xref: /freebsd/usr.bin/sort/bwstring.h (revision 3c4ba5f55438f7afd4f4b0b56f88f2bb505fd6a6)
1 /*	$FreeBSD$	*/
2 
3 /*-
4  * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
5  *
6  * Copyright (C) 2009 Gabor Kovesdan <gabor@FreeBSD.org>
7  * Copyright (C) 2012 Oleg Moskalenko <mom040267@gmail.com>
8  * All rights reserved.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29  * SUCH DAMAGE.
30  */
31 
32 #if !defined(__BWSTRING_H__)
33 #define	__BWSTRING_H__
34 
35 #include <stdbool.h>
36 #include <stdio.h>
37 #include <errno.h>
38 #include <sysexits.h>
39 #include <wchar.h>
40 
41 #include "sort.h"
42 #include "mem.h"
43 
44 extern bool byte_sort;
45 
46 /* wchar_t is of 4 bytes: */
47 #define	SIZEOF_WCHAR_STRING(LEN) ((LEN)*sizeof(wchar_t))
48 
49 struct wstr {
50 	size_t len;
51 	wchar_t str[];
52 };
53 
54 struct cstr {
55 	size_t len;
56 	char str[];
57 };
58 
59 /*
60  * Binary "wide" string
61  */
62 struct bwstring
63 {
64 	union {
65 		struct wstr wdata;
66 		struct cstr cdata;
67 	};
68 };
69 
70 typedef void *bwstring_iterator;
71 
72 #define	BWSLEN(s) ((mb_cur_max == 1) ? (s)->cdata.len : (s)->wdata.len)
73 struct bwstring *bwsalloc(size_t sz);
74 
75 size_t bwsrawlen(const struct bwstring *bws);
76 const void* bwsrawdata(const struct bwstring *bws);
77 void bws_setlen(struct bwstring *bws, size_t newlen);
78 size_t bws_memsize(const struct bwstring *bws);
79 double bwstod(struct bwstring *s0, bool *empty);
80 int bws_month_score(const struct bwstring *s0);
81 
82 struct bwstring *ignore_leading_blanks(struct bwstring *str);
83 struct bwstring *ignore_nonprinting(struct bwstring *str);
84 struct bwstring *dictionary_order(struct bwstring *str);
85 struct bwstring *ignore_case(struct bwstring *str);
86 
87 void bwsprintf(FILE*, struct bwstring*, const char *prefix, const char *suffix);
88 void bws_disorder_warnx(struct bwstring *s, const char *fn, size_t pos);
89 
90 struct bwstring *bwsdup(const struct bwstring *s);
91 struct bwstring *bwssbdup(const wchar_t *str, size_t size);
92 struct bwstring *bwscsbdup(const unsigned char *str, size_t size);
93 void bwsfree(const struct bwstring *s);
94 struct bwstring *bwsnocpy(struct bwstring *dst, const struct bwstring *src, size_t offset, size_t size);
95 int bwscmp(const struct bwstring *bws1, const struct bwstring *bws2, size_t offset);
96 int bwsncmp(const struct bwstring *bws1, const struct bwstring *bws2, size_t offset, size_t len);
97 int bwscoll(const struct bwstring *bws1, const struct bwstring *bws2, size_t offset);
98 size_t bwsfwrite(struct bwstring *bws, FILE *f, bool zero_ended);
99 
100 static inline bwstring_iterator
101 bws_begin(struct bwstring *bws)
102 {
103 
104 	return ((bwstring_iterator)bws->wdata.str);
105 }
106 
107 static inline bwstring_iterator
108 bws_end(struct bwstring *bws)
109 {
110 
111 	return ((mb_cur_max == 1) ?
112 	    (bwstring_iterator) (bws->cdata.str + bws->cdata.len) :
113 	    (bwstring_iterator) (bws->wdata.str + bws->wdata.len));
114 }
115 
116 static inline bwstring_iterator
117 bws_iterator_inc(bwstring_iterator iter, size_t pos)
118 {
119 
120 	if (mb_cur_max == 1)
121 		return ((unsigned char *) iter) + pos;
122 	else
123 		return ((wchar_t*) iter) + pos;
124 }
125 
126 static inline wchar_t
127 bws_get_iter_value(bwstring_iterator iter)
128 {
129 
130 	if (mb_cur_max == 1)
131 		return *((unsigned char *) iter);
132 	else
133 		return *((wchar_t*) iter);
134 }
135 
136 int
137 bws_iterator_cmp(bwstring_iterator iter1, bwstring_iterator iter2, size_t len);
138 
139 #define	BWS_GET(bws, pos) ((mb_cur_max == 1) ? (bws->cdata.str[(pos)]) : bws->wdata.str[(pos)])
140 
141 void initialise_months(void);
142 
143 #endif /* __BWSTRING_H__ */
144