Skip to content

Commit 3984ac9

Browse files
ChenPi11ChenPi11
committed
Optimize the project structure.
The compile may be a little slower, but the project structure is more reasonable.
1 parent 5cb112a commit 3984ac9

File tree

11 files changed

+387
-231
lines changed

11 files changed

+387
-231
lines changed

‎lib/all_encodings.h‎

Lines changed: 39 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,39 @@
1+
/*
2+
* Copyright (C) 2024 The C++ Plus Project.
3+
* This file is part of the cppp-reiconv library.
4+
*
5+
* The cppp-reiconv library is free software; you can redistribute it
6+
* and/or modify it under the terms of the GNU Lesser General Public
7+
* License as published by the Free Software Foundation; either version 3
8+
* of the License, or (at your option) any later version.
9+
*
10+
* The cppp-reiconv library is distributed in the hope that it will be
11+
* useful, but WITHOUT ANY WARRANTY; without even the implied warranty of
12+
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13+
* Lesser General Public License for more details.
14+
*
15+
* You should have received a copy of the GNU Lesser General Public
16+
* License along with the cppp-reiconv library; see the file LICENSE.
17+
* If not, see <https://www.gnu.org/licenses/>.
18+
*/
19+
20+
#ifndef _ALL_ENCODINGS_H_
21+
#define _ALL_ENCODINGS_H_
22+
23+
#include "converters.h"
24+
#include "encoding.h"
25+
26+
#define DEFENCODING(xxx_names, xxx, xxx_index, xxx_ifuncs1, xxx_ifuncs2, xxx_ofuncs1, xxx_ofuncs2) \
27+
{xxx_ifuncs1, xxx_ifuncs2, xxx_ofuncs1, xxx_ofuncs2},
28+
#define DEFCODEPAGE(codepage, xxx)
29+
#define DEFINDEX(alias, name)
30+
31+
static struct encoding const all_encodings[] = {
32+
#include "encodings.h.snippet"
33+
};
34+
35+
#undef DEFINDEX
36+
#undef DEFENCODING
37+
#undef DEFCODEPAGE
38+
39+
#endif /* _ALL_ENCODINGS_H_ */

‎lib/converters.h‎

Lines changed: 78 additions & 161 deletions
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,6 @@
11
/*
22
* Copyright (C) 1999-2002, 2004-2011, 2016, 2022-2023 Free Software Foundation, Inc.
3+
* Copyright (C) 2024 The C++ Project.
34
* This file is part of the cppp-reiconv library.
45
*
56
* The cppp-reiconv library is free software; you can redistribute it
@@ -17,146 +18,44 @@
1718
* If not, see <https://www.gnu.org/licenses/>.
1819
*/
1920

20-
/* This file defines all the converters. */
21+
#ifndef _CONVERTERS_H_
22+
#define _CONVERTERS_H_
2123

