xref: /freebsd/usr.bin/sort/bwstring.h (revision 71ec05a21257e159f40d54e26ad0011bb19b5134)
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 /*
50  * Binary "wide" string
51  */
52 struct bwstring
53 {
54 	size_t				len;
55 	union
56 	{
57 		wchar_t		wstr[0];
58 		unsigned char	cstr[0];
59 	}				data;
60 };
61 
62 struct reader_buffer
63 {
64 	wchar_t			*fgetwln_z_buffer;
65 	size_t			 fgetwln_z_buffer_size;
66 };
67 
68 typedef void *bwstring_iterator;
69 
70 #define	BWSLEN(s) ((s)->len)
71 
72 struct bwstring *bwsalloc(size_t sz);
73 
74 size_t bwsrawlen(const struct bwstring *bws);
75 const void* bwsrawdata(const struct bwstring *bws);
76 void bws_setlen(struct bwstring *bws, size_t newlen);
77 size_t bws_memsize(const struct bwstring *bws);
78 double bwstod(struct bwstring *s0, bool *empty);
79 int bws_month_score(const struct bwstring *s0);
80 
81 struct bwstring *ignore_leading_blanks(struct bwstring *str);
82 struct bwstring *ignore_nonprinting(struct bwstring *str);
83 struct bwstring *dictionary_order(struct bwstring *str);
84 struct bwstring *ignore_case(struct bwstring *str);
85 
86 void bwsprintf(FILE*, struct bwstring*, const char *prefix, const char *suffix);
87 void bws_disorder_warnx(struct bwstring *s, const char *fn, size_t pos);
88 
89 struct bwstring *bwsdup(const struct bwstring *s);
90 struct bwstring *bwssbdup(const wchar_t *str, size_t size);
91 struct bwstring *bwscsbdup(const unsigned char *str, size_t size);
92 void bwsfree(const struct bwstring *s);
93 size_t bwscpy(struct bwstring *dst, const struct bwstring *src);
94 struct bwstring *bwsncpy(struct bwstring *dst, const struct bwstring *src, size_t size);
95 struct bwstring *bwsnocpy(struct bwstring *dst, const struct bwstring *src, size_t offset, size_t size);
96 int bwscmp(const struct bwstring *bws1, const struct bwstring *bws2, size_t offset);
97 int bwsncmp(const struct bwstring *bws1, const struct bwstring *bws2, size_t offset, size_t len);
98 int bwscoll(const struct bwstring *bws1, const struct bwstring *bws2, size_t offset);
99 size_t bwsfwrite(struct bwstring *bws, FILE *f, bool zero_ended);
100 struct bwstring *bwsfgetln(FILE *file, size_t *len, bool zero_ended, struct reader_buffer *rb);
101 
102 static inline bwstring_iterator
103 bws_begin(struct bwstring *bws)
104 {
105 
106 	return (bwstring_iterator) (&(bws->data));
107 }
108 
109 static inline bwstring_iterator
110 bws_end(struct bwstring *bws)
111 {
112 
113 	return ((mb_cur_max == 1) ?
114 	    (bwstring_iterator) (bws->data.cstr + bws->len) :
115 	    (bwstring_iterator) (bws->data.wstr + bws->len));
116 }
117 
118 static inline bwstring_iterator
119 bws_iterator_inc(bwstring_iterator iter, size_t pos)
120 {
121 
122 	if (mb_cur_max == 1)
123 		return ((unsigned char *) iter) + pos;
124 	else
125 		return ((wchar_t*) iter) + pos;
126 }
127 
128 static inline wchar_t
129 bws_get_iter_value(bwstring_iterator iter)
130 {
131 
132 	if (mb_cur_max == 1)
133 		return *((unsigned char *) iter);
134 	else
135 		return *((wchar_t*) iter);
136 }
137 
138 int
139 bws_iterator_cmp(bwstring_iterator iter1, bwstring_iterator iter2, size_t len);
140 
141 #define	BWS_GET(bws, pos) ((mb_cur_max == 1) ? ((bws)->data.cstr[(pos)]) : (bws)->data.wstr[(pos)])
142 
143 void initialise_months(void);
144 
145 #endif /* __BWSTRING_H__ */
146