blob: 7198dec6185837f335e978a8b2239cc6bc4dd457 [file] [log] [blame]
// Copyright (c) 2011 The Chromium Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#include "base/string_util.h"
#include "build/build_config.h"
#include <ctype.h>
#include <errno.h>
#include <math.h>
#include <stdarg.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <time.h>
#include <wchar.h>
#include <wctype.h>
#include <algorithm>
#include <vector>
#include "base/basictypes.h"
#include "base/logging.h"
#include "base/memory/singleton.h"
#include "base/third_party/dmg_fp/dmg_fp.h"
#include "base/utf_string_conversion_utils.h"
#include "base/utf_string_conversions.h"
#include "base/third_party/icu/icu_utf.h"
namespace {
// Force the singleton used by Empty[W]String[16] to be a unique type. This
// prevents other code that might accidentally use Singleton<string> from
// getting our internal one.
struct EmptyStrings {
EmptyStrings() {}
const std::string s;
const std::wstring ws;
const string16 s16;
static EmptyStrings* GetInstance() {
return Singleton<EmptyStrings>::get();
}
};
// Used by ReplaceStringPlaceholders to track the position in the string of
// replaced parameters.
struct ReplacementOffset {
ReplacementOffset(uintptr_t parameter, size_t offset)
: parameter(parameter),
offset(offset) {}
// Index of the parameter.
uintptr_t parameter;
// Starting position in the string.
size_t offset;
};
static bool CompareParameter(const ReplacementOffset& elem1,
const ReplacementOffset& elem2) {
return elem1.parameter < elem2.parameter;
}
} // namespace
namespace base {
bool IsWprintfFormatPortable(const wchar_t* format) {
for (const wchar_t* position = format; *position != '\0'; ++position) {
if (*position == '%') {
bool in_specification = true;
bool modifier_l = false;
while (in_specification) {
// Eat up characters until reaching a known specifier.
if (*++position == '\0') {
// The format string ended in the middle of a specification. Call
// it portable because no unportable specifications were found. The
// string is equally broken on all platforms.
return true;
}
if (*position == 'l') {
// 'l' is the only thing that can save the 's' and 'c' specifiers.
modifier_l = true;
} else if (((*position == 's' || *position == 'c') && !modifier_l) ||
*position == 'S' || *position == 'C' || *position == 'F' ||
*position == 'D' || *position == 'O' || *position == 'U') {
// Not portable.
return false;
}
if (wcschr(L"diouxXeEfgGaAcspn%", *position)) {
// Portable, keep scanning the rest of the format string.
in_specification = false;
}
}
}
}
return true;
}
} // namespace base
const std::string& EmptyString() {
return EmptyStrings::GetInstance()->s;
}
const std::wstring& EmptyWString() {
return EmptyStrings::GetInstance()->ws;
}
const string16& EmptyString16() {
return EmptyStrings::GetInstance()->s16;
}
#define WHITESPACE_UNICODE \
0x0009, /* <control-0009> to <control-000D> */ \
0x000A, \
0x000B, \
0x000C, \
0x000D, \
0x0020, /* Space */ \
0x0085, /* <control-0085> */ \
0x00A0, /* No-Break Space */ \
0x1680, /* Ogham Space Mark */ \
0x180E, /* Mongolian Vowel Separator */ \
0x2000, /* En Quad to Hair Space */ \
0x2001, \
0x2002, \
0x2003, \
0x2004, \
0x2005, \
0x2006, \
0x2007, \
0x2008, \
0x2009, \
0x200A, \
0x200C, /* Zero Width Non-Joiner */ \
0x2028, /* Line Separator */ \
0x2029, /* Paragraph Separator */ \
0x202F, /* Narrow No-Break Space */ \
0x205F, /* Medium Mathematical Space */ \
0x3000, /* Ideographic Space */ \
0
const wchar_t kWhitespaceWide[] = {
WHITESPACE_UNICODE
};
const char16 kWhitespaceUTF16[] = {
WHITESPACE_UNICODE
};
const char kWhitespaceASCII[] = {
0x09, // <control-0009> to <control-000D>
0x0A,
0x0B,
0x0C,
0x0D,
0x20, // Space
0
};
const char kUtf8ByteOrderMark[] = "\xEF\xBB\xBF";
template<typename STR>
bool RemoveCharsT(const STR& input,
const typename STR::value_type remove_chars[],
STR* output) {
bool removed = false;
size_t found;
*output = input;
found = output->find_first_of(remove_chars);
while (found != STR::npos) {
removed = true;
output->replace(found, 1, STR());
found = output->find_first_of(remove_chars, found);
}
return removed;
}
bool RemoveChars(const std::wstring& input,
const wchar_t remove_chars[],
std::wstring* output) {
return RemoveCharsT(input, remove_chars, output);
}
#if !defined(WCHAR_T_IS_UTF16)
bool RemoveChars(const string16& input,
const char16 remove_chars[],
string16* output) {
return RemoveCharsT(input, remove_chars, output);
}
#endif
bool RemoveChars(const std::string& input,
const char remove_chars[],
std::string* output) {
return RemoveCharsT(input, remove_chars, output);
}
template<typename STR>
TrimPositions TrimStringT(const STR& input,
const typename STR::value_type trim_chars[],
TrimPositions positions,
STR* output) {
// Find the edges of leading/trailing whitespace as desired.
