xref: /aoo41x/main/sal/textenc/textenc.cxx (revision 21651e5a)
1*87d2adbcSAndrew Rist /**************************************************************
2cdf0e10cSrcweir  *
3*87d2adbcSAndrew Rist  * Licensed to the Apache Software Foundation (ASF) under one
4*87d2adbcSAndrew Rist  * or more contributor license agreements.  See the NOTICE file
5*87d2adbcSAndrew Rist  * distributed with this work for additional information
6*87d2adbcSAndrew Rist  * regarding copyright ownership.  The ASF licenses this file
7*87d2adbcSAndrew Rist  * to you under the Apache License, Version 2.0 (the
8*87d2adbcSAndrew Rist  * "License"); you may not use this file except in compliance
9*87d2adbcSAndrew Rist  * with the License.  You may obtain a copy of the License at
10*87d2adbcSAndrew Rist  *
11*87d2adbcSAndrew Rist  *   http://www.apache.org/licenses/LICENSE-2.0
12*87d2adbcSAndrew Rist  *
13*87d2adbcSAndrew Rist  * Unless required by applicable law or agreed to in writing,
14*87d2adbcSAndrew Rist  * software distributed under the License is distributed on an
15*87d2adbcSAndrew Rist  * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
16*87d2adbcSAndrew Rist  * KIND, either express or implied.  See the License for the
17*87d2adbcSAndrew Rist  * specific language governing permissions and limitations
18*87d2adbcSAndrew Rist  * under the License.
19*87d2adbcSAndrew Rist  *
20*87d2adbcSAndrew Rist  *************************************************************/
21*87d2adbcSAndrew Rist 
22*87d2adbcSAndrew Rist 
23cdf0e10cSrcweir 
24cdf0e10cSrcweir // MARKER(update_precomp.py): autogen include statement, do not remove
25cdf0e10cSrcweir #include "precompiled_sal.hxx"
26cdf0e10cSrcweir 
27cdf0e10cSrcweir /* TODO! This file should not be called textenc.c, because it is not the
28cdf0e10cSrcweir    implementation of rtl/textenc.h.  Rather, it should be called
29cdf0e10cSrcweir    gettextencodingdata.c. */
30cdf0e10cSrcweir #include "gettextencodingdata.h"
31cdf0e10cSrcweir #include "tenchelp.h"
32cdf0e10cSrcweir #include "rtl/textenc.h"
33cdf0e10cSrcweir 
34cdf0e10cSrcweir #ifndef INCLUDED_STDDEF_H
35cdf0e10cSrcweir #include <stddef.h>
36cdf0e10cSrcweir #define INCLUDED_STDDEF_H
37cdf0e10cSrcweir #endif
38cdf0e10cSrcweir 
39cdf0e10cSrcweir #define NOTABUNI_START 0xFF
40cdf0e10cSrcweir #define NOTABUNI_END 0x00
41cdf0e10cSrcweir 
42cdf0e10cSrcweir #define NOTABCHAR_START 0xFFFF
43cdf0e10cSrcweir #define NOTABCHAR_END 0x0000
44cdf0e10cSrcweir 
45cdf0e10cSrcweir #define SAME8090UNI_START 0x80
46cdf0e10cSrcweir #define SAME8090UNI_END 0x9F
47cdf0e10cSrcweir static sal_uInt16 const aImpl8090SameToUniTab[SAME8090UNI_END
48cdf0e10cSrcweir                                                   - SAME8090UNI_START
49cdf0e10cSrcweir                                                   + 1]
50cdf0e10cSrcweir = { 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, /* 0x80 */
