/src/libreoffice/sal/textenc/tcvtutf8.cxx

Source (jump to first uncovered line)
/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */
/*
 * This file is part of the LibreOffice project.
 *
 * This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/.
 *
 * This file incorporates work covered by the following license notice:
 *
 *   Licensed to the Apache Software Foundation (ASF) under one or more
 *   contributor license agreements. See the NOTICE file distributed
 *   with this work for additional information regarding copyright
 *   ownership. The ASF licenses this file to you under the Apache
 *   License, Version 2.0 (the "License"); you may not use this file
 *   except in compliance with the License. You may obtain a copy of
 *   the License at http://www.apache.org/licenses/LICENSE-2.0 .
 */

#include <sal/config.h>

#include <cassert>

#include <sal/types.h>
#include <rtl/character.hxx>
#include <rtl/textcvt.h>

#include "converter.hxx"
#include "tcvtutf8.hxx"

namespace {

struct ImplUtf8ToUnicodeContext
{
    sal_uInt32 nUtf32;
    int nBytes;
    int nShift;
    bool bCheckBom;
};

struct ImplUnicodeToUtf8Context
{
    sal_Unicode nHighSurrogate; /* 0xFFFF: write BOM */
};

}

void * ImplCreateUtf8ToUnicodeContext()
{
    ImplUtf8ToUnicodeContext * p = new ImplUtf8ToUnicodeContext;
    ImplResetUtf8ToUnicodeContext(p);
    return p;
}

void ImplResetUtf8ToUnicodeContext(void * pContext)
{
    if (pContext != nullptr)
    {
        static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nBytes = 1;
        static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nShift = -1;
        static_cast< ImplUtf8ToUnicodeContext * >(pContext)->bCheckBom = true;
    }
}

void ImplDestroyUtf8ToUnicodeContext(void * pContext)
{
    delete static_cast< ImplUtf8ToUnicodeContext * >(pContext);
}

sal_Size ImplConvertUtf8ToUnicode(
    void const * pData, void * pContext, char const * pSrcBuf,
    sal_Size nSrcBytes, sal_Unicode * pDestBuf, sal_Size nDestChars,
    sal_uInt32 nFlags, sal_uInt32 * pInfo, sal_Size * pSrcCvtBytes)
{
    bool bJavaUtf8 = pData != nullptr;
    sal_uInt32 nUtf32 = 0;
    int nBytes = 1;
    int nShift = -1;
    bool bCheckBom = true;
    sal_uInt32 nInfo = 0;
    unsigned char const * pSrcBufPtr = reinterpret_cast<unsigned char const *>(pSrcBuf);
    unsigned char const * pSrcBufEnd = pSrcBufPtr + nSrcBytes;
    sal_Unicode * pDestBufPtr = pDestBuf;
    sal_Unicode * pDestBufEnd = pDestBufPtr + nDestChars;
    unsigned char const * startOfCurrentChar = pSrcBufPtr;

    if (pContext != nullptr)
    {
        nUtf32 = static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nUtf32;
        nBytes = static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nBytes;
        nShift = static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nShift;
        bCheckBom = static_cast< ImplUtf8ToUnicodeContext * >(pContext)->bCheckBom;
    }

    while (pSrcBufPtr < pSrcBufEnd)
    {
        bool bConsume = true;
        sal_uInt32 nChar = *pSrcBufPtr++;
        if (nShift < 0)
            // Allow (illegal) 5 and 6 byte sequences, so they are read as a
            // single individual bad character:
            if (nChar <= 0x7F)
            {
                nUtf32 = nChar;
                nBytes = 1;
                goto transform;
            }
            else if (nChar <= 0xBF)
                goto bad_input;
            else if (nChar <= 0xDF)
            {
                nUtf32 = (nChar & 0x1F) << 6;
                nBytes = 2;
                nShift = 0;
            }
            else if (nChar <= 0xEF)
            {
                nUtf32 = (nChar & 0x0F) << 12;
                nBytes = 3;
                nShift = 6;
            }
            else if (nChar <= 0xF7)
            {
                nUtf32 = (nChar & 0x07) << 18;
                nBytes = 4;
                nShift = 12;
            }
            else if (nChar <= 0xFB)
            {
                nUtf32 = (nChar & 0x03) << 24;
                nBytes = 5;
                nShift = 18;
            }
            else if (nChar <= 0xFD)
            {
                nUtf32 = (nChar & 0x01) << 30;
                nBytes = 6;
                nShift = 24;
            }
            else
                goto bad_input;
        else if ((nChar & 0xC0) == 0x80)
        {
            nUtf32 |= (nChar & 0x3F) << nShift;
            if (nShift == 0)
                goto transform;
            else
                nShift -= 6;
        }
        else
        {
            /*
             This byte is preceded by a broken UTF-8 sequence; if this byte
             is neither in the range [0x80..0xBF] nor in the range
             [0xFE..0xFF], assume that this byte does not belong to that
             broken sequence, but instead starts a new, legal UTF-8 sequence:
             */
            bConsume = nChar >= 0xFE;
            goto bad_input;
        }
        continue;

