| 1 | /* $Id: string.cpp 55942 2015-05-19 23:00:07Z vboxsync $ */
|
|---|
| 2 | /** @file
|
|---|
| 3 | * MS COM / XPCOM Abstraction Layer - UTF-8 and UTF-16 string classes.
|
|---|
| 4 | */
|
|---|
| 5 |
|
|---|
| 6 | /*
|
|---|
| 7 | * Copyright (C) 2006-2012 Oracle Corporation
|
|---|
| 8 | *
|
|---|
| 9 | * This file is part of VirtualBox Open Source Edition (OSE), as
|
|---|
| 10 | * available from http://www.virtualbox.org. This file is free software;
|
|---|
| 11 | * you can redistribute it and/or modify it under the terms of the GNU
|
|---|
| 12 | * General Public License (GPL) as published by the Free Software
|
|---|
| 13 | * Foundation, in version 2 as it comes in the "COPYING" file of the
|
|---|
| 14 | * VirtualBox OSE distribution. VirtualBox OSE is distributed in the
|
|---|
| 15 | * hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
|
|---|
| 16 | */
|
|---|
| 17 |
|
|---|
| 18 | #include "VBox/com/string.h"
|
|---|
| 19 |
|
|---|
| 20 | #include <iprt/err.h>
|
|---|
| 21 | #include <iprt/path.h>
|
|---|
| 22 | #include <iprt/log.h>
|
|---|
| 23 |
|
|---|
| 24 | namespace com
|
|---|
| 25 | {
|
|---|
| 26 |
|
|---|
| 27 | // BSTR representing a null wide char with 32 bits of length prefix (0);
|
|---|
| 28 | // this will work on Windows as well as other platforms where BSTR does
|
|---|
| 29 | // not use length prefixes
|
|---|
| 30 | const OLECHAR g_achEmptyBstr[3] = { 0, 0, 0 };
|
|---|
| 31 | const BSTR g_bstrEmpty = (BSTR)&g_achEmptyBstr[2];
|
|---|
| 32 |
|
|---|
| 33 | /* static */
|
|---|
| 34 | const Bstr Bstr::Empty; /* default ctor is OK */
|
|---|
| 35 |
|
|---|
| 36 | void Bstr::copyFromN(const char *a_pszSrc, size_t a_cchMax)
|
|---|
| 37 | {
|
|---|
| 38 | /*
|
|---|
| 39 | * Initialie m_bstr first in case of throws further down in the code, then
|
|---|
| 40 | * check for empty input (m_bstr == NULL means empty, there are no NULL
|
|---|
| 41 | * strings).
|
|---|
| 42 | */
|
|---|
| 43 | m_bstr = NULL;
|
|---|
| 44 | if (!a_cchMax || !a_pszSrc || !*a_pszSrc)
|
|---|
| 45 | return;
|
|---|
| 46 |
|
|---|
| 47 | /*
|
|---|
| 48 | * Calculate the length and allocate a BSTR string buffer of the right
|
|---|
| 49 | * size, i.e. optimize heap usage.
|
|---|
| 50 | */
|
|---|
| 51 | size_t cwc;
|
|---|
| 52 | int vrc = ::RTStrCalcUtf16LenEx(a_pszSrc, a_cchMax, &cwc);
|
|---|
| 53 | if (RT_SUCCESS(vrc))
|
|---|
| 54 | {
|
|---|
| 55 | m_bstr = ::SysAllocStringByteLen(NULL, (unsigned)(cwc * sizeof(OLECHAR)));
|
|---|
| 56 | if (RT_LIKELY(m_bstr))
|
|---|
| 57 | {
|
|---|
| 58 | PRTUTF16 pwsz = (PRTUTF16)m_bstr;
|
|---|
| 59 | vrc = ::RTStrToUtf16Ex(a_pszSrc, a_cchMax, &pwsz, cwc + 1, NULL);
|
|---|
| 60 | if (RT_SUCCESS(vrc))
|
|---|
| 61 | return;
|
|---|
| 62 |
|
|---|
| 63 | /* This should not happen! */
|
|---|
| 64 | AssertRC(vrc);
|
|---|
| 65 | cleanup();
|
|---|
| 66 | }
|
|---|
| 67 | }
|
|---|
| 68 | else /* ASSUME: input is valid Utf-8. Fake out of memory error. */
|
|---|
| 69 | AssertLogRelMsgFailed(("%Rrc %.*Rhxs\n", vrc, RTStrNLen(a_pszSrc, a_cchMax), a_pszSrc));
|
|---|
| 70 | throw std::bad_alloc();
|
|---|
| 71 | }
|
|---|
| 72 |
|
