CC_UPPER = 10, CC_XDIGIT = 11,
78 /* Character class to which a character (returned by get_next) belonged;
79 but it is set only if the construct from which the character was obtained
80 was one of the character classes [:upper:] or [:lower:]. The value
81 is used only when translating and then, only to make sure that upper
82 and lower class constructs have the same relative positions in string1
84 enum Upper_Lower_class
91 /* A shortcut to ensure that when constructing the translation array,
92 one of the values returned by paired calls to get_next (from s1 and s2)
93 is from [:upper:] and the other is from [:lower:], or neither is from
94 upper or lower. By default, GNU tr permits the identity mappings: from
95 [:upper:] to [:upper:] and [:lower:] to [:lower:]. But when
96 POSIXLY_CORRECT is set, those evoke diagnostics. This array is indexed
97 by values of type enum Upper_Lower_class. */
98 static int const class_ok[3][3] =
105 /* The type of a List_element. See build_spec_list for more details. */
106 enum Range_element_type
116 /* One construct in one of tr's argument strings.
117 For example, consider the POSIX version of the classic tr command:
118 tr -cs 'a-zA-Z_' '[\n*]'
119 String1 has 3 constructs, two of which are ranges (a-z and A-Z),
120 and a single normal character, `_'. String2 has one construct. */
123 enum Range_element_type type;
124 struct List_element *next;
130 unsigned int first_char;
131 unsigned int last_char;
134 enum Char_class char_class;
138 unsigned int the_repeated_char;
146 /* Each of tr's argument strings is parsed into a form that is easier
147 to work with: a linked list of constructs (struct List_element).
148 Each Spec_list structure also encapsulates various attributes of
149 the corresponding argument string. The attributes are used mainly
150 to verify that the strings are valid in the context of any options
151 specified (like -s, -d, or -c). The main exception is the member
152 `tail', which is first used to construct the list. After construction,
153 it is used by get_next to save its state when traversing the list.
154 The member `state' serves a similar function. */
157 /* Points to the head of the list of range elements.
158 The first struct is a dummy; its members are never used. */
159 struct List_element *head;
161 /* When appending, points to the last element. When traversing via
162 get_next(), points to the element to process next. Setting
163 Spec_list.state to the value BEGIN_STATE before calling get_next
164 signals get_next to initialize tail to point to head->next. */
165 struct List_element *tail;
167 /* Used to save state between calls to get_next(). */
170 /* Length, in the sense that length ('a-z[:digit:]123abc')
171 is 42 ( = 26 + 10 + 6). */
174 /* The number of [c*] and [c*0] constructs that appear in this spec. */
175 int n_indefinite_repeats;
177 /* If n_indefinite_repeats is nonzero, this points to the List_element
178 corresponding to the last [c*] or [c*0] construct encountered in
179 this spec. Otherwise it is undefined. */
180 struct List_element *indefinite_repeat_element;
182 /* Non-zero if this spec contains at least one equivalence
183 class construct e.g. [=c=]. */
186 /* Non-zero if this spec contains at least one character class
187 construct. E.g. [:digit:]. */
190 /* Non-zero if this spec contains at least one of the character class
191 constructs (all but upper and lower) that aren't allowed in s2. */
192 int has_restricted_char_class;
195 /* A representation for escaped string1 or string2. As a string is parsed,
196 any backslash-escaped characters (other than octal or \a, \b, \f, \n,
197 etc.) are marked as such in this structure by setting the corresponding
198 entry in the ESCAPED vector. */
206 /* Return nonzero if the Ith character of escaped string ES matches C
207 and is not escaped itself. */
208 #define ES_MATCH(ES, I, C) ((ES)->s[(I)] == (C) && !(ES)->escaped[(I)])
210 /* The name by which this program was run. */
213 /* When nonzero, each sequence in the input of a repeated character
214 (call it c) is replaced (in the output) by a single occurrence of c
215 for every c in the squeeze set. */
216 static int squeeze_repeats = 0;
218 /* When nonzero, removes characters in the delete set from input. */
219 static int delete = 0;
221 /* Use the complement of set1 in place of set1. */
222 static int complement = 0;
224 /* When nonzero, this flag causes GNU tr to provide strict
225 compliance with POSIX draft 1003.2.11.2. The POSIX spec
226 says that when -d is used without -s, string2 (if present)
227 must be ignored. Silently ignoring arguments is a bad idea.
228 The default GNU behavior is to give a usage message and exit.
229 Additionally, when this flag is nonzero, tr prints warnings
230 on stderr if it is being used in a manner that is not portable.
231 Applicable warnings are given by default, but are suppressed
232 if the environment variable `POSIXLY_CORRECT' is set, since
233 being POSIX conformant means we can't issue such messages.
234 Warnings on the following topics are suppressed when this
236 1. Ambiguous octal escapes. */
237 static int posix_pedantic;
239 /* When tr is performing translation and string1 is longer than string2,
240 POSIX says that the result is undefined. That gives the implementor
241 of a POSIX conforming version of tr two reasonable choices for the
242 semantics of this case.
244 * The BSD tr pads string2 to the length of string1 by
245 repeating the last character in string2.
247 * System V tr ignores characters in string1 that have no
248 corresponding character in string2. That is, string1 is effectively
249 truncated to the length of string2.
251 When nonzero, this flag causes GNU tr to imitate the behavior
252 of System V tr when translating with string1 longer than string2.
253 The default is to emulate BSD tr. This flag is ignored in modes where
254 no translation is performed. Emulating the System V tr
255 in this exceptional case causes the relatively common BSD idiom:
257 tr -cs A-Za-z0-9 '\012'
259 to break (it would convert only zero bytes, rather than all
260 non-alphanumerics, to newlines).
262 WARNING: This switch does not provide general BSD or System V
263 compatibility. For example, it doesn't disable the interpretation
264 of the POSIX constructs [:alpha:], [=c=], and [c*10], so if by
265 some unfortunate coincidence you use such constructs in scripts
266 expecting to use some other version of tr, the scripts will break. */
267 static int truncate_set1 = 0;
269 /* An alias for (!delete && non_option_args == 2).
