xref: /freebsd/usr.bin/sort/bwstring.h (revision 42249ef2)
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 "mem.h"
42 
43 extern bool byte_sort;
44 
45 /* wchar_t is of 4 bytes: */
46 #define	SIZEOF_WCHAR_STRING(LEN) ((LEN)*sizeof(wchar_t))
47 
48 /*
49  * Binary "wide" string
50  */
51 struct bwstring
52 {
53 	size_t				len;
54 	union
55 	{
56 		wchar_t		wstr[0];
57 		unsigned char	cstr[0];
58 	}				data;
59 };
60 
61 struct reader_buffer
62 {
63 	wchar_t			*fgetwln_z_buffer;
64 	size_t			 fgetwln_z_buffer_size;
65 };
66 
67 typedef void *bwstring_iterator;
68 
69 #define	BWSLEN(s) ((s)->len)
70 
71 struct bwstring *bwsalloc(size_t sz);
72 
73 size_t bwsrawlen(const struct bwstring *bws);
74 const void* bwsrawdata(const struct bwstring *bws);
75 void bws_setlen(struct bwstring *bws, size_t newlen);
76 size_t bws_memsize(const struct bwstring *bws);
77 double bwstod(struct bwstring *s0, bool *empty);
78 int bws_month_score(const struct bwstring *s0);
79 
80 struct bwstring *ignore_leading_blanks(struct bwstring *str);
81 struct bwstring *ignore_nonprinting(struct bwstring *str);
82 struct bwstring *dictionary_order(struct bwstring *str);
83 struct bwstring *ignore_case(struct bwstring *str);
84 
85 void bwsprintf(FILE*, struct bwstring*, const char *prefix, const char *suffix);
86 void bws_disorder_warnx(struct bwstring *s, const char *fn, size_t pos);
87 
88 struct bwstring *bwsdup(const struct bwstring *s);
89 struct bwstring *bwssbdup(const wchar_t *str, size_t size);
90 struct bwstring *bwscsbdup(const unsigned char *str, size_t size);
91 void bwsfree(const struct bwstring *s);
92 size_t bwscpy(struct bwstring *dst, const struct bwstring *src);
93 struct bwstring *bwsncpy(struct bwstring *dst, const struct bwstring *src, size_t size);
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 struct bwstring *bwsfgetln(FILE *file, size_t *len, bool zero_ended, struct reader_buffer *rb);
100 
101 static inline bwstring_iterator
102 bws_begin(struct bwstring *bws)
103 {
104 
105 	return (bwstring_iterator) (&(bws->data));
106 }
107 
108 static inline bwstring_iterator
109 bws_end(struct bwstring *bws)
110 {
111 
112 	return ((MB_CUR_MAX == 1) ?
113 	    (bwstring_iterator) (bws->data.cstr + bws->len) :
114 	    (bwstring_iterator) (bws->data.wstr + bws->len));
115 }
116 
117 static inline bwstring_iterator
118 bws_iterator_inc(bwstring_iterator iter, size_t pos)
119 {
120 
121 	if (MB_CUR_MAX == 1)
122 		return ((unsigned char *) iter) + pos;
123 	else
124 		return ((wchar_t*) iter) + pos;
125 }
126 
127 static inline wchar_t
128 bws_get_iter_value(bwstring_iterator iter)
129 {
130 
131 	if (MB_CUR_MAX == 1)
132 		return *((unsigned char *) iter);
133 	else
134 		return *((wchar_t*) iter);
135 }
136 
137 int
138 bws_iterator_cmp(bwstring_iterator iter1, bwstring_iterator iter2, size_t len);
139 
140 #define	BWS_GET(bws, pos) ((MB_CUR_MAX == 1) ? ((bws)->data.cstr[(pos)]) : (bws)->data.wstr[(pos)])
141 
142 void initialise_months(void);
143 
144 #endif /* __BWSTRING_H__ */
145