1 /*************************************************************************
2  *
3  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
4  *
5  * Copyright 2000, 2010 Oracle and/or its affiliates.
6  *
7  * OpenOffice.org - a multi-platform office productivity suite
8  *
9  * This file is part of OpenOffice.org.
10  *
11  * OpenOffice.org is free software: you can redistribute it and/or modify
12  * it under the terms of the GNU Lesser General Public License version 3
13  * only, as published by the Free Software Foundation.
14  *
15  * OpenOffice.org is distributed in the hope that it will be useful,
16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
18  * GNU Lesser General Public License version 3 for more details
19  * (a copy is included in the LICENSE file that accompanied this code).
20  *
21  * You should have received a copy of the GNU Lesser General Public License
22  * version 3 along with OpenOffice.org.  If not, see
23  * <http://www.openoffice.org/license.html>
24  * for a copy of the LGPLv3 License.
25  *
26  ************************************************************************/
27 
28 // MARKER(update_precomp.py): autogen include statement, do not remove
29 #include "precompiled_sal.hxx"
30 
31 #include "context.h"
32 #include "converter.h"
33 #include "convertsinglebytetobmpunicode.hxx"
34 #include "unichars.h"
35 
36 #include "osl/diagnose.h"
37 #include "rtl/textcvt.h"
38 #include "sal/types.h"
39 
40 #include <cstddef>
41 
42 sal_Size rtl_textenc_convertSingleByteToBmpUnicode(
43     ImplTextConverterData const * data, void *, sal_Char const * srcBuf,
44     sal_Size srcBytes, sal_Unicode * destBuf, sal_Size destChars,
45     sal_uInt32 flags, sal_uInt32 * info, sal_Size * srcCvtBytes)
46 {
47     sal_Unicode const * map = static_cast<
48         rtl::textenc::BmpUnicodeToSingleByteConverterData const * >(
49             data)->byteToUnicode;
50     sal_uInt32 infoFlags = 0;
51     sal_Size converted = 0;
52     sal_Unicode * destBufPtr = destBuf;
53     sal_Unicode * destBufEnd = destBuf + destChars;
54     for (; converted < srcBytes; ++converted) {
55         bool undefined = true;
56         sal_Char b = *srcBuf++;
57         sal_Unicode c = map[static_cast< sal_uInt8 >(b)];
58         if (c == 0xFFFF) {
59             goto bad_input;
60         }
61         if (destBufEnd - destBufPtr < 1) {
62             goto no_output;
63         }
64         *destBufPtr++ = c;
65         continue;
66     bad_input:
67         switch (ImplHandleBadInputTextToUnicodeConversion(
68                     undefined, false, b, flags, &destBufPtr, destBufEnd,
69                     &infoFlags))
70         {
71         case IMPL_BAD_INPUT_STOP:
72             break;
73 
74         case IMPL_BAD_INPUT_CONTINUE:
75             continue;
76 
77         case IMPL_BAD_INPUT_NO_OUTPUT:
78             goto no_output;
79         }
80         break;
81     no_output:
82         --srcBuf;
83         infoFlags |= RTL_TEXTTOUNICODE_INFO_DESTBUFFERTOSMALL;
84         break;
85     }
86     if (info != 0) {
87         *info = infoFlags;
88     }
89     if (srcCvtBytes != 0) {
90         *srcCvtBytes = converted;
91     }
92     return destBufPtr - destBuf;
93 }
94 
95 sal_Size rtl_textenc_convertBmpUnicodeToSingleByte(
96     ImplTextConverterData const * data, void * context,
97     sal_Unicode const * srcBuf, sal_Size srcChars, sal_Char * destBuf,
98     sal_Size destBytes, sal_uInt32 flags, sal_uInt32 * info,
99     sal_Size * srcCvtChars)
100 {
101     std::size_t entries = static_cast<
102         rtl::textenc::BmpUnicodeToSingleByteConverterData const * >(
103             data)->unicodeToByteEntries;
104     rtl::textenc::BmpUnicodeToSingleByteRange const * ranges = static_cast<
105         rtl::textenc::BmpUnicodeToSingleByteConverterData const * >(
106             data)->unicodeToByte;