270 It is set in main and used there and in validate(). */
271 static int translating;
273 static unsigned char io_buf[BUFSIZ];
275 static char const *const char_class_name[] =
277 "alnum", "alpha", "blank", "cntrl", "digit", "graph",
278 "lower", "print", "punct", "space", "upper", "xdigit"
280 #define N_CHAR_CLASSES (sizeof(char_class_name) / sizeof(char_class_name[0]))
282 typedef char SET_TYPE;
284 /* Array of boolean values. A character `c' is a member of the
285 squeeze set if and only if in_squeeze_set[c] is true. The squeeze
286 set is defined by the last (possibly, the only) string argument
287 on the command line when the squeeze option is given. */
288 static SET_TYPE in_squeeze_set[N_CHARS];
290 /* Array of boolean values. A character `c' is a member of the
291 delete set if and only if in_delete_set[c] is true. The delete
292 set is defined by the first (or only) string argument on the
293 command line when the delete option is given. */
294 static SET_TYPE in_delete_set[N_CHARS];
296 /* Array of character values defining the translation (if any) that
297 tr is to perform. Translation is performed only when there are
298 two specification strings and the delete switch is not given. */
299 static char xlate[N_CHARS];
301 static struct option const long_options[] =
303 {"complement", no_argument, NULL, 'c'},
304 {"delete", no_argument, NULL, 'd'},
305 {"squeeze-repeats", no_argument, NULL, 's'},
306 {"truncate-set1", no_argument, NULL, 't'},
307 {GETOPT_HELP_OPTION_DECL},
308 {GETOPT_VERSION_OPTION_DECL},
316 fprintf (stderr, _("Try `%s --help' for more information.\n"),
321 Usage: %s [OPTION]... SET1 [SET2]\n\
325 Translate, squeeze, and/or delete characters from standard input,\n\
326 writing to standard output.\n\
328 -c, --complement first complement SET1\n\
329 -d, --delete delete characters in SET1, do not translate\n\
330 -s, --squeeze-repeats replace each input sequence of a repeated character\n\
331 that is listed in SET1 with a single occurrence\n\
333 -t, --truncate-set1 first truncate SET1 to length of SET2\n\
335 fputs (HELP_OPTION_DESCRIPTION, stdout);
336 fputs (VERSION_OPTION_DESCRIPTION, stdout);
339 SETs are specified as strings of characters. Most represent themselves.\n\
340 Interpreted sequences are:\n\
342 \\NNN character with octal value NNN (1 to 3 octal digits)\n\
349 \\t horizontal tab\n\
353 CHAR1-CHAR2 all characters from CHAR1 to CHAR2 in ascending order\n\
354 [CHAR*] in SET2, copies of CHAR until length of SET1\n\
355 [CHAR*REPEAT] REPEAT copies of CHAR, REPEAT octal if starting with 0\n\
356 [:alnum:] all letters and digits\n\
357 [:alpha:] all letters\n\
358 [:blank:] all horizontal whitespace\n\
359 [:cntrl:] all control characters\n\
360 [:digit:] all digits\n\
363 [:graph:] all printable characters, not including space\n\
364 [:lower:] all lower case letters\n\
365 [:print:] all printable characters, including space\n\
366 [:punct:] all punctuation characters\n\
367 [:space:] all horizontal or vertical whitespace\n\
368 [:upper:] all upper case letters\n\
369 [:xdigit:] all hexadecimal digits\n\
370 [=CHAR=] all characters which are equivalent to CHAR\n\
374 Translation occurs if -d is not given and both SET1 and SET2 appear.\n\
375 -t may be used only when translating. SET2 is extended to length of\n\
376 SET1 by repeating its last character as necessary. \
380 of SET2 are ignored. Only [:lower:] and [:upper:] are guaranteed to\n\
381 expand in ascending order; used in SET2 while translating, they may\n\
382 only be used in pairs to specify case conversion. \
385 -s uses SET1 if not\n\
386 translating nor deleting; else squeezing uses SET2 and occurs after\n\
387 translation or deletion.\n\
389 printf (_("\nReport bugs to <%s>.\n"), PACKAGE_BUGREPORT);
391 exit (status == 0 ? EXIT_SUCCESS : EXIT_FAILURE);
394 /* Return nonzero if the character C is a member of the
395 equivalence class containing the character EQUIV_CLASS. */
398 is_equiv_class_member (unsigned int equiv_class, unsigned int c)
400 return (equiv_class == c);
403 /* Return nonzero if the character C is a member of the
404 character class CHAR_CLASS. */
407 is_char_class_member (enum Char_class char_class, unsigned int c)
414 result = ISALNUM (c);
417 result = ISALPHA (c);
420 result = ISBLANK (c);
423 result = ISCNTRL (c);
426 result = ISDIGIT_LOCALE (c);
429 result = ISGRAPH (c);
432 result = ISLOWER (c);
435 result = ISPRINT (c);
438 result = ISPUNCT (c);
441 result = ISSPACE (c);
444 result = ISUPPER (c);
447 result = ISXDIGIT (c);
457 es_free (struct E_string *es)
463 /* Perform the first pass over each range-spec argument S, converting all
464 \c and \ddd escapes to their one-byte representations. The conversion
465 is done in-place, so S must point to writable storage. If an invalid
466 quote sequence is found print an error message and return nonzero.