const typename STR::size_type last_char = input.length() - 1;
const typename STR::size_type first_good_char = (positions & TRIM_LEADING) ?
input.find_first_not_of(trim_chars) : 0;
const typename STR::size_type last_good_char = (positions & TRIM_TRAILING) ?
input.find_last_not_of(trim_chars) : last_char;
// When the string was all whitespace, report that we stripped off whitespace
// from whichever position the caller was interested in. For empty input, we
// stripped no whitespace, but we still need to clear |output|.
if (input.empty() ||
(first_good_char == STR::npos) || (last_good_char == STR::npos)) {
bool input_was_empty = input.empty(); // in case output == &input
output->clear();
return input_was_empty ? TRIM_NONE : positions;
}
// Trim the whitespace.
*output =
input.substr(first_good_char, last_good_char - first_good_char + 1);
// Return where we trimmed from.
return static_cast<TrimPositions>(
((first_good_char == 0) ? TRIM_NONE : TRIM_LEADING) |
((last_good_char == last_char) ? TRIM_NONE : TRIM_TRAILING));
}
bool TrimString(const std::wstring& input,
const wchar_t trim_chars[],
std::wstring* output) {
return TrimStringT(input, trim_chars, TRIM_ALL, output) != TRIM_NONE;
}
#if !defined(WCHAR_T_IS_UTF16)
bool TrimString(const string16& input,
const char16 trim_chars[],
string16* output) {
return TrimStringT(input, trim_chars, TRIM_ALL, output) != TRIM_NONE;
}
#endif
bool TrimString(const std::string& input,
const char trim_chars[],
std::string* output) {
return TrimStringT(input, trim_chars, TRIM_ALL, output) != TRIM_NONE;
}
void TruncateUTF8ToByteSize(const std::string& input,
const size_t byte_size,
std::string* output) {
DCHECK(output);
if (byte_size > input.length()) {
*output = input;
return;
}
DCHECK_LE(byte_size, static_cast<uint32>(kint32max));
// Note: This cast is necessary because CBU8_NEXT uses int32s.
int32 truncation_length = static_cast<int32>(byte_size);
int32 char_index = truncation_length - 1;
const char* data = input.data();
// Using CBU8, we will move backwards from the truncation point
// to the beginning of the string looking for a valid UTF8
// character. Once a full UTF8 character is found, we will
// truncate the string to the end of that character.
while (char_index >= 0) {
int32 prev = char_index;
uint32 code_point = 0;
CBU8_NEXT(data, char_index, truncation_length, code_point);
if (!base::IsValidCharacter(code_point) ||
!base::IsValidCodepoint(code_point)) {
char_index = prev - 1;
} else {
break;
}
}
if (char_index >= 0 )
*output = input.substr(0, char_index);
else
output->clear();
}
TrimPositions TrimWhitespace(const std::wstring& input,
TrimPositions positions,
std::wstring* output) {
return TrimStringT(input, kWhitespaceWide, positions, output);
}
#if !defined(WCHAR_T_IS_UTF16)
TrimPositions TrimWhitespace(const string16& input,
TrimPositions positions,
string16* output) {
return TrimStringT(input, kWhitespaceUTF16, positions, output);
}
#endif
TrimPositions TrimWhitespaceASCII(const std::string& input,
TrimPositions positions,
std::string* output) {
return TrimStringT(input, kWhitespaceASCII, positions, output);
}
// This function is only for backward-compatibility.
// To be removed when all callers are updated.
TrimPositions TrimWhitespace(const std::string& input,
TrimPositions positions,
std::string* output) {
return TrimWhitespaceASCII(input, positions, output);
}
template<typename STR>
STR CollapseWhitespaceT(const STR& text,
bool trim_sequences_with_line_breaks) {
STR result;
result.resize(text.size());
// Set flags to pretend we're already in a trimmed whitespace sequence, so we
// will trim any leading whitespace.
bool in_whitespace = true;
bool already_trimmed = true;
int chars_written = 0;
for (typename STR::const_iterator i(text.begin()); i != text.end(); ++i) {
if (IsWhitespace(*i)) {
if (!in_whitespace) {
// Reduce all whitespace sequences to a single space.
in_whitespace = true;
result[chars_written++] = L' ';
}
if (trim_sequences_with_line_breaks && !already_trimmed &&
((*i == '\n') || (*i == '\r'))) {
// Whitespace sequences containing CR or LF are eliminated entirely.
already_trimmed = true;
--chars_written;
}
} else {
// Non-whitespace chracters are copied straight across.
in_whitespace = false;
already_trimmed = false;
result[chars_written++] = *i;
}
}
if (in_whitespace && !already_trimmed) {
// Any trailing whitespace is eliminated.