    transform:
        if (!bCheckBom || nUtf32 != 0xFEFF || nBytes != 3
            || (nFlags & RTL_TEXTTOUNICODE_FLAGS_GLOBAL_SIGNATURE) == 0
            || bJavaUtf8)
        {
            switch (nBytes) {
            case 1:
                if (bJavaUtf8 && nUtf32 == 0) {
                    goto bad_input;
                }
                break;
            case 2:
                if (nUtf32 < 0x80 && !(bJavaUtf8 && nUtf32 == 0)) {
                    goto bad_input;
                }
                break;
            case 3:
                if (nUtf32 < 0x800 || (!bJavaUtf8 && rtl::isSurrogate(nUtf32)))
                {
                    goto bad_input;
                }
                break;
            case 4:
                if (nUtf32 < 0x10000 || !rtl::isUnicodeCodePoint(nUtf32)
                    || bJavaUtf8)
                {
                    goto bad_input;
                }
                break;
            default:
                goto bad_input;
            }
            if (nUtf32 <= 0xFFFF)
                if (pDestBufPtr != pDestBufEnd)
                    *pDestBufPtr++ = static_cast<sal_Unicode>(nUtf32);
                else
                    goto no_output;
            else if (pDestBufEnd - pDestBufPtr >= 2)
                pDestBufPtr += rtl::splitSurrogates(nUtf32, pDestBufPtr);
            else
                goto no_output;
        }
        nShift = -1;
        bCheckBom = false;
        startOfCurrentChar = pSrcBufPtr;
        continue;

    bad_input:
        switch (sal::detail::textenc::handleBadInputTextToUnicodeConversion(
                    false, nBytes != 1, 0, nFlags, &pDestBufPtr, pDestBufEnd,
                    &nInfo))
        {
        case sal::detail::textenc::BAD_INPUT_STOP:
            nShift = -1;
            bCheckBom = false;
            if ((nFlags & RTL_TEXTTOUNICODE_FLAGS_FLUSH) == 0) {
                if (!bConsume)
                    --pSrcBufPtr;
            } else {
                pSrcBufPtr = startOfCurrentChar;
            }
            break;

        case sal::detail::textenc::BAD_INPUT_CONTINUE:
            nShift = -1;
            bCheckBom = false;
            if (!bConsume)
                --pSrcBufPtr;
            startOfCurrentChar = pSrcBufPtr;
            continue;

        case sal::detail::textenc::BAD_INPUT_NO_OUTPUT:
            goto no_output;
        }
        break;

    no_output:
        --pSrcBufPtr;
        nInfo |= RTL_TEXTTOUNICODE_INFO_DESTBUFFERTOOSMALL;
        break;
    }

    if (nShift >= 0
        && (nInfo & (RTL_TEXTTOUNICODE_INFO_ERROR
                         | RTL_TEXTTOUNICODE_INFO_DESTBUFFERTOOSMALL))
               == 0)
    {
        if ((nFlags & RTL_TEXTTOUNICODE_FLAGS_FLUSH) == 0)
            nInfo |= RTL_TEXTTOUNICODE_INFO_SRCBUFFERTOOSMALL;
        else
            switch (sal::detail::textenc::handleBadInputTextToUnicodeConversion(
                        false, true, 0, nFlags, &pDestBufPtr, pDestBufEnd,
                        &nInfo))
            {
            case sal::detail::textenc::BAD_INPUT_STOP:
                if ((nFlags & RTL_TEXTTOUNICODE_FLAGS_FLUSH) != 0) {
                    pSrcBufPtr = startOfCurrentChar;
                }
                [[fallthrough]];
            case sal::detail::textenc::BAD_INPUT_CONTINUE:
                nShift = -1;
                bCheckBom = false;
                break;

            case sal::detail::textenc::BAD_INPUT_NO_OUTPUT:
                nInfo |= RTL_TEXTTOUNICODE_INFO_DESTBUFFERTOOSMALL;
                break;
            }
    }

    if (pContext != nullptr)
    {
        static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nUtf32 = nUtf32;
        static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nBytes = nBytes;
        static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nShift = nShift;
        static_cast< ImplUtf8ToUnicodeContext * >(pContext)->bCheckBom = bCheckBom;
    }
    if (pInfo != nullptr)
        *pInfo = nInfo;
    if (pSrcCvtBytes != nullptr)
        *pSrcCvtBytes = reinterpret_cast< char const * >(pSrcBufPtr) - pSrcBuf;
    return pDestBufPtr - pDestBuf;
}

