1 /**************************************************************
2  *
3  * Licensed to the Apache Software Foundation (ASF) under one
4  * or more contributor license agreements.  See the NOTICE file
5  * distributed with this work for additional information
6  * regarding copyright ownership.  The ASF licenses this file
7  * to you under the Apache License, Version 2.0 (the
8  * "License"); you may not use this file except in compliance
9  * with the License.  You may obtain a copy of the License at
10  *
11  *   http://www.apache.org/licenses/LICENSE-2.0
12  *
13  * Unless required by applicable law or agreed to in writing,
14  * software distributed under the License is distributed on an
15  * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
16  * KIND, either express or implied.  See the License for the
17  * specific language governing permissions and limitations
18  * under the License.
19  *
20  *************************************************************/
21 
22 
23 
24 // MARKER(update_precomp.py): autogen include statement, do not remove
25 #include "precompiled_sal.hxx"
26 
27 #include "context.h"
28 #include "converter.h"
29 #include "convertsinglebytetobmpunicode.hxx"
30 #include "unichars.h"
31 
32 #include "osl/diagnose.h"
33 #include "rtl/textcvt.h"
34 #include "sal/types.h"
35 
36 #include <cstddef>
37 
rtl_textenc_convertSingleByteToBmpUnicode(ImplTextConverterData const * data,void *,sal_Char const * srcBuf,sal_Size srcBytes,sal_Unicode * destBuf,sal_Size destChars,sal_uInt32 flags,sal_uInt32 * info,sal_Size * srcCvtBytes)38 sal_Size rtl_textenc_convertSingleByteToBmpUnicode(
39     ImplTextConverterData const * data, void *, sal_Char const * srcBuf,
40     sal_Size srcBytes, sal_Unicode * destBuf, sal_Size destChars,
41     sal_uInt32 flags, sal_uInt32 * info, sal_Size * srcCvtBytes)
42 {
43     sal_Unicode const * map = static_cast<
44         rtl::textenc::BmpUnicodeToSingleByteConverterData const * >(
45             data)->byteToUnicode;
46     sal_uInt32 infoFlags = 0;
47     sal_Size converted = 0;
48     sal_Unicode * destBufPtr = destBuf;
49     sal_Unicode * destBufEnd = destBuf + destChars;
50     for (; converted < srcBytes; ++converted) {
51         bool undefined = true;
52         sal_Char b = *srcBuf++;
53         sal_Unicode c = map[static_cast< sal_uInt8 >(b)];
54         if (c == 0xFFFF) {
55             goto bad_input;
56         }
57         if (destBufEnd - destBufPtr < 1) {
58             goto no_output;
59         }
60         *destBufPtr++ = c;
61         continue;
62     bad_input:
63         switch (ImplHandleBadInputTextToUnicodeConversion(
64                     undefined, false, b, flags, &destBufPtr, destBufEnd,
65                     &infoFlags))
66         {
67         case IMPL_BAD_INPUT_STOP:
68             break;
69 
70         case IMPL_BAD_INPUT_CONTINUE:
71             continue;
72 
73         case IMPL_BAD_INPUT_NO_OUTPUT:
74             goto no_output;
75         }
76         break;
77     no_output:
78         --srcBuf;
79         infoFlags |= RTL_TEXTTOUNICODE_INFO_DESTBUFFERTOSMALL;
80         break;
81     }
82     if (info != 0) {
83         *info = infoFlags;
84     }
85     if (srcCvtBytes != 0) {
86         *srcCvtBytes = converted;
87     }
88     return destBufPtr - destBuf;
89 }
90 
rtl_textenc_convertBmpUnicodeToSingleByte(ImplTextConverterData const * data,void * context,sal_Unicode const * srcBuf,sal_Size srcChars,sal_Char * destBuf,sal_Size destBytes,sal_uInt32 flags,sal_uInt32 * info,sal_Size * srcCvtChars)91 sal_Size rtl_textenc_convertBmpUnicodeToSingleByte(
92     ImplTextConverterData const * data, void * context,
93     sal_Unicode const * srcBuf, sal_Size srcChars, sal_Char * destBuf,
94     sal_Size destBytes, sal_uInt32 flags, sal_uInt32 * info,
95     sal_Size * srcCvtChars)
96 {
97     std::size_t entries = static_cast<
98         rtl::textenc::BmpUnicodeToSingleByteConverterData const * >(
99             data)->unicodeToByteEntries;