467 Otherwise set *LEN to the length of the resulting string and return
468 zero. The resulting array of characters may contain zero-bytes;
469 however, on input, S is assumed to be null-terminated, and hence
470 cannot contain actual (non-escaped) zero bytes. */
473 unquote (const unsigned char *s, struct E_string *es)
478 len = strlen ((char *) s);
480 es->s = xmalloc (len);
481 es->escaped = xmalloc (len * sizeof (es->escaped[0]));
482 for (i = 0; i < len; i++)
486 for (i = 0; s[i]; i++)
528 oct_digit = s[i + 2] - '0';
529 if (0 <= oct_digit && oct_digit <= 7)
531 c = 8 * c + oct_digit;
533 oct_digit = s[i + 2] - '0';
534 if (0 <= oct_digit && oct_digit <= 7)
536 if (8 * c + oct_digit < N_CHARS)
538 c = 8 * c + oct_digit;
541 else if (!posix_pedantic)
543 /* A 3-digit octal number larger than \377 won't
544 fit in 8 bits. So we stop when adding the
545 next digit would put us over the limit and
546 give a warning about the ambiguity. POSIX
547 isn't clear on this, but one person has said
548 that in his interpretation, POSIX says tr
549 can't even give a warning. */
550 error (0, 0, _("warning: the ambiguous octal escape \
551 \\%c%c%c is being\n\tinterpreted as the 2-byte sequence \\0%c%c, `%c'"),
552 s[i], s[i + 1], s[i + 2],
553 s[i], s[i + 1], s[i + 2]);
559 error (0, 0, _("invalid backslash escape at end of string"));
565 error (0, 0, _("invalid backslash escape `\\%c'"), s[i + 1]);
586 /* If CLASS_STR is a valid character class string, return its index
587 in the global char_class_name array. Otherwise, return CC_NO_CLASS. */
589 static enum Char_class
590 look_up_char_class (const unsigned char *class_str, size_t len)
594 for (i = 0; i < N_CHAR_CLASSES; i++)
595 if (strncmp ((const char *) class_str, char_class_name[i], len) == 0
596 && strlen (char_class_name[i]) == len)
597 return (enum Char_class) i;
601 /* Return a newly allocated string with a printable version of C.
602 This function is used solely for formatting error messages. */
605 make_printable_char (unsigned int c)
607 char *buf = xmalloc (5);
609 assert (c < N_CHARS);
617 sprintf (buf, "\\%03o", c);
622 /* Return a newly allocated copy of S which is suitable for printing.
623 LEN is the number of characters in S. Most non-printing
624 (isprint) characters are represented by a backslash followed by
625 3 octal digits. However, the characters represented by \c escapes
626 where c is one of [abfnrtv] are represented by their 2-character \c
627 sequences. This function is used solely for printing error messages. */
630 make_printable_str (const unsigned char *s, size_t len)
632 /* Worst case is that every character expands to a backslash
633 followed by a 3-character octal escape sequence. */
634 char *printable_buf = xmalloc (4 * len + 1);
635 char *p = printable_buf;
638 for (i = 0; i < len; i++)
676 sprintf (buf, "\\%03o", s[i]);
682 return printable_buf;
685 /* Append a newly allocated structure representing a
686 character C to the specification list LIST. */
689 append_normal_char (struct Spec_list *list, unsigned int c)
691 struct List_element *new;
693 new = xmalloc (sizeof (struct List_element));
695 new->type = RE_NORMAL_CHAR;
696 new->u.normal_char = c;
698 list->tail->next = new;
702 /* Append a newly allocated structure representing the range
703 of characters from FIRST to LAST to the specification list LIST.
704 Return nonzero if LAST precedes FIRST in the collating sequence,
705 zero otherwise. This means that '[c-c]' is acceptable. */
708 append_range (struct Spec_list *list, unsigned int first, unsigned int last)
710 struct List_element *new;
712 if (ORD (first) > ORD (last))
714 char *tmp1 = make_printable_char (first);
715 char *tmp2 = make_printable_char (last);
718 _("range-endpoints of `%s-%s' are in reverse collating sequence order"),
724 new = xmalloc (sizeof (struct List_element));
726 new->type = RE_RANGE;
727 new->u.range.first_char = first;
728 new->u.range.last_char = last;
730 list->tail->next = new;
735 /* If CHAR_CLASS_STR is a valid character class string, append a
736 newly allocated structure representing that character class to the end
737 of the specification list LIST and return 0. If CHAR_CLASS_STR is not
738 a valid string return nonzero. */
741 append_char_class (struct Spec_list *list,
742 const unsigned char *char_class_str, size_t len)
744 enum Char_class char_class;
745 struct List_element *new;
747 char_class = look_up_char_class (char_class_str, len);
748 if (char_class == CC_NO_CLASS)
750 new = xmalloc (sizeof (struct List_element));
752 new->type = RE_CHAR_CLASS;
753 new->u.char_class = char_class;
755 list->tail->next = new;
760 /* Append a newly allocated structure representing a [c*n]
761 repeated character construct to the specification list LIST.
762 THE_CHAR is the single character to be repeated, and REPEAT_COUNT
763 is a non-negative repeat count. */
766 append_repeated_char (struct Spec_list *list, unsigned int the_char,
769 struct List_element *new;
771 new = xmalloc (sizeof (struct List_element));
773 new->type = RE_REPEATED_CHAR;
774 new->u.repeated_char.the_repeated_char = the_char;
775 new->u.repeated_char.repeat_count = repeat_count;
777 list->tail->next = new;
781 /* Given a string, EQUIV_CLASS_STR, from a [=str=] context and
782 the length of that string, LEN, if LEN is exactly one, append
783 a newly allocated structure representing the specified
784 equivalence class to the specification list, LIST and return zero.
785 If LEN is not 1, return nonzero. */
788 append_equiv_class (struct Spec_list *list,
789 const unsigned char *equiv_class_str, size_t len)
791 struct List_element *new;
795 new = xmalloc (sizeof (struct List_element));
797 new->type = RE_EQUIV_CLASS;
798 new->u.equiv_code = *equiv_class_str;
800 list->tail->next = new;
805 /* Return a newly allocated copy of the LEN-byte prefix of P.