24+
#include "reiconv_defines.h"
2225

23-
/* Our own notion of wide character, as UCS-4, according to ISO-10646-1. */
24-
typedef unsigned int ucs4_t;
25-
26-
/* State used by a conversion. 0 denotes the initial state. */
27-
typedef unsigned int state_t;
28-
29-
/* iconv_t is an opaque type. This is the real iconv_t type. */
30-
typedef struct conv_struct * conv_t;
31-
32-
/*
33-
* Data type for conversion multibyte -> unicode
34-
*/
35-
struct mbtowc_funcs {
36-
int (*xxx_mbtowc) (conv_t conv, ucs4_t *pwc, unsigned char const *s, size_t n);
37-
/*
38-
* int xxx_mbtowc (conv_t conv, ucs4_t *pwc, unsigned char const *s, size_t n)
39-
* converts the byte sequence starting at s to a wide character. Up to n bytes
40-
* are available at s. n is >= 1.
41-
* Result is number of bytes consumed (if a wide character was read),
42-
* or -1 if invalid, or -2 if n too small,
43-
* or RET_SHIFT_ILSEQ(number of bytes consumed) if invalid input after a shift
44-
* sequence was read,
45-
* or RET_TOOFEW(number of bytes consumed) if only a shift sequence was read.
46-
*/
47-
int (*xxx_flushwc) (conv_t conv, ucs4_t *pwc);
48-
/*
49-
* int xxx_flushwc (conv_t conv, ucs4_t *pwc)
50-
* returns to the initial state and stores the pending wide character, if any.
51-
* Result is 1 (if a wide character was read) or 0 if none was pending.
52-
*/
53-
};
54-
55-
/* Return code if invalid input after a shift sequence of n bytes was read.
56-
(xxx_mbtowc) */
57-
#define RET_SHIFT_ILSEQ(n) (-1-2*(n))
58-
/* Return code if invalid. (xxx_mbtowc) */
59-
#define RET_ILSEQ RET_SHIFT_ILSEQ(0)
60-
/* Return code if only a shift sequence of n bytes was read. (xxx_mbtowc) */
61-
#define RET_TOOFEW(n) (-2-2*(n))
62-
/* Retrieve the n from the encoded RET_... value. */
63-
#define DECODE_SHIFT_ILSEQ(r) ((unsigned int)(RET_SHIFT_ILSEQ(0) - (r)) / 2)
64-
#define DECODE_TOOFEW(r) ((unsigned int)(RET_TOOFEW(0) - (r)) / 2)
65-
/* Maximum value of n that may be used as argument to RET_SHIFT_ILSEQ or RET_TOOFEW. */
66-
#define RET_COUNT_MAX ((INT_MAX / 2) - 1)
67-
68-
/*
69-
* Data type for conversion unicode -> multibyte
70-
*/
71-
struct wctomb_funcs {
72-
int (*xxx_wctomb) (conv_t conv, unsigned char *r, ucs4_t wc, size_t n);
73-
/*
74-
* int xxx_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, size_t n)
75-
* converts the wide character wc to the character set xxx, and stores the
76-
* result beginning at r. Up to n bytes may be written at r. n is >= 1.
77-
* Result is number of bytes written, or -1 if invalid, or -2 if n too small.
78-
*/
79-
int (*xxx_reset) (conv_t conv, unsigned char *r, size_t n);
80-
/*
81-
* int xxx_reset (conv_t conv, unsigned char *r, size_t n)
82-
* stores a shift sequences returning to the initial state beginning at r.
83-
* Up to n bytes may be written at r. n is >= 0.
84-
* Result is number of bytes written, or -2 if n too small.
85-
*/
86-
};
87-
88-
/* Return code if invalid. (xxx_wctomb) */
89-
#define RET_ILUNI -1
90-
/* Return code if output buffer is too small. (xxx_wctomb, xxx_reset) */
91-
#define RET_TOOSMALL -2
92-
93-
/*
94-
* Contents of a conversion descriptor.
95-
*/
96-
struct conv_struct {
97-
struct loop_funcs lfuncs;
98-
/* Input (conversion multibyte -> unicode) */
99-
int iindex;
100-
struct mbtowc_funcs ifuncs;
101-
state_t istate;
102-
/* Output (conversion unicode -> multibyte) */
103-
int oindex;
104-
struct wctomb_funcs ofuncs;
105-
state_t ostate;
106-
/* Operation flags */
107-
int discard_ilseq;
108-
};
109-
110-
/*
111-
* Include all the converters.
112-
*/
26+
// Include all the converters.
11327

11428
#pragma region ASCII
11529
#include "converters/ascii.h"
11630
#pragma endregion
11731

11832
#pragma region General multi-byte encodings
119-
#include "converters/utf8.h"
33+
#include "converters/c99.h"
34+
#include "converters/java.h"
12035
#include "converters/ucs2.h"
12136
#include "converters/ucs2be.h"
37+
#include "converters/ucs2internal.h"
12238
#include "converters/ucs2le.h"
39+
#include "converters/ucs2swapped.h"
12340
#include "converters/ucs4.h"
12441
#include "converters/ucs4be.h"
42+
#include "converters/ucs4internal.h"
12543
#include "converters/ucs4le.h"
44+
#include "converters/ucs4swapped.h"
12645
#include "converters/utf16.h"
12746
#include "converters/utf16be.h"
12847
#include "converters/utf16le.h"
12948
#include "converters/utf32.h"
13049
#include "converters/utf32be.h"
13150
#include "converters/utf32le.h"
13251
#include "converters/utf7.h"
133-
#include "converters/ucs2internal.h"
134-
#include "converters/ucs2swapped.h"
135-
#include "converters/ucs4internal.h"
136-
#include "converters/ucs4swapped.h"
137-
#include "converters/c99.h"
138-
#include "converters/java.h"
52+
#include "converters/utf8.h"
13953
#pragma endregion
14054