100     rtl::textenc::BmpUnicodeToSingleByteRange const * ranges = static_cast<
101         rtl::textenc::BmpUnicodeToSingleByteConverterData const * >(
102             data)->unicodeToByte;
103     sal_Unicode highSurrogate = 0;
104     sal_uInt32 infoFlags = 0;
105     sal_Size converted = 0;
106     sal_Char * destBufPtr = destBuf;
107     sal_Char * destBufEnd = destBuf + destBytes;
108     if (context != 0) {
109         highSurrogate = static_cast< ImplUnicodeToTextContext * >(context)->
110             m_nHighSurrogate;
111     }
112     for (; converted < srcChars; ++converted) {
113         bool undefined = true;
114         sal_uInt32 c = *srcBuf++;
115         if (highSurrogate == 0) {
116             if (ImplIsHighSurrogate(c)) {
117                 highSurrogate = static_cast< sal_Unicode >(c);
118                 continue;
119             }
120         } else if (ImplIsLowSurrogate(c)) {
121             c = ImplCombineSurrogates(highSurrogate, c);
122         } else {
123             undefined = false;
124             goto bad_input;
125         }
126         if (ImplIsLowSurrogate(c) || ImplIsNoncharacter(c)) {
127             undefined = false;
128             goto bad_input;
129         }
130         // Linearly searching through the ranges if probably fastest, assuming
131         // that most converted characters belong to the ASCII subset:
132         for (std::size_t i = 0; i < entries; ++i) {
133             if (c < ranges[i].unicode) {
134                 break;
135             } else if (c <= sal::static_int_cast< sal_uInt32 >(
136                            ranges[i].unicode + ranges[i].range))
137             {
138                 if (destBufEnd - destBufPtr < 1) {
139                     goto no_output;
140                 }
141                 *destBufPtr++ = static_cast< sal_Char >(
142                     ranges[i].byte + (c - ranges[i].unicode));
143                 goto done;
144             }
145         }
146         goto bad_input;
147     done:
148         highSurrogate = 0;
149         continue;
150     bad_input:
151         switch (ImplHandleBadInputUnicodeToTextConversion(
152                     undefined, c, flags, &destBufPtr, destBufEnd, &infoFlags, 0,
153                     0, 0))
154         {
155         case IMPL_BAD_INPUT_STOP:
156             highSurrogate = 0;
157             break;
158 
159         case IMPL_BAD_INPUT_CONTINUE:
160             highSurrogate = 0;
161             continue;
162 
163         case IMPL_BAD_INPUT_NO_OUTPUT:
164             goto no_output;
165         }
166         break;
167     no_output:
168         --srcBuf;
169         infoFlags |= RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL;
170         break;
171     }
172     if (highSurrogate != 0
173         && ((infoFlags
174              & (RTL_UNICODETOTEXT_INFO_ERROR
175                 | RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL))
176             == 0))
177     {
178         if ((flags & RTL_UNICODETOTEXT_FLAGS_FLUSH) != 0) {
179             infoFlags |= RTL_UNICODETOTEXT_INFO_SRCBUFFERTOSMALL;
180         } else {
181             switch (ImplHandleBadInputUnicodeToTextConversion(
182                         false, 0, flags, &destBufPtr, destBufEnd, &infoFlags, 0,
183                         0, 0))
184             {
185             case IMPL_BAD_INPUT_STOP:
186             case IMPL_BAD_INPUT_CONTINUE:
187                 highSurrogate = 0;
188                 break;
189 
190             case IMPL_BAD_INPUT_NO_OUTPUT:
191                 infoFlags |= RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL;
192                 break;
193             }
194         }
195     }
196     if (context != 0) {
197         static_cast< ImplUnicodeToTextContext * >(context)->m_nHighSurrogate
198             = highSurrogate;
199     }
200     if (info != 0) {
201         *info = infoFlags;
202     }
203     if (srcCvtChars != 0) {
204         *srcCvtChars = converted;
205     }
206     return destBufPtr - destBuf;
207 }
208