806 The returned string may contain NUL bytes and is *not* NUL-terminated. */
808 static unsigned char *
809 xmemdup (const unsigned char *p, size_t len)
811 unsigned char *tmp = xmalloc (len);
813 /* Use memcpy rather than strncpy because `p' may contain zero-bytes. */
814 memcpy (tmp, p, len);
818 /* Search forward starting at START_IDX for the 2-char sequence
819 (PRE_BRACKET_CHAR,']') in the string P of length P_LEN. If such
820 a sequence is found, set *RESULT_IDX to the index of the first
821 character and return nonzero. Otherwise return zero. P may contain
825 find_closing_delim (const struct E_string *es, size_t start_idx,
826 unsigned int pre_bracket_char, size_t *result_idx)
830 for (i = start_idx; i < es->len - 1; i++)
831 if (es->s[i] == pre_bracket_char && es->s[i + 1] == ']'
832 && !es->escaped[i] && !es->escaped[i + 1])
840 /* Parse the bracketed repeat-char syntax. If the P_LEN characters
841 beginning with P[ START_IDX ] comprise a valid [c*n] construct,
842 then set *CHAR_TO_REPEAT, *REPEAT_COUNT, and *CLOSING_BRACKET_IDX
843 and return zero. If the second character following
844 the opening bracket is not `*' or if no closing bracket can be
845 found, return -1. If a closing bracket is found and the
846 second char is `*', but the string between the `*' and `]' isn't
847 empty, an octal number, or a decimal number, print an error message
851 find_bracketed_repeat (const struct E_string *es, size_t start_idx,
852 unsigned int *char_to_repeat, size_t *repeat_count,
853 size_t *closing_bracket_idx)
857 assert (start_idx + 1 < es->len);
858 if (!ES_MATCH (es, start_idx + 1, '*'))
861 for (i = start_idx + 2; i < es->len; i++)
863 if (ES_MATCH (es, i, ']'))
865 size_t digit_str_len = i - start_idx - 2;
867 *char_to_repeat = es->s[start_idx];
868 if (digit_str_len == 0)
870 /* We've matched [c*] -- no explicit repeat count. */
872 *closing_bracket_idx = i;
876 /* Here, we have found [c*s] where s should be a string
877 of octal (if it starts with `0') or decimal digits. */
879 const char *digit_str = (const char *) &es->s[start_idx + 2];
880 unsigned long int tmp_ulong;
883 /* Select the base manually so we can be sure it's either 8 or 10.
884 If the spec allowed it to be interpreted as hexadecimal, we
885 could have used `0' and let xstrtoul decide. */
886 if (*digit_str == '0')
892 if (xstrtoul (digit_str, &d_end, base, &tmp_ulong, NULL)
894 || BEGIN_STATE < tmp_ulong
895 || digit_str + digit_str_len != d_end)
897 char *tmp = make_printable_str (es->s + start_idx + 2,
899 error (0, 0, _("invalid repeat count `%s' in [c*n] construct"),
904 *repeat_count = tmp_ulong;
906 *closing_bracket_idx = i;
910 return -1; /* No bracket found. */
913 /* Return nonzero if the string at ES->s[IDX] matches the regular
914 expression `\*[0-9]*\]', zero otherwise. To match, the `*' and
915 the `]' must not be escaped. */
918 star_digits_closebracket (const struct E_string *es, size_t idx)
922 if (!ES_MATCH (es, idx, '*'))
925 for (i = idx + 1; i < es->len; i++)
927 if (!ISDIGIT (es->s[i]))
929 if (ES_MATCH (es, i, ']'))
937 /* Convert string UNESCAPED_STRING (which has been preprocessed to
938 convert backslash-escape sequences) of length LEN characters into
939 a linked list of the following 5 types of constructs:
940 - [:str:] Character class where `str' is one of the 12 valid strings.
941 - [=c=] Equivalence class where `c' is any single character.
942 - [c*n] Repeat the single character `c' `n' times. n may be omitted.
943 However, if `n' is present, it must be a non-negative octal or
945 - r-s Range of characters from `r' to `s'. The second endpoint must
946 not precede the first in the current collating sequence.
947 - c Any other character is interpreted as itself. */
950 build_spec_list (const struct E_string *es, struct Spec_list *result)
952 const unsigned char *p;
957 /* The main for-loop below recognizes the 4 multi-character constructs.
958 A character that matches (in its context) none of the multi-character
959 constructs is classified as `normal'. Since all multi-character
960 constructs have at least 3 characters, any strings of length 2 or
961 less are composed solely of normal characters. Hence, the index of
962 the outer for-loop runs only as far as LEN-2. */
964 for (i = 0; i + 2 < es->len; /* empty */)
966 if (ES_MATCH (es, i, '['))
968 int matched_multi_char_construct;
969 size_t closing_bracket_idx;
970 unsigned int char_to_repeat;
974 matched_multi_char_construct = 1;
975 if (ES_MATCH (es, i + 1, ':')
976 || ES_MATCH (es, i + 1, '='))
978 size_t closing_delim_idx;
981 found = find_closing_delim (es, i + 2, p[i + 1],
986 size_t opnd_str_len = closing_delim_idx - 1 - (i + 2) + 1;
987 unsigned char *opnd_str;
989 if (opnd_str_len == 0)
992 error (0, 0, _("missing character class name `[::]'"));
995 _("missing equivalence class character `[==]'"));
999 opnd_str = xmemdup (p + i + 2, opnd_str_len);
1001 if (p[i + 1] == ':')
