2 * Copyright (C) 1999 Lars Knoll (knoll@kde.org)
3 * Copyright (C) 2005, 2006, 2007, 2008, 2009, 2010 Apple Inc. All rights reserved.
4 * Copyright (C) 2009 Google Inc. All rights reserved.
6 * This library is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Library General Public
8 * License as published by the Free Software Foundation; either
9 * version 2 of the License, or (at your option) any later version.
11 * This library is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Library General Public License for more details.
16 * You should have received a copy of the GNU Library General Public License
17 * along with this library; see the file COPYING.LIB. If not, write to
18 * the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
19 * Boston, MA 02110-1301, USA.
27 #include <wtf/ASCIICType.h>
28 #include <wtf/CrossThreadRefCounted.h>
29 #include <wtf/Forward.h>
30 #include <wtf/OwnFastMallocPtr.h>
31 #include <wtf/StdLibExtras.h>
32 #include <wtf/StringHasher.h>
33 #include <wtf/Vector.h>
34 #include <wtf/text/StringImplBase.h>
35 #include <wtf/unicode/Unicode.h>
38 typedef const struct __CFString * CFStringRef;
45 // FIXME: This is a temporary layering violation while we move string code to WTF.
46 // Landing the file moves in one patch, will follow on with patches to change the namespaces.
48 struct IdentifierCStringTranslator;
49 struct IdentifierUCharBufferTranslator;
54 struct CStringTranslator;
55 struct HashAndCharactersTranslator;
56 struct HashAndUTF8CharactersTranslator;
57 struct UCharBufferTranslator;
59 enum TextCaseSensitivity { TextCaseSensitive, TextCaseInsensitive };
61 typedef OwnFastMallocPtr<const UChar> SharableUChar;
62 typedef CrossThreadRefCounted<SharableUChar> SharedUChar;
63 typedef bool (*CharacterMatchFunctionPtr)(UChar);
64 typedef bool (*IsWhiteSpaceFunctionPtr)(UChar);
66 class StringImpl : public StringImplBase {
67 friend struct JSC::IdentifierCStringTranslator;
68 friend struct JSC::IdentifierUCharBufferTranslator;
69 friend struct WTF::CStringTranslator;
70 friend struct WTF::HashAndCharactersTranslator;
71 friend struct WTF::HashAndUTF8CharactersTranslator;
72 friend struct WTF::UCharBufferTranslator;
73 friend class AtomicStringImpl;
75 // Used to construct static strings, which have an special refCount that can never hit zero.
76 // This means that the static string will never be destroyed, which is important because
77 // static strings will be shared across threads & ref-counted in a non-threadsafe manner.
78 StringImpl(const UChar* characters, unsigned length, StaticStringConstructType)
79 : StringImplBase(length, ConstructStaticString)
84 // Ensure that the hash is computed so that AtomicStringHash can call existingHash()
85 // with impunity. The empty string is special because it is never entered into
86 // AtomicString's HashKey, but still needs to compare correctly.
90 // Create a normal string with internal storage (BufferInternal)
91 StringImpl(unsigned length)
92 : StringImplBase(length, BufferInternal)
93 , m_data(reinterpret_cast<const UChar*>(this + 1))
101 // Create a StringImpl adopting ownership of the provided buffer (BufferOwned)
102 StringImpl(const UChar* characters, unsigned length)
103 : StringImplBase(length, BufferOwned)
112 // Used to create new strings that are a substring of an existing StringImpl (BufferSubstring)
113 StringImpl(const UChar* characters, unsigned length, PassRefPtr<StringImpl> base)
114 : StringImplBase(length, BufferSubstring)
116 , m_substringBuffer(base.leakRef())
121 ASSERT(m_substringBuffer->bufferOwnership() != BufferSubstring);
124 // Used to construct new strings sharing an existing SharedUChar (BufferShared)
125 StringImpl(const UChar* characters, unsigned length, PassRefPtr<SharedUChar> sharedBuffer)
126 : StringImplBase(length, BufferShared)
128 , m_sharedBuffer(sharedBuffer.leakRef())
135 // For use only by AtomicString's XXXTranslator helpers.