|---|
| 73 |
|
|---|
| 74 | /* static */
|
|---|
| 75 | const Utf8Str Utf8Str::Empty; /* default ctor is OK */
|
|---|
| 76 |
|
|---|
| 77 | #if defined(VBOX_WITH_XPCOM)
|
|---|
| 78 | void Utf8Str::cloneTo(char **pstr) const
|
|---|
| 79 | {
|
|---|
| 80 | size_t cb = length() + 1;
|
|---|
| 81 | *pstr = (char *)nsMemory::Alloc(cb);
|
|---|
| 82 | if (RT_LIKELY(*pstr))
|
|---|
| 83 | memcpy(*pstr, c_str(), cb);
|
|---|
| 84 | else
|
|---|
| 85 | throw std::bad_alloc();
|
|---|
| 86 | }
|
|---|
| 87 |
|
|---|
| 88 | HRESULT Utf8Str::cloneToEx(char **pstr) const
|
|---|
| 89 | {
|
|---|
| 90 | size_t cb = length() + 1;
|
|---|
| 91 | *pstr = (char *)nsMemory::Alloc(cb);
|
|---|
| 92 | if (RT_LIKELY(*pstr))
|
|---|
| 93 | {
|
|---|
| 94 | memcpy(*pstr, c_str(), cb);
|
|---|
| 95 | return S_OK;
|
|---|
| 96 | }
|
|---|
| 97 | return E_OUTOFMEMORY;
|
|---|
| 98 | }
|
|---|
| 99 | #endif
|
|---|
| 100 |
|
|---|
| 101 | Utf8Str& Utf8Str::stripTrailingSlash()
|
|---|
| 102 | {
|
|---|
| 103 | if (length())
|
|---|
| 104 | {
|
|---|
| 105 | ::RTPathStripTrailingSlash(m_psz);
|
|---|
| 106 | jolt();
|
|---|
| 107 | }
|
|---|
| 108 | return *this;
|
|---|
| 109 | }
|
|---|
| 110 |
|
|---|
| 111 | Utf8Str& Utf8Str::stripFilename()
|
|---|
| 112 | {
|
|---|
| 113 | if (length())
|
|---|
| 114 | {
|
|---|
| 115 | RTPathStripFilename(m_psz);
|
|---|
| 116 | jolt();
|
|---|
| 117 | }
|
|---|
| 118 | return *this;
|
|---|
| 119 | }
|
|---|
| 120 |
|
|---|
| 121 | Utf8Str& Utf8Str::stripPath()
|
|---|
| 122 | {
|
|---|
| 123 | if (length())
|
|---|
| 124 | {
|
|---|
| 125 | char *pszName = ::RTPathFilename(m_psz);
|
|---|
| 126 | if (pszName)
|
|---|
| 127 | {
|
|---|
| 128 | size_t cchName = length() - (pszName - m_psz);
|
|---|
| 129 | memmove(m_psz, pszName, cchName + 1);
|
|---|
| 130 | jolt();
|
|---|
| 131 | }
|
|---|
| 132 | else
|
|---|
| 133 | cleanup();
|
|---|
| 134 | }
|
|---|
| 135 | return *this;
|
|---|
| 136 | }
|
|---|
| 137 |
|
|---|
| 138 | Utf8Str& Utf8Str::stripSuffix()
|
|---|
| 139 | {
|
|---|
| 140 | if (length())
|
|---|
| 141 | {
|
|---|
| 142 | RTPathStripSuffix(m_psz);
|
|---|
| 143 | jolt();
|
|---|
| 144 | }
|
|---|
| 145 | return *this;
|
|---|
| 146 | }
|
|---|
| 147 |
|
|---|
| 148 | size_t Utf8Str::parseKeyValue(Utf8Str &key, Utf8Str &value, size_t pos, const Utf8Str &pairSeparator, const Utf8Str &keyValueSeparator) const
|
|---|
| 149 | {
|
|---|
| 150 | size_t start = pos;
|
|---|
| 151 | while(start == (pos = find(pairSeparator.c_str(), pos)))
|
|---|
| 152 | start = ++pos;
|
|---|
| 153 |
|
|---|
| 154 | size_t kvSepPos = find(keyValueSeparator.c_str(), start);
|
|---|
| 155 | if (kvSepPos < pos)
|
|---|
| 156 | {
|
|---|
| 157 | key = substr(start, kvSepPos - start);
|
|---|
| 158 | value = substr(kvSepPos + 1, pos - kvSepPos - 1);
|
|---|
| 159 | }
|
|---|
| 160 | else
|
|---|
| 161 | {
|
|---|
| 162 | key = value = "";
|
|---|
| 163 | }
|
|---|
| 164 | return pos;
|
|---|
| 165 | }
|
|---|
| 166 |
|
|---|
| 167 | /**
|
|---|
| 168 | * Internal function used in Utf8Str copy constructors and assignment when
|
|---|
| 169 | * copying from a UTF-16 string.