void * ImplCreateUnicodeToUtf8Context()
{
    ImplUnicodeToUtf8Context * p = new ImplUnicodeToUtf8Context;
    ImplResetUnicodeToUtf8Context(p);
    return p;
}

void ImplResetUnicodeToUtf8Context(void * pContext)
{
    if (pContext != nullptr)
        static_cast< ImplUnicodeToUtf8Context * >(pContext)->nHighSurrogate = 0xFFFF;
}

void ImplDestroyUnicodeToUtf8Context(void * pContext)
{
    delete static_cast< ImplUnicodeToUtf8Context * >(pContext);
}

sal_Size ImplConvertUnicodeToUtf8(
    void const * pData, void * pContext, sal_Unicode const * pSrcBuf,
    sal_Size nSrcChars, char * pDestBuf, sal_Size nDestBytes, sal_uInt32 nFlags,
    sal_uInt32 * pInfo, sal_Size * pSrcCvtChars)
{
    bool bJavaUtf8 = pData != nullptr;
    sal_Unicode nHighSurrogate = 0xFFFF;
    sal_uInt32 nInfo = 0;
    sal_Unicode const * pSrcBufPtr = pSrcBuf;
    sal_Unicode const * pSrcBufEnd = pSrcBufPtr + nSrcChars;
    char * pDestBufPtr = pDestBuf;
    char * pDestBufEnd = pDestBufPtr + nDestBytes;

    if (pContext != nullptr)
        nHighSurrogate
            = static_cast< ImplUnicodeToUtf8Context * >(pContext)->nHighSurrogate;

    if (nHighSurrogate == 0xFFFF)
    {
        if ((nFlags & RTL_UNICODETOTEXT_FLAGS_GLOBAL_SIGNATURE) != 0
            && !bJavaUtf8)
        {
            if (pDestBufEnd - pDestBufPtr >= 3)
            {
                /* Write BOM (U+FEFF) as UTF-8: */
                *pDestBufPtr++ = static_cast< char >(static_cast< unsigned char >(0xEF));
                *pDestBufPtr++ = static_cast< char >(static_cast< unsigned char >(0xBB));
                *pDestBufPtr++ = static_cast< char >(static_cast< unsigned char >(0xBF));
            }
            else
            {
                nInfo |= RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL;
                goto done;
            }
        }
        nHighSurrogate = 0;
    }

    while (pSrcBufPtr < pSrcBufEnd)
    {
        sal_uInt32 nChar = *pSrcBufPtr++;
        if (nHighSurrogate == 0)
        {
            if (rtl::isHighSurrogate(nChar) && !bJavaUtf8)
            {
                nHighSurrogate = static_cast<sal_Unicode>(nChar);
                continue;
            }
            else if (rtl::isLowSurrogate(nChar) && !bJavaUtf8)
            {
                goto bad_input;
            }
        }
        else if (rtl::isLowSurrogate(nChar) && !bJavaUtf8)
            nChar = rtl::combineSurrogates(nHighSurrogate, nChar);
        else
            goto bad_input;

        assert(bJavaUtf8 ? nChar <= 0xFFFF : rtl::isUnicodeScalarValue(nChar));

        if (nChar <= 0x7F && (!bJavaUtf8 || nChar != 0))
            if (pDestBufPtr != pDestBufEnd)
                *pDestBufPtr++ = static_cast< char >(nChar);
            else
                goto no_output;
        else if (nChar <= 0x7FF)
            if (pDestBufEnd - pDestBufPtr >= 2)
            {
                *pDestBufPtr++ = static_cast< char >(0xC0 | (nChar >> 6));
                *pDestBufPtr++ = static_cast< char >(0x80 | (nChar & 0x3F));
            }
            else
                goto no_output;
        else if (nChar <= 0xFFFF)
            if (pDestBufEnd - pDestBufPtr >= 3)
            {
                *pDestBufPtr++ = static_cast< char >(0xE0 | (nChar >> 12));
                *pDestBufPtr++ = static_cast< char >(0x80 | ((nChar >> 6) & 0x3F));
                *pDestBufPtr++ = static_cast< char >(0x80 | (nChar & 0x3F));
            }
            else
                goto no_output;
        else if (pDestBufEnd - pDestBufPtr >= 4)
        {
            *pDestBufPtr++ = static_cast< char >(0xF0 | (nChar >> 18));
            *pDestBufPtr++ = static_cast< char >(0x80 | ((nChar >> 12) & 0x3F));
            *pDestBufPtr++ = static_cast< char >(0x80 | ((nChar >> 6) & 0x3F));
            *pDestBufPtr++ = static_cast< char >(0x80 | (nChar & 0x3F));
        }
        else
            goto no_output;
        nHighSurrogate = 0;
        continue;