136 void setHash(unsigned hash)
140 ASSERT(hash == StringHasher::computeHash(m_data, m_length));
147 static PassRefPtr<StringImpl> create(const UChar*, unsigned length);
148 static PassRefPtr<StringImpl> create(const char*, unsigned length);
149 static PassRefPtr<StringImpl> create(const char*);
150 static PassRefPtr<StringImpl> create(const UChar*, unsigned length, PassRefPtr<SharedUChar> sharedBuffer);
151 static ALWAYS_INLINE PassRefPtr<StringImpl> create(PassRefPtr<StringImpl> rep, unsigned offset, unsigned length)
154 ASSERT(length <= rep->length());
159 StringImpl* ownerRep = (rep->bufferOwnership() == BufferSubstring) ? rep->m_substringBuffer : rep.get();
160 return adoptRef(new StringImpl(rep->m_data + offset, length, ownerRep));
163 static PassRefPtr<StringImpl> createUninitialized(unsigned length, UChar*& data);
164 static ALWAYS_INLINE PassRefPtr<StringImpl> tryCreateUninitialized(unsigned length, UChar*& output)
171 if (length > ((std::numeric_limits<unsigned>::max() - sizeof(StringImpl)) / sizeof(UChar))) {
175 StringImpl* resultImpl;
176 if (!tryFastMalloc(sizeof(UChar) * length + sizeof(StringImpl)).getValue(resultImpl)) {
180 output = reinterpret_cast<UChar*>(resultImpl + 1);
181 return adoptRef(new(resultImpl) StringImpl(length));
184 static unsigned dataOffset() { return OBJECT_OFFSETOF(StringImpl, m_data); }
185 static PassRefPtr<StringImpl> createWithTerminatingNullCharacter(const StringImpl&);
186 static PassRefPtr<StringImpl> createStrippingNullCharacters(const UChar*, unsigned length);
188 template<size_t inlineCapacity>
189 static PassRefPtr<StringImpl> adopt(Vector<UChar, inlineCapacity>& vector)
191 if (size_t size = vector.size()) {
192 ASSERT(vector.data());
193 if (size > std::numeric_limits<unsigned>::max())
195 return adoptRef(new StringImpl(vector.releaseBuffer(), size));
199 static PassRefPtr<StringImpl> adopt(StringBuffer&);
201 SharedUChar* sharedBuffer();
202 const UChar* characters() const { return m_data; }
206 // For substrings, return the cost of the base string.
207 if (bufferOwnership() == BufferSubstring)
208 return m_substringBuffer->cost();
210 if (m_refCountAndFlags & s_refCountFlagShouldReportedCost) {
211 m_refCountAndFlags &= ~s_refCountFlagShouldReportedCost;
217 bool isIdentifier() const { return m_refCountAndFlags & s_refCountFlagIsIdentifier; }
218 void setIsIdentifier(bool isIdentifier)
222 m_refCountAndFlags |= s_refCountFlagIsIdentifier;
224 m_refCountAndFlags &= ~s_refCountFlagIsIdentifier;
227 bool hasTerminatingNullCharacter() const { return m_refCountAndFlags & s_refCountFlagHasTerminatingNullCharacter; }
229 bool isAtomic() const { return m_refCountAndFlags & s_refCountFlagIsAtomic; }
230 void setIsAtomic(bool isIdentifier)
234 m_refCountAndFlags |= s_refCountFlagIsAtomic;
236 m_refCountAndFlags &= ~s_refCountFlagIsAtomic;
239 unsigned hash() const { if (!m_hash) m_hash = StringHasher::computeHash(m_data, m_length); return m_hash; }
240 unsigned existingHash() const { ASSERT(m_hash); return m_hash; }
241 bool hasHash() const { return m_hash; }
243 ALWAYS_INLINE void deref() { m_refCountAndFlags -= s_refCountIncrement; if (!(m_refCountAndFlags & (s_refCountMask | s_refCountFlagStatic))) delete this; }
244 ALWAYS_INLINE bool hasOneRef() const { return (m_refCountAndFlags & (s_refCountMask | s_refCountFlagStatic)) == s_refCountIncrement; }
246 static StringImpl* empty();
248 static void copyChars(UChar* destination, const UChar* source, unsigned numCharacters)
250 if (numCharacters <= s_copyCharsInlineCutOff) {
251 for (unsigned i = 0; i < numCharacters; ++i)
252 destination[i] = source[i];
254 memcpy(destination, source, numCharacters * sizeof(UChar));
257 // Returns a StringImpl suitable for use on another thread.