|
|---|
| 170 | *
|
|---|
| 171 | * As with the RTCString::copyFrom() variants, this unconditionally sets the
|
|---|
| 172 | * members to a copy of the given other strings and makes no assumptions about
|
|---|
| 173 | * previous contents. This can therefore be used both in copy constructors,
|
|---|
| 174 | * when member variables have no defined value, and in assignments after having
|
|---|
| 175 | * called cleanup().
|
|---|
| 176 | *
|
|---|
| 177 | * This variant converts from a UTF-16 string, most probably from
|
|---|
| 178 | * a Bstr assignment.
|
|---|
| 179 | *
|
|---|
| 180 | * @param a_pbstr The source string. The caller guarantees that this
|
|---|
| 181 | * is valid UTF-16.
|
|---|
| 182 | * @param a_cwcMax The number of characters to be copied. If set to RTSTR_MAX,
|
|---|
| 183 | * the entire string will be copied.
|
|---|
| 184 | *
|
|---|
| 185 | * @sa RTCString::copyFromN
|
|---|
| 186 | */
|
|---|
| 187 | void Utf8Str::copyFrom(CBSTR a_pbstr, size_t a_cwcMax)
|
|---|
| 188 | {
|
|---|
| 189 | if (a_pbstr && *a_pbstr)
|
|---|
| 190 | {
|
|---|
| 191 | int vrc = RTUtf16ToUtf8Ex((PCRTUTF16)a_pbstr,
|
|---|
| 192 | a_cwcMax, // size_t cwcString: translate entire string
|
|---|
| 193 | &m_psz, // char **ppsz: output buffer
|
|---|
| 194 | 0, // size_t cch: if 0, func allocates buffer in *ppsz
|
|---|
| 195 | &m_cch); // size_t *pcch: receives the size of the output string, excluding the terminator.
|
|---|
| 196 | if (RT_SUCCESS(vrc))
|
|---|
| 197 | m_cbAllocated = m_cch + 1;
|
|---|
| 198 | else
|
|---|
| 199 | {
|
|---|
| 200 | if ( vrc != VERR_NO_STR_MEMORY
|
|---|
| 201 | && vrc != VERR_NO_MEMORY)
|
|---|
| 202 | {
|
|---|
| 203 | /* ASSUME: input is valid Utf-16. Fake out of memory error. */
|
|---|
| 204 | AssertLogRelMsgFailed(("%Rrc %.*Rhxs\n", vrc, RTUtf16Len(a_pbstr) * sizeof(RTUTF16), a_pbstr));
|
|---|
| 205 | }
|
|---|
| 206 |
|
|---|
| 207 | m_cch = 0;
|
|---|
| 208 | m_cbAllocated = 0;
|
|---|
| 209 | m_psz = NULL;
|
|---|
| 210 |
|
|---|
| 211 | throw std::bad_alloc();
|
|---|
| 212 | }
|
|---|
| 213 | }
|
|---|
| 214 | else
|
|---|
| 215 | {
|
|---|
| 216 | m_cch = 0;
|
|---|
| 217 | m_cbAllocated = 0;
|
|---|
| 218 | m_psz = NULL;
|
|---|
| 219 | }
|
|---|
| 220 | }
|
|---|
| 221 |
|
|---|
| 222 | /**
|
|---|
| 223 | * A variant of Utf8Str::copyFrom that does not throw any exceptions but returns
|
|---|
| 224 | * E_OUTOFMEMORY instead.
|
|---|
| 225 | *
|
|---|
| 226 | * @param a_pbstr The source string.
|
|---|
| 227 | * @returns S_OK or E_OUTOFMEMORY.