    bad_input:
        switch (sal::detail::textenc::handleBadInputUnicodeToTextConversion(
                    false, 0, nFlags, &pDestBufPtr, pDestBufEnd, &nInfo, nullptr,
                    0, nullptr))
        {
        case sal::detail::textenc::BAD_INPUT_STOP:
            nHighSurrogate = 0;
            break;

        case sal::detail::textenc::BAD_INPUT_CONTINUE:
            nHighSurrogate = 0;
            continue;

        case sal::detail::textenc::BAD_INPUT_NO_OUTPUT:
            goto no_output;
        }
        break;

    no_output:
        --pSrcBufPtr;
        nInfo |= RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL;
        break;
    }

    if (nHighSurrogate != 0
        && (nInfo & (RTL_UNICODETOTEXT_INFO_ERROR
                         | RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL))
               == 0)
    {
        if ((nFlags & RTL_UNICODETOTEXT_FLAGS_FLUSH) != 0)
            nInfo |= RTL_UNICODETOTEXT_INFO_SRCBUFFERTOSMALL;
        else
            switch (sal::detail::textenc::handleBadInputUnicodeToTextConversion(
                        false, 0, nFlags, &pDestBufPtr, pDestBufEnd, &nInfo,
                        nullptr, 0, nullptr))
            {
            case sal::detail::textenc::BAD_INPUT_STOP:
            case sal::detail::textenc::BAD_INPUT_CONTINUE:
                nHighSurrogate = 0;
                break;

            case sal::detail::textenc::BAD_INPUT_NO_OUTPUT:
                nInfo |= RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL;
                break;
            }
    }

 done:
    if (pContext != nullptr)
        static_cast< ImplUnicodeToUtf8Context * >(pContext)->nHighSurrogate
            = nHighSurrogate;
    if (pInfo != nullptr)
        *pInfo = nInfo;
    if (pSrcCvtChars != nullptr)
        *pSrcCvtChars = pSrcBufPtr - pSrcBuf;
    return pDestBufPtr - pDestBuf;
}