14155
#pragma region 8-bit encodings
142-
#include "converters/iso8859_1.h"
143-
#include "converters/iso8859_2.h"
144-
#include "converters/iso8859_3.h"
145-
#include "converters/iso8859_4.h"
146-
#include "converters/iso8859_5.h"
147-
#include "converters/iso8859_6.h"
148-
#include "converters/iso8859_7.h"
149-
#include "converters/iso8859_8.h"
150-
#include "converters/iso8859_9.h"
151-
#include "converters/iso8859_10.h"
152-
#include "converters/iso8859_11.h"
153-
#include "converters/iso8859_13.h"
154-
#include "converters/iso8859_14.h"
155-
#include "converters/iso8859_15.h"
156-
#include "converters/iso8859_16.h"
157-
#include "converters/koi8_r.h"
158-
#include "converters/koi8_u.h"
159-
#include "converters/koi8_ru.h"
56+
#include "converters/armscii_8.h"
57+
#include "converters/cp1131.h"
58+
#include "converters/cp1133.h"
16059
#include "converters/cp1250.h"
16160
#include "converters/cp1251.h"
16261
#include "converters/cp1252.h"
@@ -169,39 +68,55 @@ struct conv_struct {
16968
#include "converters/cp850.h"
17069
#include "converters/cp862.h"
17170
#include "converters/cp866.h"
172-
#include "converters/cp1131.h"
173-
#include "converters/mac_roman.h"
71+
#include "converters/cp874.h"
72+
#include "converters/georgian_academy.h"
73+
#include "converters/georgian_ps.h"
74+
#include "converters/hp_roman8.h"
75+
#include "converters/iso8859_1.h"
76+
#include "converters/iso8859_10.h"
77+
#include "converters/iso8859_11.h"
78+
#include "converters/iso8859_13.h"
79+
#include "converters/iso8859_14.h"
80+
#include "converters/iso8859_15.h"
81+
#include "converters/iso8859_16.h"
82+
#include "converters/iso8859_2.h"
83+
#include "converters/iso8859_3.h"
84+
#include "converters/iso8859_4.h"
85+
#include "converters/iso8859_5.h"
86+
#include "converters/iso8859_6.h"
87+
#include "converters/iso8859_7.h"
88+
#include "converters/iso8859_8.h"
89+
#include "converters/iso8859_9.h"
90+
#include "converters/koi8_r.h"
91+
#include "converters/koi8_ru.h"
92+
#include "converters/koi8_t.h"
93+
#include "converters/koi8_u.h"
94+
#include "converters/mac_arabic.h"
17495
#include "converters/mac_centraleurope.h"
175-
#include "converters/mac_iceland.h"
17696
#include "converters/mac_croatian.h"
177-
#include "converters/mac_romania.h"
17897
#include "converters/mac_cyrillic.h"
179-
#include "converters/mac_ukraine.h"
18098
#include "converters/mac_greek.h"
181-
#include "converters/mac_turkish.h"
18299
#include "converters/mac_hebrew.h"
183-
#include "converters/mac_arabic.h"
100+
#include "converters/mac_iceland.h"
101+
#include "converters/mac_roman.h"
102+
#include "converters/mac_romania.h"
184103
#include "converters/mac_thai.h"
185-
#include "converters/hp_roman8.h"
104+
#include "converters/mac_turkish.h"
105+
#include "converters/mac_ukraine.h"
106+
#include "converters/mulelao.h"
186107
#include "converters/nextstep.h"
187-
#include "converters/armscii_8.h"
188-
#include "converters/georgian_academy.h"
189-
#include "converters/georgian_ps.h"
190-
#include "converters/koi8_t.h"
191108
#include "converters/pt154.h"
192109
#include "converters/rk1048.h"
193-
#include "converters/mulelao.h"
194-
#include "converters/cp1133.h"
110+
#include "converters/tcvn.h"
195111
#include "converters/tis620.h"
196-
#include "converters/cp874.h"
197112
#include "converters/viscii.h"
198-
#include "converters/tcvn.h"
199113
#pragma endregion
200114

201115
/* [CCS = coded character set] [CJKV.INF chapter 3] */
202-
typedef struct {
203-
unsigned short indx; /* Index into big table. */
204-
unsigned short used; /* Bitmask of used entries. */
116+
typedef struct
117+
{
118+
unsigned short indx; /* Index into big table. */
119+
unsigned short used; /* Bitmask of used entries. */
205120
} Summary16;
206121

207122
#pragma region CJK character sets
@@ -210,64 +125,65 @@ typedef struct {
210125
#include "converters/jisx0208.h"
211126
#include "converters/jisx0212.h"
212127