107     sal_Unicode highSurrogate = 0;
108     sal_uInt32 infoFlags = 0;
109     sal_Size converted = 0;
110     sal_Char * destBufPtr = destBuf;
111     sal_Char * destBufEnd = destBuf + destBytes;
112     if (context != 0) {
113         highSurrogate = static_cast< ImplUnicodeToTextContext * >(context)->
114             m_nHighSurrogate;
115     }
116     for (; converted < srcChars; ++converted) {
117         bool undefined = true;
118         sal_uInt32 c = *srcBuf++;
119         if (highSurrogate == 0) {
120             if (ImplIsHighSurrogate(c)) {
121                 highSurrogate = static_cast< sal_Unicode >(c);
122                 continue;
123             }
124         } else if (ImplIsLowSurrogate(c)) {
125             c = ImplCombineSurrogates(highSurrogate, c);
126         } else {
127             undefined = false;
128             goto bad_input;
129         }
130         if (ImplIsLowSurrogate(c) || ImplIsNoncharacter(c)) {
131             undefined = false;
132             goto bad_input;
133         }
134         // Linearly searching through the ranges if probably fastest, assuming
135         // that most converted characters belong to the ASCII subset:
136         for (std::size_t i = 0; i < entries; ++i) {
137             if (c < ranges[i].unicode) {
138                 break;
139             } else if (c <= sal::static_int_cast< sal_uInt32 >(
140                            ranges[i].unicode + ranges[i].range))
141             {
142                 if (destBufEnd - destBufPtr < 1) {
143                     goto no_output;
144                 }
145                 *destBufPtr++ = static_cast< sal_Char >(
146                     ranges[i].byte + (c - ranges[i].unicode));
147                 goto done;
148             }
149         }
150         goto bad_input;
151     done:
152         highSurrogate = 0;
153         continue;
154     bad_input:
155         switch (ImplHandleBadInputUnicodeToTextConversion(
156                     undefined, c, flags, &destBufPtr, destBufEnd, &infoFlags, 0,
157                     0, 0))
158         {
159         case IMPL_BAD_INPUT_STOP:
160             highSurrogate = 0;
161             break;
162 
163         case IMPL_BAD_INPUT_CONTINUE:
164             highSurrogate = 0;
165             continue;
166 
167         case IMPL_BAD_INPUT_NO_OUTPUT:
168             goto no_output;
169         }
170         break;
171     no_output:
172         --srcBuf;
173         infoFlags |= RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL;
174         break;
175     }
176     if (highSurrogate != 0
177         && ((infoFlags
178              & (RTL_UNICODETOTEXT_INFO_ERROR
179                 | RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL))
180             == 0))
181     {
182         if ((flags & RTL_UNICODETOTEXT_FLAGS_FLUSH) != 0) {
183             infoFlags |= RTL_UNICODETOTEXT_INFO_SRCBUFFERTOSMALL;
184         } else {
185             switch (ImplHandleBadInputUnicodeToTextConversion(
186                         false, 0, flags, &destBufPtr, destBufEnd, &infoFlags, 0,
187                         0, 0))
188             {
189             case IMPL_BAD_INPUT_STOP:
190             case IMPL_BAD_INPUT_CONTINUE:
191                 highSurrogate = 0;
192                 break;
193 
194             case IMPL_BAD_INPUT_NO_OUTPUT:
195                 infoFlags |= RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL;
196                 break;
197             }
198         }
199     }
200     if (context != 0) {
201         static_cast< ImplUnicodeToTextContext * >(context)->m_nHighSurrogate
202             = highSurrogate;
203     }
204     if (info != 0) {
205         *info = infoFlags;
206     }
207     if (srcCvtChars != 0) {
208         *srcCvtChars = converted;
209     }
210     return destBufPtr - destBuf;
211 }
212