/* vim:set shiftwidth=4 softtabstop=4 expandtab: */

Coverage Report

Created: 2025-07-07 10:01

Line	Count	Source (jump to first uncovered line)
1		/* -- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -- */
2		/*
3		* This file is part of the LibreOffice project.
4		*
5		* This Source Code Form is subject to the terms of the Mozilla Public
6		* License, v. 2.0. If a copy of the MPL was not distributed with this
7		* file, You can obtain one at http://mozilla.org/MPL/2.0/.
8		*
9		* This file incorporates work covered by the following license notice:
10		*
11		* Licensed to the Apache Software Foundation (ASF) under one or more
12		* contributor license agreements. See the NOTICE file distributed
13		* with this work for additional information regarding copyright
14		* ownership. The ASF licenses this file to you under the Apache
15		* License, Version 2.0 (the "License"); you may not use this file
16		* except in compliance with the License. You may obtain a copy of
17		* the License at http://www.apache.org/licenses/LICENSE-2.0 .
18		*/
19
20		#include <sal/config.h>
21
22		#include <cassert>
23
24		#include <sal/types.h>
25		#include <rtl/character.hxx>
26		#include <rtl/textcvt.h>
27
28		#include "converter.hxx"
29		#include "tcvtutf8.hxx"
30
31		namespace {
32
33		struct ImplUtf8ToUnicodeContext
34		{
35		sal_uInt32 nUtf32;
36		int nBytes;
37		int nShift;
38		bool bCheckBom;
39		};
40
41		struct ImplUnicodeToUtf8Context
42		{
43		sal_Unicode nHighSurrogate; /* 0xFFFF: write BOM */
44		};
45
46		}
47
48		void * ImplCreateUtf8ToUnicodeContext()
49	115k	{
50	115k	ImplUtf8ToUnicodeContext * p = new ImplUtf8ToUnicodeContext;
51	115k	ImplResetUtf8ToUnicodeContext(p);
52	115k	return p;
53	115k	}
54
55		void ImplResetUtf8ToUnicodeContext(void * pContext)
56	191k	{
57	191k	if (pContext != nullptr)
58	191k	{
59	191k	static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nBytes = 1;
60	191k	static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nShift = -1;
61	191k	static_cast< ImplUtf8ToUnicodeContext * >(pContext)->bCheckBom = true;
62	191k	}
63	191k	}
64
65		void ImplDestroyUtf8ToUnicodeContext(void * pContext)
66	115k	{
67	115k	delete static_cast< ImplUtf8ToUnicodeContext * >(pContext);
68	115k	}
69
70		sal_Size ImplConvertUtf8ToUnicode(
71		void const * pData, void * pContext, char const * pSrcBuf,
72		sal_Size nSrcBytes, sal_Unicode * pDestBuf, sal_Size nDestChars,
73		sal_uInt32 nFlags, sal_uInt32 * pInfo, sal_Size * pSrcCvtBytes)
74	95.2M	{
75	95.2M	bool bJavaUtf8 = pData != nullptr;
76	95.2M	sal_uInt32 nUtf32 = 0;
77	95.2M	int nBytes = 1;
78	95.2M	int nShift = -1;
79	95.2M	bool bCheckBom = true;
80	95.2M	sal_uInt32 nInfo = 0;
81	95.2M	unsigned char const * pSrcBufPtr = reinterpret_cast<unsigned char const *>(pSrcBuf);
82	95.2M	unsigned char const * pSrcBufEnd = pSrcBufPtr + nSrcBytes;
83	95.2M	sal_Unicode * pDestBufPtr = pDestBuf;
84	95.2M	sal_Unicode * pDestBufEnd = pDestBufPtr + nDestChars;
85	95.2M	unsigned char const * startOfCurrentChar = pSrcBufPtr;
86
87	95.2M	if (pContext != nullptr)
88	92.5M	{
89	92.5M	nUtf32 = static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nUtf32;
90	92.5M	nBytes = static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nBytes;
91	92.5M	nShift = static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nShift;
92	92.5M	bCheckBom = static_cast< ImplUtf8ToUnicodeContext * >(pContext)->bCheckBom;
93	92.5M	}
94
95	1.42G	while (pSrcBufPtr < pSrcBufEnd)
96	1.32G	{
97	1.32G	bool bConsume = true;
98	1.32G	sal_uInt32 nChar = *pSrcBufPtr++;
99	1.32G	if (nShift < 0)
100		// Allow (illegal) 5 and 6 byte sequences, so they are read as a
101		// single individual bad character:
102	1.10G	if (nChar <= 0x7F)