51cdf0e10cSrcweir     0x0088, 0x0089, 0x008A, 0x008B, 0x008C, 0x008D, 0x008E, 0x008F,
52cdf0e10cSrcweir     0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, /* 0x90 */
53cdf0e10cSrcweir     0x0098, 0x0099, 0x009A, 0x009B, 0x009C, 0x009D, 0x009E, 0x009F };
54cdf0e10cSrcweir 
55cdf0e10cSrcweir #define SAME8090CHAR_START 0x0080
56cdf0e10cSrcweir #define SAME8090CHAR_END 0x009F
57cdf0e10cSrcweir static sal_uChar const aImpl8090SameToCharTab[SAME8090CHAR_END
58cdf0e10cSrcweir                                                   - SAME8090CHAR_START
59cdf0e10cSrcweir                                                   + 1]
60cdf0e10cSrcweir     = { 0x80, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, /* 0x0080 */
61cdf0e10cSrcweir         0x88, 0x89, 0x8A, 0x8B, 0x8C, 0x8D, 0x8E, 0x8F,
62cdf0e10cSrcweir         0x90, 0x91, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, /* 0x0090 */
63cdf0e10cSrcweir         0x98, 0x99, 0x9A, 0x9B, 0x9C, 0x9D, 0x9E, 0x9F };
64cdf0e10cSrcweir 
65cdf0e10cSrcweir #define SAMEA0FFCHAR_START 0x00A0
66cdf0e10cSrcweir #define SAMEA0FFCHAR_END 0x00FF
67cdf0e10cSrcweir static sal_uChar const aImplA0FFSameToCharTab[SAMEA0FFCHAR_END
68cdf0e10cSrcweir                                                   - SAMEA0FFCHAR_START
69cdf0e10cSrcweir                                                   + 1]
70cdf0e10cSrcweir     = { 0xA0, 0xA1, 0xA2, 0xA3, 0xA4, 0xA5, 0xA6, 0xA7, /* 0x00A0 */
71cdf0e10cSrcweir         0xA8, 0xA9, 0xAA, 0xAB, 0xAC, 0xAD, 0xAE, 0xAF,
72cdf0e10cSrcweir         0xB0, 0xB1, 0xB2, 0xB3, 0xB4, 0xB5, 0xB6, 0xB7, /* 0x00B0 */
73cdf0e10cSrcweir         0xB8, 0xB9, 0xBA, 0xBB, 0xBC, 0xBD, 0xBE, 0xBF,
74cdf0e10cSrcweir         0xC0, 0xC1, 0xC2, 0xC3, 0xC4, 0xC5, 0xC6, 0xC7, /* 0x00C0 */
75cdf0e10cSrcweir         0xC8, 0xC9, 0xCA, 0xCB, 0xCC, 0xCD, 0xCE, 0xCF,
76cdf0e10cSrcweir         0xD0, 0xD1, 0xD2, 0xD3, 0xD4, 0xD5, 0xD6, 0xD7, /* 0x00D0 */
77cdf0e10cSrcweir         0xD8, 0xD9, 0xDA, 0xDB, 0xDC, 0xDD, 0xDE, 0xDF,
78cdf0e10cSrcweir         0xE0, 0xE1, 0xE2, 0xE3, 0xE4, 0xE5, 0xE6, 0xE7, /* 0x00E0 */
79cdf0e10cSrcweir         0xE8, 0xE9, 0xEA, 0xEB, 0xEC, 0xED, 0xEE, 0xEF,
80cdf0e10cSrcweir         0xF0, 0xF1, 0xF2, 0xF3, 0xF4, 0xF5, 0xF6, 0xF7, /* 0x00F0 */
81cdf0e10cSrcweir         0xF8, 0xF9, 0xFA, 0xFB, 0xFC, 0xFD, 0xFE, 0xFF };
82cdf0e10cSrcweir 
83cdf0e10cSrcweir static sal_uInt16 const aImplDoubleByteIdentifierTab[1] = { 0 };
84cdf0e10cSrcweir 
85cdf0e10cSrcweir #include "tcvtarb1.tab"
86cdf0e10cSrcweir #include "tcvteas1.tab"
87cdf0e10cSrcweir #include "tcvtest1.tab"
88cdf0e10cSrcweir #include "tcvtjp1.tab"
89cdf0e10cSrcweir #include "tcvtjp2.tab"
90cdf0e10cSrcweir #include "tcvtjp3.tab"
91cdf0e10cSrcweir #include "tcvtjp4.tab"
92cdf0e10cSrcweir #include "tcvtjp5.tab"
93cdf0e10cSrcweir #include "tcvtjp6.tab"
94cdf0e10cSrcweir #include "tcvtkr1.tab"
95cdf0e10cSrcweir #include "tcvtkr2.tab"