1003 parse_failed = append_char_class (result, opnd_str,
1006 /* FIXME: big comment. */
1009 if (star_digits_closebracket (es, i + 2))
1012 goto try_bracketed_repeat;
1016 char *tmp = make_printable_str (opnd_str,
1018 error (0, 0, _("invalid character class `%s'"),
1027 parse_failed = append_equiv_class (result, opnd_str,
1030 /* FIXME: big comment. */
1033 if (star_digits_closebracket (es, i + 2))
1036 goto try_bracketed_repeat;
1040 char *tmp = make_printable_str (opnd_str,
1043 _("%s: equivalence class operand must be a single character"),
1052 /* Return nonzero if append_*_class reports a problem. */
1056 i = closing_delim_idx + 2;
1059 /* Else fall through. This could be [:*] or [=*]. */
1062 try_bracketed_repeat:
1064 /* Determine whether this is a bracketed repeat range
1065 matching the RE \[.\*(dec_or_oct_number)?\]. */
1066 err = find_bracketed_repeat (es, i + 1, &char_to_repeat,
1068 &closing_bracket_idx);
1071 append_repeated_char (result, char_to_repeat, repeat_count);
1072 i = closing_bracket_idx + 1;
1076 matched_multi_char_construct = 0;
1080 /* Found a string that looked like [c*n] but the
1081 numeric part was invalid. */
1085 if (matched_multi_char_construct)
1088 /* We reach this point if P does not match [:str:], [=c=],
1089 [c*n], or [c*]. Now, see if P looks like a range `[-c'
1090 (from `[' to `c'). */
1093 /* Look ahead one char for ranges like a-z. */
1094 if (ES_MATCH (es, i + 1, '-'))
1096 if (append_range (result, p[i], p[i + 2]))
1102 append_normal_char (result, p[i]);
1107 /* Now handle the (2 or fewer) remaining characters p[i]..p[es->len - 1]. */
1108 for (; i < es->len; i++)
1109 append_normal_char (result, p[i]);
1114 /* Given a Spec_list S (with its saved state implicit in the values
1115 of its members `tail' and `state'), return the next single character
1116 in the expansion of S's constructs. If the last character of S was
1117 returned on the previous call or if S was empty, this function
1118 returns -1. For example, successive calls to get_next where S
1119 represents the spec-string 'a-d[y*3]' will return the sequence
1120 of values a, b, c, d, y, y, y, -1. Finally, if the construct from
1121 which the returned character comes is [:upper:] or [:lower:], the
1122 parameter CLASS is given a value to indicate which it was. Otherwise
1123 CLASS is set to UL_NONE. This value is used only when constructing
1124 the translation table to verify that any occurrences of upper and
1125 lower class constructs in the spec-strings appear in the same relative
1129 get_next (struct Spec_list *s, enum Upper_Lower_class *class)
1131 struct List_element *p;
1138 if (s->state == BEGIN_STATE)
1140 s->tail = s->head->next;
1141 s->state = NEW_ELEMENT;
1150 case RE_NORMAL_CHAR:
1151 return_val = p->u.normal_char;
1152 s->state = NEW_ELEMENT;
1157 if (s->state == NEW_ELEMENT)
1158 s->state = ORD (p->u.range.first_char);
1161 return_val = CHR (s->state);
1162 if (s->state == ORD (p->u.range.last_char))
1165 s->state = NEW_ELEMENT;
1173 switch (p->u.char_class)
1191 s->state = NEW_ELEMENT;
1197 if (s->state == NEW_ELEMENT)
1199 for (i = 0; i < N_CHARS; i++)
1200 if (is_char_class_member (p->u.char_class, i))
1202 assert (i < N_CHARS);
1205 assert (is_char_class_member (p->u.char_class, s->state));
1206 return_val = CHR (s->state);
1207 for (i = s->state + 1; i < N_CHARS; i++)
1208 if (is_char_class_member (p->u.char_class, i))
1215 s->state = NEW_ELEMENT;
1219 case RE_EQUIV_CLASS:
1220 /* FIXME: this assumes that each character is alone in its own
1221 equivalence class (which appears to be correct for my
1222 LC_COLLATE. But I don't know of any function that allows
1223 one to determine a character's equivalence class. */
1225 return_val = p->u.equiv_code;
1226 s->state = NEW_ELEMENT;
1230 case RE_REPEATED_CHAR:
1231 /* Here, a repeat count of n == 0 means don't repeat at all. */
1232 if (p->u.repeated_char.repeat_count == 0)
1235 s->state = NEW_ELEMENT;
1236 return_val = get_next (s, class);
1240 if (s->state == NEW_ELEMENT)
1245 return_val = p->u.repeated_char.the_repeated_char;
1246 if (p->u.repeated_char.repeat_count > 0
1247 && s->state == p->u.repeated_char.repeat_count)
1250 s->state = NEW_ELEMENT;
1267 /* This is a minor kludge. This function is called from
1268 get_spec_stats to determine the cardinality of a set derived
1269 from a complemented string. It's a kludge in that some of the
1270 same operations are (duplicated) performed in set_initialize. */
1273 card_of_complement (struct Spec_list *s)
1276 int cardinality = N_CHARS;
1277 SET_TYPE in_set[N_CHARS];
1279 memset (in_set, 0, N_CHARS * sizeof (in_set[0]));
1280 s->state = BEGIN_STATE;
1281 while ((c = get_next (s, NULL)) != -1)
1287 /* Gather statistics about the spec-list S in preparation for the tests
1288 in validate that determine the consistency of the specs. This function
1289 is called at most twice; once for string1, and again for any string2.