103	843M	{
104	843M	nUtf32 = nChar;
105	843M	nBytes = 1;
106	843M	goto transform;
107	843M	}
108	260M	else if (nChar <= 0xBF)
109	80.4M	goto bad_input;
110	179M	else if (nChar <= 0xDF)
111	55.2M	{
112	55.2M	nUtf32 = (nChar & 0x1F) << 6;
113	55.2M	nBytes = 2;
114	55.2M	nShift = 0;
115	55.2M	}
116	124M	else if (nChar <= 0xEF)
117	76.8M	{
118	76.8M	nUtf32 = (nChar & 0x0F) << 12;
119	76.8M	nBytes = 3;
120	76.8M	nShift = 6;
121	76.8M	}
122	47.5M	else if (nChar <= 0xF7)
123	14.5M	{
124	14.5M	nUtf32 = (nChar & 0x07) << 18;
125	14.5M	nBytes = 4;
126	14.5M	nShift = 12;
127	14.5M	}
128	32.9M	else if (nChar <= 0xFB)
129	5.56M	{
130	5.56M	nUtf32 = (nChar & 0x03) << 24;
131	5.56M	nBytes = 5;
132	5.56M	nShift = 18;
133	5.56M	}
134	27.4M	else if (nChar <= 0xFD)
135	2.99M	{
136	2.99M	nUtf32 = (nChar & 0x01) << 30;
137	2.99M	nBytes = 6;
138	2.99M	nShift = 24;
139	2.99M	}
140	24.4M	else
141	24.4M	goto bad_input;
142	225M	else if ((nChar & 0xC0) == 0x80)
143	144M	{
144	144M	nUtf32 \|= (nChar & 0x3F) << nShift;
145	144M	if (nShift == 0)
146	74.3M	goto transform;
147	70.5M	else
148	70.5M	nShift -= 6;
149	144M	}
150	80.8M	else
151	80.8M	{
152		/*
153		This byte is preceded by a broken UTF-8 sequence; if this byte
154		is neither in the range [0x80..0xBF] nor in the range
155		[0xFE..0xFF], assume that this byte does not belong to that
156		broken sequence, but instead starts a new, legal UTF-8 sequence:
157		*/
158	80.8M	bConsume = nChar >= 0xFE;
159	80.8M	goto bad_input;
160	80.8M	}
161	225M	continue;
162
163	918M	transform:
164	918M	if (!bCheckBom \|\| nUtf32 != 0xFEFF \|\| nBytes != 3
165	918M	\|\| (nFlags & RTL_TEXTTOUNICODE_FLAGS_GLOBAL_SIGNATURE) == 0
166	918M	\|\| bJavaUtf8)
167	918M	{
168	918M	switch (nBytes) {
169	843M	case 1:
170	843M	if (bJavaUtf8 && nUtf32 == 0) {
171	14.5k	goto bad_input;
172	14.5k	}
173	843M	break;
174	843M	case 2:
175	17.4M	if (nUtf32 < 0x80 && !(bJavaUtf8 && nUtf32 == 0)) {
176	710k	goto bad_input;
177	710k	}
178	16.7M	break;
179	53.3M	case 3:
180	53.3M	if (nUtf32 < 0x800 \|\| (!bJavaUtf8 && rtl::isSurrogate(nUtf32)))
181	175k	{
182	175k	goto bad_input;
183	175k	}
184	53.1M	break;
185	53.1M	case 4:
186	3.47M	if (nUtf32 < 0x10000 \|\| !rtl::isUnicodeCodePoint(nUtf32)
187	3.47M	\|\| bJavaUtf8)
188	219k	{
189	219k	goto bad_input;
190	219k	}
191	3.25M	break;
192	3.25M	default:
193	37.3k	goto bad_input;
194	918M	}
195	916M	if (nUtf32 <= 0xFFFF)
196	913M	if (pDestBufPtr != pDestBufEnd)
197	913M	*pDestBufPtr++ = static_cast<sal_Unicode>(nUtf32);
198	181k	else
199	181k	goto no_output;
200	3.25M	else if (pDestBufEnd - pDestBufPtr >= 2)
201	3.21M	pDestBufPtr += rtl::splitSurrogates(nUtf32, pDestBufPtr);
202	36.5k	else
203	36.5k	goto no_output;
204	916M	}
205	916M	nShift = -1;
206	916M	bCheckBom = false;
207	916M	startOfCurrentChar = pSrcBufPtr;
208	916M	continue;
209
210	186M	bad_input:
211	186M	switch (sal::detail::textenc::handleBadInputTextToUnicodeConversion(
212	186M	false, nBytes != 1, 0, nFlags, &pDestBufPtr, pDestBufEnd,
213	186M	&nInfo))
214	186M	{
215	1.19M	case sal::detail::textenc::BAD_INPUT_STOP:
216	1.19M	nShift = -1;
217	1.19M	bCheckBom = false;
218	1.19M	if ((nFlags & RTL_TEXTTOUNICODE_FLAGS_FLUSH) == 0) {
219	1.19M	if (!bConsume)
220	315k	--pSrcBufPtr;
221	1.19M	} else {
222	0	pSrcBufPtr = startOfCurrentChar;
223	0	}
224	1.19M	break;
225
226	185M	case sal::detail::textenc::BAD_INPUT_CONTINUE:
227	185M	nShift = -1;
228	185M	bCheckBom = false;
229	185M	if (!bConsume)
230	79.6M	--pSrcBufPtr;
231	185M	startOfCurrentChar = pSrcBufPtr;
232	185M	continue;
233
234	28.6k	case sal::detail::textenc::BAD_INPUT_NO_OUTPUT:
235	28.6k	goto no_output;
236	186M	}
237	1.19M	break;
238
239	1.19M	no_output:
240	246k	--pSrcBufPtr;