96cdf0e10cSrcweir #include "tcvtkr4.tab"
97cdf0e10cSrcweir #include "tcvtkr5.tab"
98cdf0e10cSrcweir #include "tcvtkr6.tab"
99cdf0e10cSrcweir #include "tcvtlat1.tab"
100cdf0e10cSrcweir #include "tcvtscn1.tab"
101cdf0e10cSrcweir #include "tcvtscn2.tab"
102cdf0e10cSrcweir #include "tcvtscn3.tab"
103cdf0e10cSrcweir #include "tcvtscn4.tab"
104cdf0e10cSrcweir #include "tcvtscn5.tab"
105cdf0e10cSrcweir #include "tcvtscn6.tab"
106cdf0e10cSrcweir #include "tcvtsym1.tab"
107cdf0e10cSrcweir #include "tcvttcn1.tab"
108cdf0e10cSrcweir #include "tcvttcn2.tab"
109cdf0e10cSrcweir #include "tcvttcn6.tab"
110cdf0e10cSrcweir #include "tcvtuni1.tab"
111cdf0e10cSrcweir 
112cdf0e10cSrcweir #include "convertbig5hkscs.tab"
113cdf0e10cSrcweir #include "converteuctw.tab"
114cdf0e10cSrcweir #include "convertgb18030.tab"
115cdf0e10cSrcweir #include "convertiso2022cn.tab"
116cdf0e10cSrcweir #include "convertiso2022jp.tab"
117cdf0e10cSrcweir #include "convertiso2022kr.tab"
118cdf0e10cSrcweir #include "convertadobe.tab"
119cdf0e10cSrcweir 
120cdf0e10cSrcweir ImplTextEncodingData const *
Impl_getTextEncodingData(rtl_TextEncoding nEncoding)121cdf0e10cSrcweir Impl_getTextEncodingData(rtl_TextEncoding nEncoding) SAL_THROW_EXTERN_C()
122cdf0e10cSrcweir {
123cdf0e10cSrcweir     static ImplTextEncodingData const * const aData[]
124cdf0e10cSrcweir         = { NULL, /* DONTKNOW */
125cdf0e10cSrcweir             &aImplMS1252TextEncodingData, /* MS_1252 */
126cdf0e10cSrcweir             &aImplAPPLEROMANTextEncodingData, /* APPLE_ROMAN */
127cdf0e10cSrcweir             &aImplIBM437TextEncodingData, /* IBM_437 */
128cdf0e10cSrcweir             &aImplIBM850TextEncodingData, /* IBM_850 */
129cdf0e10cSrcweir             &aImplIBM860TextEncodingData, /* IBM_860 */
130cdf0e10cSrcweir             &aImplIBM861TextEncodingData, /* IBM_861 */
131cdf0e10cSrcweir             &aImplIBM863TextEncodingData, /* IBM_863 */
132cdf0e10cSrcweir             &aImplIBM865TextEncodingData, /* IBM_865 */
133cdf0e10cSrcweir             NULL, /* reserved (SYSTEM) */
134cdf0e10cSrcweir             &aImplSYMBOLTextEncodingData, /* SYMBOL */
135cdf0e10cSrcweir             &aImplUSASCIITextEncodingData, /* ASCII_US */
136cdf0e10cSrcweir             &aImplISO88591TextEncodingData, /* ISO_8859_1 */
137cdf0e10cSrcweir             &aImplISO88592TextEncodingData, /* ISO_8859_2 */
138cdf0e10cSrcweir             &aImplISO88593TextEncodingData, /* ISO_8859_3 */
139cdf0e10cSrcweir             &aImplISO88594TextEncodingData, /* ISO_8859_4 */
140cdf0e10cSrcweir             &aImplISO88595TextEncodingData, /* ISO_8859_5 */
141cdf0e10cSrcweir             &aImplISO88596TextEncodingData, /* ISO_8859_6 */
142cdf0e10cSrcweir             &aImplISO88597TextEncodingData, /* ISO_8859_7 */
143cdf0e10cSrcweir             &aImplISO88598TextEncodingData, /* ISO_8859_8 */
144cdf0e10cSrcweir             &aImplISO88599TextEncodingData, /* ISO_8859_9 */