258 PassRefPtr<StringImpl> crossThreadString();
259 // Makes a deep copy. Helpful only if you need to use a String on another thread
260 // (use crossThreadString if the method call doesn't need to be threadsafe).
261 // Since StringImpl objects are immutable, there's no other reason to make a copy.
262 PassRefPtr<StringImpl> threadsafeCopy() const;
264 PassRefPtr<StringImpl> substring(unsigned pos, unsigned len = UINT_MAX);
266 UChar operator[](unsigned i) { ASSERT(i < m_length); return m_data[i]; }
267 UChar32 characterStartingAt(unsigned);
269 bool containsOnlyWhitespace();
271 int toIntStrict(bool* ok = 0, int base = 10);
272 unsigned toUIntStrict(bool* ok = 0, int base = 10);
273 int64_t toInt64Strict(bool* ok = 0, int base = 10);
274 uint64_t toUInt64Strict(bool* ok = 0, int base = 10);
275 intptr_t toIntPtrStrict(bool* ok = 0, int base = 10);
277 int toInt(bool* ok = 0); // ignores trailing garbage
278 unsigned toUInt(bool* ok = 0); // ignores trailing garbage
279 int64_t toInt64(bool* ok = 0); // ignores trailing garbage
280 uint64_t toUInt64(bool* ok = 0); // ignores trailing garbage
281 intptr_t toIntPtr(bool* ok = 0); // ignores trailing garbage
283 double toDouble(bool* ok = 0, bool* didReadNumber = 0);
284 float toFloat(bool* ok = 0, bool* didReadNumber = 0);
286 PassRefPtr<StringImpl> lower();
287 PassRefPtr<StringImpl> upper();
289 PassRefPtr<StringImpl> fill(UChar);
290 PassRefPtr<StringImpl> foldCase();
292 PassRefPtr<StringImpl> stripWhiteSpace();
293 PassRefPtr<StringImpl> stripWhiteSpace(IsWhiteSpaceFunctionPtr);
294 PassRefPtr<StringImpl> simplifyWhiteSpace();
295 PassRefPtr<StringImpl> simplifyWhiteSpace(IsWhiteSpaceFunctionPtr);
297 PassRefPtr<StringImpl> removeCharacters(CharacterMatchFunctionPtr);
299 size_t find(UChar, unsigned index = 0);
300 size_t find(CharacterMatchFunctionPtr, unsigned index = 0);
301 size_t find(const char*, unsigned index = 0);
302 size_t find(StringImpl*, unsigned index = 0);
303 size_t findIgnoringCase(const char*, unsigned index = 0);
304 size_t findIgnoringCase(StringImpl*, unsigned index = 0);
306 size_t reverseFind(UChar, unsigned index = UINT_MAX);
307 size_t reverseFind(StringImpl*, unsigned index = UINT_MAX);
308 size_t reverseFindIgnoringCase(StringImpl*, unsigned index = UINT_MAX);
310 bool startsWith(StringImpl* str, bool caseSensitive = true) { return (caseSensitive ? reverseFind(str, 0) : reverseFindIgnoringCase(str, 0)) == 0; }
311 bool endsWith(StringImpl*, bool caseSensitive = true);
313 PassRefPtr<StringImpl> replace(UChar, UChar);
314 PassRefPtr<StringImpl> replace(UChar, StringImpl*);
315 PassRefPtr<StringImpl> replace(StringImpl*, StringImpl*);
316 PassRefPtr<StringImpl> replace(unsigned index, unsigned len, StringImpl*);
318 WTF::Unicode::Direction defaultWritingDirection(bool* hasStrongDirectionality = 0);
321 CFStringRef createCFString();
324 operator NSString*();
328 // This number must be at least 2 to avoid sharing empty, null as well as 1 character strings from SmallStrings.