|
|---|
| 228 | */
|
|---|
| 229 | HRESULT Utf8Str::copyFromEx(CBSTR a_pbstr)
|
|---|
| 230 | {
|
|---|
| 231 | if (a_pbstr && *a_pbstr)
|
|---|
| 232 | {
|
|---|
| 233 | int vrc = RTUtf16ToUtf8Ex((PCRTUTF16)a_pbstr,
|
|---|
| 234 | RTSTR_MAX, // size_t cwcString: translate entire string
|
|---|
| 235 | &m_psz, // char **ppsz: output buffer
|
|---|
| 236 | 0, // size_t cch: if 0, func allocates buffer in *ppsz
|
|---|
| 237 | &m_cch); // size_t *pcch: receives the size of the output string, excluding the terminator.
|
|---|
| 238 | if (RT_SUCCESS(vrc))
|
|---|
| 239 | m_cbAllocated = m_cch + 1;
|
|---|
| 240 | else
|
|---|
| 241 | {
|
|---|
| 242 | if ( vrc != VERR_NO_STR_MEMORY
|
|---|
| 243 | && vrc != VERR_NO_MEMORY)
|
|---|
| 244 | {
|
|---|
| 245 | /* ASSUME: input is valid Utf-16. Fake out of memory error. */
|
|---|
| 246 | AssertLogRelMsgFailed(("%Rrc %.*Rhxs\n", vrc, RTUtf16Len(a_pbstr) * sizeof(RTUTF16), a_pbstr));
|
|---|
| 247 | }
|
|---|
| 248 |
|
|---|
| 249 | m_cch = 0;
|
|---|
| 250 | m_cbAllocated = 0;
|
|---|
| 251 | m_psz = NULL;
|
|---|
| 252 |
|
|---|
| 253 | return E_OUTOFMEMORY;
|
|---|
| 254 | }
|
|---|
| 255 | }
|
|---|
| 256 | else
|
|---|
| 257 | {
|
|---|
| 258 | m_cch = 0;
|
|---|
| 259 | m_cbAllocated = 0;
|
|---|
| 260 | m_psz = NULL;
|
|---|
| 261 | }
|
|---|
| 262 | return S_OK;
|
|---|
| 263 | }
|
|---|
| 264 |
|
|---|
| 265 |
|
|---|
| 266 | /**
|
|---|
| 267 | * A variant of Utf8Str::copyFromN that does not throw any exceptions but
|
|---|
| 268 | * returns E_OUTOFMEMORY instead.
|
|---|
| 269 | *
|
|---|
| 270 | * @param a_pcszSrc The source string.
|
|---|
| 271 | * @param a_offSrc Start offset to copy from.
|
|---|
| 272 | * @param a_cchSrc The source string.
|
|---|
| 273 | * @returns S_OK or E_OUTOFMEMORY.
|
|---|
| 274 | *
|
|---|
| 275 | * @remarks This calls cleanup() first, so the caller doesn't have to. (Saves
|
|---|
| 276 | * code space.)
|
|---|
| 277 | */
|
|---|
| 278 | HRESULT Utf8Str::copyFromExNComRC(const char *a_pcszSrc, size_t a_offSrc, size_t a_cchSrc)
|
|---|
| 279 | {
|
|---|
| 280 | cleanup();
|
|---|
| 281 | if (a_cchSrc)
|
|---|
| 282 | {
|
|---|
| 283 | m_psz = RTStrAlloc(a_cchSrc + 1);
|
|---|
| 284 | if (RT_LIKELY(m_psz))
|
|---|
| 285 | {
|
|---|
| 286 | m_cch = a_cchSrc;
|
|---|
| 287 | m_cbAllocated = a_cchSrc + 1;
|
|---|
| 288 | memcpy(m_psz, a_pcszSrc + a_offSrc, a_cchSrc);
|
|---|
| 289 | m_psz[a_cchSrc] = '\0';
|
|---|
| 290 | }
|
|---|
| 291 | else
|
|---|
| 292 | {
|
|---|
| 293 | m_cch = 0;
|
|---|
| 294 | m_cbAllocated = 0;
|
|---|
| 295 | return E_OUTOFMEMORY;
|
|---|
| 296 | }
|
|---|
| 297 | }
|
|---|
| 298 | else
|
|---|
| 299 | {
|
|---|
| 300 | m_cch = 0;
|
|---|
| 301 | m_cbAllocated = 0;
|
|---|
| 302 | m_psz = NULL;
|
|---|
| 303 | }
|
|---|
| 304 | return S_OK;
|
|---|
| 305 | }
|
|---|
| 306 |
|
|---|
| 307 | } /* namespace com */
|
|---|