145cdf0e10cSrcweir             &aImplISO885914TextEncodingData, /* ISO_8859_14 */
146cdf0e10cSrcweir             &aImplISO885915TextEncodingData, /* ISO_8859_15 */
147cdf0e10cSrcweir             &aImplIBM737TextEncodingData, /* IBM_737 */
148cdf0e10cSrcweir             &aImplIBM775TextEncodingData, /* IBM_775 */
149cdf0e10cSrcweir             &aImplIBM852TextEncodingData, /* IBM_852 */
150cdf0e10cSrcweir             &aImplIBM855TextEncodingData, /* IBM_855 */
151cdf0e10cSrcweir             &aImplIBM857TextEncodingData, /* IBM_857 */
152cdf0e10cSrcweir             &aImplIBM862TextEncodingData, /* IBM_862 */
153cdf0e10cSrcweir             &aImplIBM864TextEncodingData, /* IBM_864 */
154cdf0e10cSrcweir             &aImplIBM866TextEncodingData, /* IBM_866 */
155cdf0e10cSrcweir             &aImplIBM869TextEncodingData, /* IBM_869 */
156cdf0e10cSrcweir             &aImplMS874TextEncodingData, /* MS_874 */
157cdf0e10cSrcweir             &aImplMS1250TextEncodingData, /* MS_1250 */
158cdf0e10cSrcweir             &aImplMS1251TextEncodingData, /* MS_1251 */
159cdf0e10cSrcweir             &aImplMS1253TextEncodingData, /* MS_1253 */
160cdf0e10cSrcweir             &aImplMS1254TextEncodingData, /* MS_1254 */
161cdf0e10cSrcweir             &aImplMS1255TextEncodingData, /* MS_1255 */
162cdf0e10cSrcweir             &aImplMS1256TextEncodingData, /* MS_1256 */
163cdf0e10cSrcweir             &aImplMS1257TextEncodingData, /* MS_1257 */
164cdf0e10cSrcweir             &aImplMS1258TextEncodingData, /* MS_1258 */
165cdf0e10cSrcweir             NULL, /* TODO! APPLE_ARABIC */
166cdf0e10cSrcweir             &aImplAPPLECENTEUROTextEncodingData, /* APPLE_CENTEURO */
167cdf0e10cSrcweir             &aImplAPPLECROATIANTextEncodingData, /* APPLE_CROATIAN */
168cdf0e10cSrcweir             &aImplAPPLECYRILLICTextEncodingData, /* APPLE_CYRILLIC */
169cdf0e10cSrcweir             NULL, /* TODO! APPLE_DEVANAGARI */
170cdf0e10cSrcweir             NULL, /* TODO! APPLE_FARSI */
171cdf0e10cSrcweir             &aImplAPPLEGREEKTextEncodingData, /* APPLE_GREEK */
172cdf0e10cSrcweir             NULL, /* TODO! APPLE_GUJARATI */
173cdf0e10cSrcweir             NULL, /* TODO! APPLE_GURMUKHI */
174cdf0e10cSrcweir             NULL, /* TODO! APPLE_HEBREW */
175cdf0e10cSrcweir             &aImplAPPLEICELANDTextEncodingData, /* APPLE_ICELAND */
176cdf0e10cSrcweir             &aImplAPPLEROMANIANTextEncodingData, /* APPLE_ROMANIAN */
177cdf0e10cSrcweir             NULL, /* TODO! APPLE_THAI */
178cdf0e10cSrcweir             &aImplAPPLETURKISHTextEncodingData, /* APPLE_TURKISH */
179cdf0e10cSrcweir             &aImplAPPLEUKRAINIANTextEncodingData, /* APPLE_UKRAINIAN */
180cdf0e10cSrcweir             &aImplAPPLECHINSIMPTextEncodingData, /* APPLE_CHINSIMP */
181cdf0e10cSrcweir             &aImplAPPLECHINTRADTextEncodingData, /* APPLE_CHINTRAD */