241	246k	nInfo \|= RTL_TEXTTOUNICODE_INFO_DESTBUFFERTOOSMALL;
242	246k	break;
243	186M	}
244
245	95.2M	if (nShift >= 0
246	95.2M	&& (nInfo & (RTL_TEXTTOUNICODE_INFO_ERROR
247	949k	\| RTL_TEXTTOUNICODE_INFO_DESTBUFFERTOOSMALL))
248	949k	== 0)
249	897k	{
250	897k	if ((nFlags & RTL_TEXTTOUNICODE_FLAGS_FLUSH) == 0)
251	896k	nInfo \|= RTL_TEXTTOUNICODE_INFO_SRCBUFFERTOOSMALL;
252	1.55k	else
253	1.55k	switch (sal::detail::textenc::handleBadInputTextToUnicodeConversion(
254	1.55k	false, true, 0, nFlags, &pDestBufPtr, pDestBufEnd,
255	1.55k	&nInfo))
256	1.55k	{
257	0	case sal::detail::textenc::BAD_INPUT_STOP:
258	0	if ((nFlags & RTL_TEXTTOUNICODE_FLAGS_FLUSH) != 0) {
259	0	pSrcBufPtr = startOfCurrentChar;
260	0	}
261	0	[[fallthrough]];
262	1.39k	case sal::detail::textenc::BAD_INPUT_CONTINUE:
263	1.39k	nShift = -1;
264	1.39k	bCheckBom = false;
265	1.39k	break;
266
267	160	case sal::detail::textenc::BAD_INPUT_NO_OUTPUT:
268	160	nInfo \|= RTL_TEXTTOUNICODE_INFO_DESTBUFFERTOOSMALL;
269	160	break;
270	1.55k	}
271	897k	}
272
273	95.2M	if (pContext != nullptr)
274	92.5M	{
275	92.5M	static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nUtf32 = nUtf32;
276	92.5M	static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nBytes = nBytes;
277	92.5M	static_cast< ImplUtf8ToUnicodeContext * >(pContext)->nShift = nShift;
278	92.5M	static_cast< ImplUtf8ToUnicodeContext * >(pContext)->bCheckBom = bCheckBom;
279	92.5M	}
280	95.2M	if (pInfo != nullptr)
281	95.2M	*pInfo = nInfo;
282	95.2M	if (pSrcCvtBytes != nullptr)
283	95.2M	pSrcCvtBytes = reinterpret_cast< char const >(pSrcBufPtr) - pSrcBuf;
284	95.2M	return pDestBufPtr - pDestBuf;
285	95.2M	}
286
287		void * ImplCreateUnicodeToUtf8Context()
288	32.9k	{
289	32.9k	ImplUnicodeToUtf8Context * p = new ImplUnicodeToUtf8Context;
290	32.9k	ImplResetUnicodeToUtf8Context(p);
291	32.9k	return p;
292	32.9k	}
293
294		void ImplResetUnicodeToUtf8Context(void * pContext)
295	32.9k	{
296	32.9k	if (pContext != nullptr)
297	32.9k	static_cast< ImplUnicodeToUtf8Context * >(pContext)->nHighSurrogate = 0xFFFF;
298	32.9k	}
299
300		void ImplDestroyUnicodeToUtf8Context(void * pContext)
301	32.9k	{
302	32.9k	delete static_cast< ImplUnicodeToUtf8Context * >(pContext);
303	32.9k	}
304
305		sal_Size ImplConvertUnicodeToUtf8(
306		void const * pData, void * pContext, sal_Unicode const * pSrcBuf,
307		sal_Size nSrcChars, char * pDestBuf, sal_Size nDestBytes, sal_uInt32 nFlags,
308		sal_uInt32 * pInfo, sal_Size * pSrcCvtChars)
309	255k	{
310	255k	bool bJavaUtf8 = pData != nullptr;
311	255k	sal_Unicode nHighSurrogate = 0xFFFF;
312	255k	sal_uInt32 nInfo = 0;
313	255k	sal_Unicode const * pSrcBufPtr = pSrcBuf;
314	255k	sal_Unicode const * pSrcBufEnd = pSrcBufPtr + nSrcChars;
315	255k	char * pDestBufPtr = pDestBuf;
316	255k	char * pDestBufEnd = pDestBufPtr + nDestBytes;
317
318	255k	if (pContext != nullptr)
319	43.3k	nHighSurrogate
320	43.3k	= static_cast< ImplUnicodeToUtf8Context * >(pContext)->nHighSurrogate;
321
322	255k	if (nHighSurrogate == 0xFFFF)
323	242k	{
324	242k	if ((nFlags & RTL_UNICODETOTEXT_FLAGS_GLOBAL_SIGNATURE) != 0
325	242k	&& !bJavaUtf8)
326	0	{
327	0	if (pDestBufEnd - pDestBufPtr >= 3)
328	0	{
329		/* Write BOM (U+FEFF) as UTF-8: */
330	0	*pDestBufPtr++ = static_cast< char >(static_cast< unsigned char >(0xEF));
331	0	*pDestBufPtr++ = static_cast< char >(static_cast< unsigned char >(0xBB));
332	0	*pDestBufPtr++ = static_cast< char >(static_cast< unsigned char >(0xBF));
333	0	}
334	0	else
335	0	{
336	0	nInfo \|= RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL;
337	0	goto done;
338	0	}
339	0	}
340	242k	nHighSurrogate = 0;
341	242k	}
342
343	385M	while (pSrcBufPtr < pSrcBufEnd)
344	385M	{
345	385M	sal_uInt32 nChar = *pSrcBufPtr++;