1290 LEN_S1 < 0 indicates that this is the first call and that S represents
1291 string1. When LEN_S1 >= 0, it is the length of the expansion of the
1292 constructs in string1, and we can use its value to resolve any
1293 indefinite repeat construct in S (which represents string2). Hence,
1294 this function has the side-effect that it converts a valid [c*]
1295 construct in string2 to [c*n] where n is large enough (or 0) to give
1296 string2 the same length as string1. For example, with the command
1297 tr a-z 'A[\n*]Z' on the second call to get_spec_stats, LEN_S1 would
1298 be 26 and S (representing string2) would be converted to 'A[\n*24]Z'. */
1301 get_spec_stats (struct Spec_list *s)
1303 struct List_element *p;
1306 s->n_indefinite_repeats = 0;
1307 s->has_equiv_class = 0;
1308 s->has_restricted_char_class = 0;
1309 s->has_char_class = 0;
1310 for (p = s->head->next; p; p = p->next)
1315 case RE_NORMAL_CHAR:
1320 assert (p->u.range.last_char >= p->u.range.first_char);
1321 len += p->u.range.last_char - p->u.range.first_char + 1;
1325 s->has_char_class = 1;
1326 for (i = 0; i < N_CHARS; i++)
1327 if (is_char_class_member (p->u.char_class, i))
1329 switch (p->u.char_class)
1335 s->has_restricted_char_class = 1;
1340 case RE_EQUIV_CLASS:
1341 for (i = 0; i < N_CHARS; i++)
1342 if (is_equiv_class_member (p->u.equiv_code, i))
1344 s->has_equiv_class = 1;
1347 case RE_REPEATED_CHAR:
1348 if (p->u.repeated_char.repeat_count > 0)
1349 len += p->u.repeated_char.repeat_count;
1350 else if (p->u.repeated_char.repeat_count == 0)
1352 s->indefinite_repeat_element = p;
1353 ++(s->n_indefinite_repeats);
1367 get_s1_spec_stats (struct Spec_list *s1)
1369 get_spec_stats (s1);
1371 s1->length = card_of_complement (s1);
1375 get_s2_spec_stats (struct Spec_list *s2, size_t len_s1)
1377 get_spec_stats (s2);
1378 if (len_s1 >= s2->length && s2->n_indefinite_repeats == 1)
1380 s2->indefinite_repeat_element->u.repeated_char.repeat_count =
1381 len_s1 - s2->length;
1382 s2->length = len_s1;
1387 spec_init (struct Spec_list *spec_list)
1389 spec_list->head = spec_list->tail =
1390 xmalloc (sizeof (struct List_element));
1391 spec_list->head->next = NULL;
1394 /* This function makes two passes over the argument string S. The first
1395 one converts all \c and \ddd escapes to their one-byte representations.
1396 The second constructs a linked specification list, SPEC_LIST, of the
1397 characters and constructs that comprise the argument string. If either
1398 of these passes detects an error, this function returns nonzero. */
1401 parse_str (const unsigned char *s, struct Spec_list *spec_list)
1406 fail = unquote (s, &es);
1408 fail = build_spec_list (&es, spec_list);
1413 /* Given two specification lists, S1 and S2, and assuming that
1414 S1->length > S2->length, append a single [c*n] element to S2 where c
1415 is the last character in the expansion of S2 and n is the difference
1416 between the two lengths.
1417 Upon successful completion, S2->length is set to S1->length. The only
1418 way this function can fail to make S2 as long as S1 is when S2 has
1419 zero-length, since in that case, there is no last character to repeat.
1420 So S2->length is required to be at least 1.
1422 Providing this functionality allows the user to do some pretty
1423 non-BSD (and non-portable) things: For example, the command
1424 tr -cs '[:upper:]0-9' '[:lower:]'
1425 is almost guaranteed to give results that depend on your collating
1429 string2_extend (const struct Spec_list *s1, struct Spec_list *s2)
1431 struct List_element *p;
1435 assert (translating);
1436 assert (s1->length > s2->length);
1437 assert (s2->length > 0);
1442 case RE_NORMAL_CHAR:
1443 char_to_repeat = p->u.normal_char;
1446 char_to_repeat = p->u.range.last_char;
1449 for (i = N_CHARS; i >= 0; i--)
1450 if (is_char_class_member (p->u.char_class, i))
1453 char_to_repeat = CHR (i);
1456 case RE_REPEATED_CHAR:
1457 char_to_repeat = p->u.repeated_char.the_repeated_char;
1460 case RE_EQUIV_CLASS:
1461 /* This shouldn't happen, because validate exits with an error
1462 if it finds an equiv class in string2 when translating. */
1475 append_repeated_char (s2, char_to_repeat, s1->length - s2->length);
1476 s2->length = s1->length;
1479 /* Return non-zero if S is a non-empty list in which exactly one
1480 character (but potentially, many instances of it) appears.
1481 E.g. [X*] or xxxxxxxx. */
1484 homogeneous_spec_list (struct Spec_list *s)
1488 s->state = BEGIN_STATE;
1490 if ((b = get_next (s, NULL)) == -1)
1493 while ((c = get_next (s, NULL)) != -1)
1500 /* Die with an error message if S1 and S2 describe strings that
1501 are not valid with the given command line switches.
1502 A side effect of this function is that if a valid [c*] or
1503 [c*0] construct appears in string2, it is converted to [c*n]
1504 with a value for n that makes s2->length == s1->length. By
1505 the same token, if the --truncate-set1 option is not
1506 given, S2 may be extended. */
1509 validate (struct Spec_list *s1, struct Spec_list *s2)
1511 get_s1_spec_stats (s1);
1512 if (s1->n_indefinite_repeats > 0)
1514 error (EXIT_FAILURE, 0,
1515 _("the [c*] repeat construct may not appear in string1"));
1520 get_s2_spec_stats (s2, s1->length);
1522 if (s2->n_indefinite_repeats > 1)
1524 error (EXIT_FAILURE, 0,
1525 _("only one [c*] repeat construct may appear in string2"));
1530 if (s2->has_equiv_class)
1532 error (EXIT_FAILURE, 0,
1533 _("[=c=] expressions may not appear in string2 \
1534 when translating"));
1537 if (s1->length > s2->length)