182cdf0e10cSrcweir             &aImplAPPLEJAPANESETextEncodingData, /* APPLE_JAPANESE */
183cdf0e10cSrcweir             &aImplAPPLEKOREANTextEncodingData, /* APPLE_KOREAN */
184cdf0e10cSrcweir             &aImplMS932TextEncodingData, /* MS_932 */
185cdf0e10cSrcweir             &aImplMS936TextEncodingData, /* MS_936 */
186cdf0e10cSrcweir             &aImplMS949TextEncodingData, /* MS_949 */
187cdf0e10cSrcweir             &aImplMS950TextEncodingData, /* MS_950 */
188cdf0e10cSrcweir             &aImplSJISTextEncodingData, /* SHIFT_JIS */
189cdf0e10cSrcweir             &aImplGB2312TextEncodingData, /* GB_2312 */
190cdf0e10cSrcweir             &aImplGBT12345TextEncodingData, /* GBT_12345 */
191cdf0e10cSrcweir             &aImplGBKTextEncodingData, /* GBK */
192cdf0e10cSrcweir             &aImplBIG5TextEncodingData, /* BIG5 */
193cdf0e10cSrcweir             &aImplEUCJPTextEncodingData, /* EUC_JP */
194cdf0e10cSrcweir             &aImplEUCCNTextEncodingData, /* EUC_CN */
195cdf0e10cSrcweir             &aImplEucTwTextEncodingData, /* EUC_TW */
196cdf0e10cSrcweir             &aImplIso2022JpTextEncodingData, /* ISO_2022_JP */
197cdf0e10cSrcweir             &aImplIso2022CnTextEncodingData, /* ISO_2022_CN */
198cdf0e10cSrcweir             &aImplKOI8RTextEncodingData, /* KOI8_R */
199cdf0e10cSrcweir             &aImplUTF7TextEncodingData, /* UTF7 */
200cdf0e10cSrcweir             &aImplUTF8TextEncodingData, /* UTF8 */
201cdf0e10cSrcweir             &aImplISO885910TextEncodingData, /* ISO_8859_10 */
202cdf0e10cSrcweir             &aImplISO885913TextEncodingData, /* ISO_8859_13 */
203cdf0e10cSrcweir             &aImplEUCKRTextEncodingData, /* EUC_KR */
204cdf0e10cSrcweir             &aImplIso2022KrTextEncodingData, /* ISO_2022_KR */
205cdf0e10cSrcweir             &aImplJISX0201TextEncodingData, /* JIS_X_0201 */
206cdf0e10cSrcweir             &aImplJISX0208TextEncodingData, /* JIS_X_0208 */
207cdf0e10cSrcweir             &aImplJISX0212TextEncodingData, /* JIS_X_0212 */
208cdf0e10cSrcweir             &aImplMS1361TextEncodingData, /* MS_1361 */
209cdf0e10cSrcweir             &aImplGb18030TextEncodingData, /* GB_18030 */
210cdf0e10cSrcweir             &aImplBig5HkscsTextEncodingData, /* BIG5_HKSCS */
211cdf0e10cSrcweir             &aImplTis620TextEncodingData, /* TIS_620 */
212cdf0e10cSrcweir             &aImplKoi8UTextEncodingData, /* KOI8_U */
213cdf0e10cSrcweir             &aImplJavaUtf8TextEncodingData, /* JAVA_UTF8 */
214cdf0e10cSrcweir             &adobeStandardEncodingData, /* ADOBE_STANDARD */
215cdf0e10cSrcweir             &adobeSymbolEncodingData, /* ADOBE_SYMBOL */
216cdf0e10cSrcweir             &aImplPT154TextEncodingData, /* PT154 */
217cdf0e10cSrcweir             &adobeDingbatsEncodingData }; /* ADOBE_DINGBATS */
218cdf0e10cSrcweir     return
219cdf0e10cSrcweir         nEncoding < sizeof aData / sizeof aData[0] ? aData[nEncoding] : NULL;
220cdf0e10cSrcweir }
221