329 static const unsigned s_copyCharsInlineCutOff = 20;
331 static PassRefPtr<StringImpl> createStrippingNullCharactersSlowCase(const UChar*, unsigned length);
333 BufferOwnership bufferOwnership() const { return static_cast<BufferOwnership>(m_refCountAndFlags & s_refCountMaskBufferOwnership); }
334 bool isStatic() const { return m_refCountAndFlags & s_refCountFlagStatic; }
335 template <class UCharPredicate> PassRefPtr<StringImpl> stripMatchedCharacters(UCharPredicate);
336 template <class UCharPredicate> PassRefPtr<StringImpl> simplifyMatchedCharactersToSpace(UCharPredicate);
341 StringImpl* m_substringBuffer;
342 SharedUChar* m_sharedBuffer;
344 mutable unsigned m_hash;
347 bool equal(const StringImpl*, const StringImpl*);
348 bool equal(const StringImpl*, const char*);
349 inline bool equal(const char* a, StringImpl* b) { return equal(b, a); }
350 bool equal(const StringImpl*, const UChar*, unsigned);
352 bool equalIgnoringCase(StringImpl*, StringImpl*);
353 bool equalIgnoringCase(StringImpl*, const char*);
354 inline bool equalIgnoringCase(const char* a, StringImpl* b) { return equalIgnoringCase(b, a); }
355 bool equalIgnoringCase(const UChar* a, const char* b, unsigned length);
356 inline bool equalIgnoringCase(const char* a, const UChar* b, unsigned length) { return equalIgnoringCase(b, a, length); }
358 bool equalIgnoringNullity(StringImpl*, StringImpl*);
360 template<size_t inlineCapacity>
361 bool equalIgnoringNullity(const Vector<UChar, inlineCapacity>& a, StringImpl* b)
365 if (a.size() != b->length())
367 return !memcmp(a.data(), b->characters(), b->length());
370 int codePointCompare(const StringImpl*, const StringImpl*);
372 static inline bool isSpaceOrNewline(UChar c)
374 // Use isASCIISpace() for basic Latin-1.
375 // This will include newlines, which aren't included in Unicode DirWS.
376 return c <= 0x7F ? WTF::isASCIISpace(c) : WTF::Unicode::direction(c) == WTF::Unicode::WhiteSpaceNeutral;
379 // This is a hot function because it's used when parsing HTML.
380 inline PassRefPtr<StringImpl> StringImpl::createStrippingNullCharacters(const UChar* characters, unsigned length)
385 // Optimize for the case where there are no Null characters by quickly
386 // searching for nulls, and then using StringImpl::create, which will
387 // memcpy the whole buffer. This is faster than assigning character by
388 // character during the loop.
392 for (unsigned i = 0; !foundNull && i < length; i++) {
393 int c = characters[i]; // more efficient than using UChar here (at least on Intel Mac OS)
397 return StringImpl::create(characters, length);
399 return StringImpl::createStrippingNullCharactersSlowCase(characters, length);
404 // StringHash is the default hash for StringImpl* and RefPtr<StringImpl>
405 template<typename T> struct DefaultHash;
406 template<> struct DefaultHash<StringImpl*> {
407 typedef StringHash Hash;
409 template<> struct DefaultHash<RefPtr<StringImpl> > {
410 typedef StringHash Hash;
415 using WTF::StringImpl;
417 using WTF::TextCaseSensitivity;
418 using WTF::TextCaseSensitive;
419 using WTF::TextCaseInsensitive;