1541 /* string2 must be non-empty unless --truncate-set1 is
1542 given or string1 is empty. */
1544 if (s2->length == 0)
1545 error (EXIT_FAILURE, 0,
1546 _("when not truncating set1, string2 must be non-empty"));
1547 string2_extend (s1, s2);
1551 if (complement && s1->has_char_class
1552 && ! (s2->length == s1->length && homogeneous_spec_list (s2)))
1554 error (EXIT_FAILURE, 0,
1555 _("when translating with complemented character classes,\
1556 \nstring2 must map all characters in the domain to one"));
1559 if (s2->has_restricted_char_class)
1561 error (EXIT_FAILURE, 0,
1562 _("when translating, the only character classes that may \
1563 appear in\nstring2 are `upper' and `lower'"));
1567 /* Not translating. */
1569 if (s2->n_indefinite_repeats > 0)
1570 error (EXIT_FAILURE, 0,
1571 _("the [c*] construct may appear in string2 only \
1572 when translating"));
1577 /* Read buffers of SIZE bytes via the function READER (if READER is
1578 NULL, read from stdin) until EOF. When non-NULL, READER is either
1579 read_and_delete or read_and_xlate. After each buffer is read, it is
1580 processed and written to stdout. The buffers are processed so that
1581 multiple consecutive occurrences of the same character in the input
1582 stream are replaced by a single occurrence of that character if the
1583 character is in the squeeze set. */
1586 squeeze_filter (unsigned char *buf, size_t size, Filter reader)
1588 unsigned int char_to_squeeze = NOT_A_CHAR;
1600 nr = safe_read (0, (char *) buf, size);
1601 if (nr == SAFE_READ_ERROR)
1602 error (EXIT_FAILURE, errno, _("read error"));
1606 nr = (*reader) (buf, size, NULL);
1616 if (char_to_squeeze == NOT_A_CHAR)
1619 /* Here, by being a little tricky, we can get a significant
1620 performance increase in most cases when the input is
1621 reasonably large. Since tr will modify the input only
1622 if two consecutive (and identical) input characters are
1623 in the squeeze set, we can step by two through the data
1624 when searching for a character in the squeeze set. This
1625 means there may be a little more work in a few cases and
1626 perhaps twice as much work in the worst cases where most
1627 of the input is removed by squeezing repeats. But most
1628 uses of this functionality seem to remove less than 20-30%
1630 for (; i < nr && !in_squeeze_set[buf[i]]; i += 2)
1633 /* There is a special case when i == nr and we've just
1634 skipped a character (the last one in buf) that is in
1636 if (i == nr && in_squeeze_set[buf[i - 1]])
1640 out_len = nr - begin;
1643 char_to_squeeze = buf[i];
1644 /* We're about to output buf[begin..i]. */
1645 out_len = i - begin + 1;
1647 /* But since we stepped by 2 in the loop above,
1648 out_len may be one too large. */
1649 if (i > 0 && buf[i - 1] == char_to_squeeze)
1652 /* Advance i to the index of first character to be
1653 considered when looking for a char different from
1658 && fwrite ((char *) &buf[begin], 1, out_len, stdout) == 0)
1659 error (EXIT_FAILURE, errno, _("write error"));
1662 if (char_to_squeeze != NOT_A_CHAR)
1664 /* Advance i to index of first char != char_to_squeeze
1665 (or to nr if all the rest of the characters in this
1666 buffer are the same as char_to_squeeze). */
1667 for (; i < nr && buf[i] == char_to_squeeze; i++)
1670 char_to_squeeze = NOT_A_CHAR;
1671 /* If (i >= nr) we've squeezed the last character in this buffer.
1672 So now we have to read a new buffer and continue comparing
1673 characters against char_to_squeeze. */
1678 /* Read buffers of SIZE bytes from stdin until one is found that
1679 contains at least one character not in the delete set. Store
1680 in the array BUF, all characters from that buffer that are not
1681 in the delete set, and return the number of characters saved
1685 read_and_delete (unsigned char *buf, size_t size, Filter not_used)
1688 static int hit_eof = 0;
1690 assert (not_used == NULL);
1695 /* This enclosing do-while loop is to make sure that
1696 we don't return zero (indicating EOF) when we've
1697 just deleted all the characters in a buffer. */
1701 size_t nr = safe_read (0, (char *) buf, size);
1703 if (nr == SAFE_READ_ERROR)
1704 error (EXIT_FAILURE, errno, _("read error"));
1711 /* This first loop may be a waste of code, but gives much
1712 better performance when no characters are deleted in
1713 the beginning of a buffer. It just avoids the copying
1714 of buf[i] into buf[n_saved] when it would be a NOP. */
1716 for (i = 0; i < nr && !in_delete_set[buf[i]]; i++)
1720 for (++i; i < nr; i++)
1721 if (!in_delete_set[buf[i]])
1722 buf[n_saved++] = buf[i];
1724 while (n_saved == 0);
1729 /* Read at most SIZE bytes from stdin into the array BUF. Then
1730 perform the in-place and one-to-one mapping specified by the global
1731 array `xlate'. Return the number of characters read, or 0 upon EOF. */
1734 read_and_xlate (unsigned char *buf, size_t size, Filter not_used)
1736 size_t bytes_read = 0;
1737 static int hit_eof = 0;
1740 assert (not_used == NULL);
1745 bytes_read = safe_read (0, (char *) buf, size);
1746 if (bytes_read == SAFE_READ_ERROR)
1747 error (EXIT_FAILURE, errno, _("read error"));
1748 if (bytes_read == 0)
1754 for (i = 0; i < bytes_read; i++)
1755 buf[i] = xlate[buf[i]];
1760 /* Initialize a boolean membership set IN_SET with the character
1761 values obtained by traversing the linked list of constructs S
1762 using the function `get_next'. If COMPLEMENT_THIS_SET is
1763 nonzero the resulting set is complemented. */
1766 set_initialize (struct Spec_list *s, int complement_this_set, SET_TYPE *in_set)