346	385M	if (nHighSurrogate == 0)
347	385M	{
348	385M	if (rtl::isHighSurrogate(nChar) && !bJavaUtf8)
349	30.8k	{
350	30.8k	nHighSurrogate = static_cast<sal_Unicode>(nChar);
351	30.8k	continue;
352	30.8k	}
353	385M	else if (rtl::isLowSurrogate(nChar) && !bJavaUtf8)
354	28.8k	{
355	28.8k	goto bad_input;
356	28.8k	}
357	385M	}
358	30.8k	else if (rtl::isLowSurrogate(nChar) && !bJavaUtf8)
359	7.48k	nChar = rtl::combineSurrogates(nHighSurrogate, nChar);
360	23.3k	else
361	23.3k	goto bad_input;
362
363	385M	assert(bJavaUtf8 ? nChar <= 0xFFFF : rtl::isUnicodeScalarValue(nChar));
364
365	385M	if (nChar <= 0x7F && (!bJavaUtf8 \|\| nChar != 0))
366	342M	if (pDestBufPtr != pDestBufEnd)
367	342M	*pDestBufPtr++ = static_cast< char >(nChar);
368	0	else
369	0	goto no_output;
370	42.3M	else if (nChar <= 0x7FF)
371	3.51M	if (pDestBufEnd - pDestBufPtr >= 2)
372	3.51M	{
373	3.51M	*pDestBufPtr++ = static_cast< char >(0xC0 \| (nChar >> 6));
374	3.51M	*pDestBufPtr++ = static_cast< char >(0x80 \| (nChar & 0x3F));
375	3.51M	}
376	0	else
377	0	goto no_output;
378	38.8M	else if (nChar <= 0xFFFF)
379	38.8M	if (pDestBufEnd - pDestBufPtr >= 3)
380	38.8M	{
381	38.8M	*pDestBufPtr++ = static_cast< char >(0xE0 \| (nChar >> 12));
382	38.8M	*pDestBufPtr++ = static_cast< char >(0x80 \| ((nChar >> 6) & 0x3F));
383	38.8M	*pDestBufPtr++ = static_cast< char >(0x80 \| (nChar & 0x3F));
384	38.8M	}
385	0	else
386	0	goto no_output;
387	7.48k	else if (pDestBufEnd - pDestBufPtr >= 4)
388	7.48k	{
389	7.48k	*pDestBufPtr++ = static_cast< char >(0xF0 \| (nChar >> 18));
390	7.48k	*pDestBufPtr++ = static_cast< char >(0x80 \| ((nChar >> 12) & 0x3F));
391	7.48k	*pDestBufPtr++ = static_cast< char >(0x80 \| ((nChar >> 6) & 0x3F));
392	7.48k	*pDestBufPtr++ = static_cast< char >(0x80 \| (nChar & 0x3F));
393	7.48k	}
394	0	else
395	0	goto no_output;
396	385M	nHighSurrogate = 0;
397	385M	continue;
398
399	52.2k	bad_input:
400	52.2k	switch (sal::detail::textenc::handleBadInputUnicodeToTextConversion(
401	52.2k	false, 0, nFlags, &pDestBufPtr, pDestBufEnd, &nInfo, nullptr,
402	52.2k	0, nullptr))
403	52.2k	{
404	0	case sal::detail::textenc::BAD_INPUT_STOP:
405	0	nHighSurrogate = 0;
406	0	break;
407
408	52.2k	case sal::detail::textenc::BAD_INPUT_CONTINUE:
409	52.2k	nHighSurrogate = 0;
410	52.2k	continue;
411
412	0	case sal::detail::textenc::BAD_INPUT_NO_OUTPUT:
413	0	goto no_output;
414	52.2k	}
415	0	break;
416
417	0	no_output:
418	0	--pSrcBufPtr;
419	0	nInfo \|= RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL;
420	0	break;
421	52.2k	}
422
423	255k	if (nHighSurrogate != 0
424	255k	&& (nInfo & (RTL_UNICODETOTEXT_INFO_ERROR
425	35	\| RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL))
426	35	== 0)
427	35	{
428	35	if ((nFlags & RTL_UNICODETOTEXT_FLAGS_FLUSH) != 0)
429	22	nInfo \|= RTL_UNICODETOTEXT_INFO_SRCBUFFERTOSMALL;
430	13	else
431	13	switch (sal::detail::textenc::handleBadInputUnicodeToTextConversion(
432	13	false, 0, nFlags, &pDestBufPtr, pDestBufEnd, &nInfo,
433	13	nullptr, 0, nullptr))
434	13	{
435	0	case sal::detail::textenc::BAD_INPUT_STOP:
436	13	case sal::detail::textenc::BAD_INPUT_CONTINUE:
437	13	nHighSurrogate = 0;
438	13	break;
439
440	0	case sal::detail::textenc::BAD_INPUT_NO_OUTPUT:
441	0	nInfo \|= RTL_UNICODETOTEXT_INFO_DESTBUFFERTOSMALL;
442	0	break;
443	13	}
444	35	}
445
446	255k	done:
447	255k	if (pContext != nullptr)
448	43.3k	static_cast< ImplUnicodeToUtf8Context * >(pContext)->nHighSurrogate
449	43.3k	= nHighSurrogate;
450	255k	if (pInfo != nullptr)
451	255k	*pInfo = nInfo;
452	255k	if (pSrcCvtChars != nullptr)
453	255k	*pSrcCvtChars = pSrcBufPtr - pSrcBuf;
454	255k	return pDestBufPtr - pDestBuf;
455	255k	}
456
457		/* vim:set shiftwidth=4 softtabstop=4 expandtab: */