213-
#include "converters/iso646_cn.h"
214128
#include "converters/gb2312.h"
129+
#include "converters/iso646_cn.h"
215130
#include "converters/isoir165.h"
216131
/* #include "converters/gb12345.h" */
217-
#include "converters/gbk.h"
218-
#include "converters/cns11643.h"
219132
#include "converters/big5.h"
133+
#include "converters/cns11643.h"
134+
#include "converters/gbk.h"
220135

221-
#include "converters/ksc5601.h"
222136
#include "converters/johab_hangul.h"
137+
#include "converters/ksc5601.h"
223138
#pragma endregion
224139

225140
/* [CES = character encoding scheme] [CJKV.INF chapter 4] */
226141
#pragma region CJK encodings
227-
#include "converters/euc_jp.h"
228-
#include "converters/sjis.h"
229142
#include "converters/cp932.h"
143+
#include "converters/euc_jp.h"
230144
#include "converters/iso2022_jp.h"
231145
#include "converters/iso2022_jp1.h"
232146
#include "converters/iso2022_jp2.h"
233147
#include "converters/iso2022_jpms.h"
148+
#include "converters/sjis.h"
234149

235-
#include "converters/euc_cn.h"
150+
#include "converters/big5hkscs1999.h"
151+
#include "converters/big5hkscs2001.h"
152+
#include "converters/big5hkscs2004.h"
153+
#include "converters/big5hkscs2008.h"
154+
#include "converters/ces_big5.h"
236155
#include "converters/ces_gbk.h"
237156
#include "converters/cp936.h"
157+
#include "converters/cp950.h"
158+
#include "converters/euc_cn.h"
159+
#include "converters/euc_tw.h"
238160
#include "converters/gb18030_2005.h"
239161
#include "converters/gb18030_2022.h"
162+
#include "converters/hz.h"
240163
#include "converters/iso2022_cn.h"
241164
#include "converters/iso2022_cnext.h"
242-
#include "converters/hz.h"
243-
#include "converters/euc_tw.h"
244-
#include "converters/ces_big5.h"
245-
#include "converters/cp950.h"
246-
#include "converters/big5hkscs1999.h"
247-
#include "converters/big5hkscs2001.h"
248-
#include "converters/big5hkscs2004.h"
249-
#include "converters/big5hkscs2008.h"
250165

251-
#include "converters/euc_kr.h"
252166
#include "converters/cp949.h"
253-
#include "converters/johab.h"
167+
#include "converters/euc_kr.h"
254168
#include "converters/iso2022_kr.h"
169+
#include "converters/johab.h"
255170

256-
#include "converters/dec_kanji.h"
257171
#include "converters/dec_hanyu.h"
172+
#include "converters/dec_kanji.h"
258173
#pragma endregion
259174

260175
#pragma region Encodings used by system dependent locales
261-
#include "converters/cp856.h"
262-
#include "converters/cp922.h"
263-
#include "converters/cp943.h"
264176
#include "converters/cp1046.h"
265177
#include "converters/cp1124.h"
266178
#include "converters/cp1129.h"
267179
#include "converters/cp1161.h"
268180
#include "converters/cp1162.h"
269181
#include "converters/cp1163.h"
182+
#include "converters/cp856.h"
183+
#include "converters/cp922.h"
184+
#include "converters/cp943.h"
270185

186+
#include "converters/cp1125.h"
271187
#include "converters/cp437.h"
272188
#include "converters/cp737.h"
273189
#include "converters/cp775.h"
@@ -282,7 +198,6 @@ typedef struct {
282198
#include "converters/cp864.h"
283199
#include "converters/cp865.h"
284200
#include "converters/cp869.h"
285-
#include "converters/cp1125.h"
286201
#pragma endregion
287202

288203
#define DEDUPLICATE_TABLES 1
@@ -344,11 +259,13 @@ typedef struct {
344259
#pragma endregion
345260

346261
#pragma region Extra encodings
262+
#include "converters/atarist.h"
263+
#include "converters/big5_2003.h"
347264
#include "converters/euc_jisx0213.h"
348-
#include "converters/shift_jisx0213.h"
349265
#include "converters/iso2022_jp3.h"
350-
#include "converters/big5_2003.h"
351-
#include "converters/tds565.h"
352-
#include "converters/atarist.h"
353266
#include "converters/riscos1.h"
267+
#include "converters/shift_jisx0213.h"
268+
#include "converters/tds565.h"
354269
#pragma endregion
270+
271+
#endif /* _CONVERTERS_H_ */

0 commit comments

Comments
 (0)