1771 memset (in_set, 0, N_CHARS * sizeof (in_set[0]));
1772 s->state = BEGIN_STATE;
1773 while ((c = get_next (s, NULL)) != -1)
1775 if (complement_this_set)
1776 for (i = 0; i < N_CHARS; i++)
1777 in_set[i] = (!in_set[i]);
1781 main (int argc, char **argv)
1784 int non_option_args;
1785 struct Spec_list buf1, buf2;
1786 struct Spec_list *s1 = &buf1;
1787 struct Spec_list *s2 = &buf2;
1789 initialize_main (&argc, &argv);
1790 program_name = argv[0];
1791 setlocale (LC_ALL, "");
1792 bindtextdomain (PACKAGE, LOCALEDIR);
1793 textdomain (PACKAGE);
1795 atexit (close_stdout);
1797 while ((c = getopt_long (argc, argv, "cdst", long_options, NULL)) != -1)
1813 squeeze_repeats = 1;
1820 case_GETOPT_HELP_CHAR;
1822 case_GETOPT_VERSION_CHAR (PROGRAM_NAME, WRITTEN_BY);
1830 posix_pedantic = (getenv ("POSIXLY_CORRECT") != NULL);
1832 non_option_args = argc - optind;
1833 translating = (non_option_args == 2 && !delete);
1835 /* Change this test if it is valid to give tr no options and
1836 no args at all. POSIX doesn't specifically say anything
1837 either way, but it looks like they implied it's invalid
1838 by omission. If you want to make tr do a slow imitation
1839 of `cat' use `tr a a'. */
1840 if (non_option_args > 2)
1842 error (0, 0, _("too many arguments"));
1846 if (!delete && !squeeze_repeats && non_option_args != 2)
1847 error (EXIT_FAILURE, 0, _("two strings must be given when translating"));
1849 if (delete && squeeze_repeats && non_option_args != 2)
1850 error (EXIT_FAILURE, 0, _("two strings must be given when both \
1851 deleting and squeezing repeats"));
1853 /* If --delete is given without --squeeze-repeats, then
1854 only one string argument may be specified. But POSIX
1855 says to ignore any string2 in this case, so if POSIXLY_CORRECT
1856 is set, pretend we never saw string2. But I think
1857 this deserves a fatal error, so that's the default. */
1858 if ((delete && !squeeze_repeats) && non_option_args != 1)
1860 if (posix_pedantic && non_option_args == 2)
1863 error (EXIT_FAILURE, 0,
1864 _("only one string may be given when deleting \
1865 without squeezing repeats"));
1868 if (squeeze_repeats && non_option_args == 0)
1869 error (EXIT_FAILURE, 0,
1870 _("at least one string must be given when squeezing repeats"));
1873 if (parse_str ((unsigned char *) argv[optind], s1))
1874 exit (EXIT_FAILURE);
1876 if (non_option_args == 2)
1879 if (parse_str ((unsigned char *) argv[optind + 1], s2))
1880 exit (EXIT_FAILURE);
1887 /* Use binary I/O, since `tr' is sometimes used to transliterate
1888 non-printable characters, or characters which are stripped away
1889 by text-mode reads (like CR and ^Z). */
1890 SET_BINARY2 (STDIN_FILENO, STDOUT_FILENO);
1892 if (squeeze_repeats && non_option_args == 1)
1894 set_initialize (s1, complement, in_squeeze_set);
1895 squeeze_filter (io_buf, sizeof io_buf, NULL);
1897 else if (delete && non_option_args == 1)
1901 set_initialize (s1, complement, in_delete_set);
1904 nr = read_and_delete (io_buf, sizeof io_buf, NULL);
1905 if (nr > 0 && fwrite ((char *) io_buf, 1, nr, stdout) == 0)
1906 error (EXIT_FAILURE, errno, _("write error"));
1910 else if (squeeze_repeats && delete && non_option_args == 2)
1912 set_initialize (s1, complement, in_delete_set);
1913 set_initialize (s2, 0, in_squeeze_set);
1914 squeeze_filter (io_buf, sizeof io_buf, read_and_delete);
1916 else if (translating)
1921 SET_TYPE *in_s1 = in_delete_set;
1923 set_initialize (s1, 0, in_s1);
1924 s2->state = BEGIN_STATE;
1925 for (i = 0; i < N_CHARS; i++)
1927 for (i = 0; i < N_CHARS; i++)
1931 int ch = get_next (s2, NULL);
1932 assert (ch != -1 || truncate_set1);
1935 /* This will happen when tr is invoked like e.g.
1936 tr -cs A-Za-z0-9 '\012'. */
1942 assert (get_next (s2, NULL) == -1 || truncate_set1);
1948 enum Upper_Lower_class class_s1;
1949 enum Upper_Lower_class class_s2;
1951 for (i = 0; i < N_CHARS; i++)
1953 s1->state = BEGIN_STATE;
1954 s2->state = BEGIN_STATE;
1957 c1 = get_next (s1, &class_s1);
1958 c2 = get_next (s2, &class_s2);
1959 if (!class_ok[(int) class_s1][(int) class_s2])
1960 error (EXIT_FAILURE, 0,
1961 _("misaligned [:upper:] and/or [:lower:] construct"));
1963 if (class_s1 == UL_LOWER && class_s2 == UL_UPPER)
1965 for (i = 0; i < N_CHARS; i++)
1967 xlate[i] = toupper (i);
1969 else if (class_s1 == UL_UPPER && class_s2 == UL_LOWER)
1971 for (i = 0; i < N_CHARS; i++)
1973 xlate[i] = tolower (i);
1975 else if ((class_s1 == UL_LOWER && class_s2 == UL_LOWER)
1976 || (class_s1 == UL_UPPER && class_s2 == UL_UPPER))
1978 /* By default, GNU tr permits the identity mappings: from
1979 [:upper:] to [:upper:] and [:lower:] to [:lower:]. But
1980 when POSIXLY_CORRECT is set, those evoke diagnostics. */
1983 error (EXIT_FAILURE, 0,
1985 invalid identity mapping; when translating, any [:lower:] or [:upper:]\n\
1986 construct in string1 must be aligned with a corresponding construct\n\
1987 ([:upper:] or [:lower:], respectively) in string2"));
1992 /* The following should have been checked by validate... */
1993 if (c1 == -1 || c2 == -1)
1998 assert (c1 == -1 || truncate_set1);
2000 if (squeeze_repeats)
2002 set_initialize (s2, 0, in_squeeze_set);
2003 squeeze_filter (io_buf, sizeof io_buf, read_and_xlate);
2011 bytes_read = read_and_xlate (io_buf, sizeof io_buf, NULL);
2013 && fwrite ((char *) io_buf, 1, bytes_read, stdout) == 0)
2014 error (EXIT_FAILURE, errno, _("write error"));
2016 while (bytes_read > 0);
2020 if (close (STDIN_FILENO) != 0)
2021 error (EXIT_FAILURE, errno, _("standard input"));
2023 exit (EXIT_SUCCESS);