filter/source/msfilter/rtfutil.cxx

   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
  10 #include <filter/msfilter/rtfutil.hxx>
  11 #include <rtl/strbuf.hxx>
  12 #include <osl/diagnose.h>
  13 #include <svtools/rtfkeywd.hxx>
  14
  15 namespace msfilter
  16 {
  17 namespace rtfutil
  18 {
  19
  20 OString OutHex(sal_uLong nHex, sal_uInt8 nLen)
  21 {
  22     sal_Char aNToABuf[] = "0000000000000000";
  23
  24     OSL_ENSURE(nLen < sizeof(aNToABuf), "nLen is too big");
  25     if (nLen >= sizeof(aNToABuf))
  26         nLen = (sizeof(aNToABuf)-1);
  27
  28     // Set pointer to the buffer end
  29     sal_Char* pStr = aNToABuf + (sizeof(aNToABuf)-1);
  30     for (sal_uInt8 n = 0; n < nLen; ++n)
  31     {
  32         *(--pStr) = (sal_Char)(nHex & 0xf) + 48;
  33         if (*pStr > '9')
  34             *pStr += 39;
  35         nHex >>= 4;
  36     }
  37     return OString(pStr);
  38 }
  39
  40 // Ideally, this function should work on (sal_uInt32) Unicode scalar values
  41 // instead of (sal_Unicode) UTF-16 code units.  However, at least "Rich Text
  42 // Format (RTF) Specification Version 1.9.1" available at
  43 // <https://www.microsoft.com/en-us/download/details.aspx?id=10725> does not
  44 // look like it allows non-BMP Unicode characters >= 0x10000 in the \uN notation
  45 // (it only talks about "Unicode character", but then explains how values of N
  46 // greater than 32767 will be expressed as negative signed 16-bit numbers, so
  47 // that smells like \uN is limited to BMP).
  48 // However the "Mathematics" section has an example that shows the code point
  49 // U+1D44E being encoded as UTF-16 surrogate pair "\u-10187?\u-9138?", so
  50 // sal_Unicode actually works fine here.
  51 OString OutChar(sal_Unicode c, int* pUCMode, rtl_TextEncoding eDestEnc, bool* pSuccess, bool bUnicode)
  52 {
  53     if (pSuccess)
  54         *pSuccess = true;
  55     OStringBuffer aBuf;
  56     const sal_Char* pStr = nullptr;
  57     // 0x0b instead of \n, etc because of the replacements in SwWW8AttrIter::GetSnippet()
  58     switch (c)
  59     {
  60     case 0x0b:
  61         // hard line break
  62         pStr = OOO_STRING_SVTOOLS_RTF_LINE;
  63         break;
  64     case '\t':
  65         pStr = OOO_STRING_SVTOOLS_RTF_TAB;
  66         break;
  67     case '\\':
  68     case '}':
  69     case '{':
  70         aBuf.append('\\');
  71         aBuf.append((sal_Char)c);
  72         break;
  73     case 0xa0:
  74         // non-breaking space
  75         pStr = "\\~";
  76         break;
  77     case 0x1e:
  78         // non-breaking hyphen
  79         pStr = "\\_";
  80         break;
  81     case 0x1f:
  82         // optional hyphen
  83         pStr = "\\-";
  84         break;
  85     default:
  86         if (c >= ' ' && c <= '~')
  87             aBuf.append((sal_Char)c);
  88         else
  89         {
  90             OUString sBuf(&c, 1);
  91             OString sConverted;
  92             if (pSuccess)
  93                 *pSuccess &= sBuf.convertToString(&sConverted, eDestEnc, RTL_UNICODETOTEXT_FLAGS_UNDEFINED_ERROR | RTL_UNICODETOTEXT_FLAGS_INVALID_ERROR);
  94             else
  95                 sBuf.convertToString(&sConverted, eDestEnc, OUSTRING_TO_OSTRING_CVTFLAGS);
  96             const sal_Int32 nLen = sConverted.getLength();
  97
  98             if (pUCMode && bUnicode)
  99             {
 100                 if (*pUCMode != nLen)
 101                 {
 102                     aBuf.append("\\uc");
 103                     aBuf.append((sal_Int32)nLen);
 104                     // #i47831# add an additional whitespace, so that "document whitespaces" are not ignored.
 105                     aBuf.append(' ');
 106                     *pUCMode = nLen;
 107                 }
 108                 aBuf.append("\\u");
 109                 aBuf.append((sal_Int32)c);
 110             }
 111
 112             for (sal_Int32 nI = 0; nI < nLen; ++nI)
 113             {
 114                 aBuf.append("\\'");
 115                 aBuf.append(OutHex(sConverted[nI], 2));
 116             }
 117         }
 118     }
 119     if (pStr)
 120     {
 121         aBuf.append(pStr);
 122         switch (c)
 123         {
 124         case 0xa0:
 125         case 0x1e:
 126         case 0x1f:
 127             break;
 128         default:
 129             aBuf.append(' ');
 130         }
 131     }
 132     return aBuf.makeStringAndClear();
 133 }
 134
 135 OString OutString(const OUString& rStr, rtl_TextEncoding eDestEnc, bool bUnicode)
 136 {
 137     SAL_INFO("filter.ms", OSL_THIS_FUNC << ", rStr = '" << OUString(rStr) << "'");
 138     OStringBuffer aBuf;
 139     int nUCMode = 1;
 140     for (sal_Int32 n = 0; n < rStr.getLength(); ++n)
 141         aBuf.append(OutChar(rStr[n], &nUCMode, eDestEnc, nullptr, bUnicode));
 142     if (nUCMode != 1)
 143     {
 144         aBuf.append(OOO_STRING_SVTOOLS_RTF_UC);
 145         aBuf.append((sal_Int32)1);
 146         aBuf.append(" "); // #i47831# add an additional whitespace, so that "document whitespaces" are not ignored.;
 147     }
 148     return aBuf.makeStringAndClear();
 149 }
 150
 151 /// Checks if lossless conversion of the string to eDestEnc is possible or not.
 152 static bool TryOutString(const OUString& rStr, rtl_TextEncoding eDestEnc)
 153 {
 154     int nUCMode = 1;
 155     for (sal_Int32 n = 0; n < rStr.getLength(); ++n)
 156     {
 157         bool bRet;
 158         OutChar(rStr[n], &nUCMode, eDestEnc, &bRet);
 159         if (!bRet)
 160             return false;
 161     }
 162     return true;
 163 }
 164
 165 OString OutStringUpr(const sal_Char* pToken, const OUString& rStr, rtl_TextEncoding eDestEnc)
 166 {
 167     if (TryOutString(rStr, eDestEnc))
 168         return OString("{") + pToken + " " + OutString(rStr, eDestEnc) + "}";
 169
 170     OStringBuffer aRet;
 171     aRet.append("{" OOO_STRING_SVTOOLS_RTF_UPR "{");
 172     aRet.append(pToken);
 173     aRet.append(" ");
 174     aRet.append(OutString(rStr, eDestEnc, /*bUnicode =*/ false));
 175     aRet.append("}{" OOO_STRING_SVTOOLS_RTF_IGNORE OOO_STRING_SVTOOLS_RTF_UD "{");
 176     aRet.append(pToken);
 177     aRet.append(" ");
 178     aRet.append(OutString(rStr, eDestEnc));
 179     aRet.append("}}}");
 180     return aRet.makeStringAndClear();
 181 }
 182 }
 183 }
 184
 185 /* vim:set shiftwidth=4 softtabstop=4 expandtab: */