--chars_written;
}
result.resize(chars_written);
return result;
}
std::wstring CollapseWhitespace(const std::wstring& text,
bool trim_sequences_with_line_breaks) {
return CollapseWhitespaceT(text, trim_sequences_with_line_breaks);
}
#if !defined(WCHAR_T_IS_UTF16)
string16 CollapseWhitespace(const string16& text,
bool trim_sequences_with_line_breaks) {
return CollapseWhitespaceT(text, trim_sequences_with_line_breaks);
}
#endif
std::string CollapseWhitespaceASCII(const std::string& text,
bool trim_sequences_with_line_breaks) {
return CollapseWhitespaceT(text, trim_sequences_with_line_breaks);
}
bool ContainsOnlyWhitespaceASCII(const std::string& str) {
for (std::string::const_iterator i(str.begin()); i != str.end(); ++i) {
if (!IsAsciiWhitespace(*i))
return false;
}
return true;
}
bool ContainsOnlyWhitespace(const string16& str) {
for (string16::const_iterator i(str.begin()); i != str.end(); ++i) {
if (!IsWhitespace(*i))
return false;
}
return true;
}
template<typename STR>
static bool ContainsOnlyCharsT(const STR& input, const STR& characters) {
for (typename STR::const_iterator iter = input.begin();
iter != input.end(); ++iter) {
if (characters.find(*iter) == STR::npos)
return false;
}
return true;
}
bool ContainsOnlyChars(const std::wstring& input,
const std::wstring& characters) {
return ContainsOnlyCharsT(input, characters);
}
#if !defined(WCHAR_T_IS_UTF16)
bool ContainsOnlyChars(const string16& input, const string16& characters) {
return ContainsOnlyCharsT(input, characters);
}
#endif
bool ContainsOnlyChars(const std::string& input,
const std::string& characters) {
return ContainsOnlyCharsT(input, characters);
}
std::string WideToASCII(const std::wstring& wide) {
DCHECK(IsStringASCII(wide)) << wide;
return std::string(wide.begin(), wide.end());
}
std::string UTF16ToASCII(const string16& utf16) {
DCHECK(IsStringASCII(utf16)) << utf16;
return std::string(utf16.begin(), utf16.end());
}
// Latin1 is just the low range of Unicode, so we can copy directly to convert.
bool WideToLatin1(const std::wstring& wide, std::string* latin1) {
std::string output;
output.resize(wide.size());
latin1->clear();
for (size_t i = 0; i < wide.size(); i++) {
if (wide[i] > 255)
return false;
output[i] = static_cast<char>(wide[i]);
}
latin1->swap(output);
return true;
}
template<class STR>
static bool DoIsStringASCII(const STR& str) {
for (size_t i = 0; i < str.length(); i++) {
typename ToUnsigned<typename STR::value_type>::Unsigned c = str[i];
if (c > 0x7F)
return false;
}
return true;
}
bool IsStringASCII(const std::wstring& str) {
return DoIsStringASCII(str);
}
#if !defined(WCHAR_T_IS_UTF16)
bool IsStringASCII(const string16& str) {
return DoIsStringASCII(str);
}
#endif
bool IsStringASCII(const base::StringPiece& str) {
return DoIsStringASCII(str);
}
bool IsStringUTF8(const std::string& str) {
const char *src = str.data();
int32 src_len = static_cast<int32>(str.length());
int32 char_index = 0;
while (char_index < src_len) {
int32 code_point;
CBU8_NEXT(src, char_index, src_len, code_point);
if (!base::IsValidCharacter(code_point))
return false;
}
return true;
}
template<typename Iter>
static inline bool DoLowerCaseEqualsASCII(Iter a_begin,
Iter a_end,
const char* b) {
for (Iter it = a_begin; it != a_end; ++it, ++b) {
if (!*b || base::ToLowerASCII(*it) != *b)
return false;
}
return *b == 0;
}
// Front-ends for LowerCaseEqualsASCII.
bool LowerCaseEqualsASCII(const std::string& a, const char* b) {
return DoLowerCaseEqualsASCII(a.begin(), a.end(), b);
}
bool LowerCaseEqualsASCII(const std::wstring& a, const char* b) {
return DoLowerCaseEqualsASCII(a.begin(), a.end(), b);
}
#if !defined(WCHAR_T_IS_UTF16)
bool LowerCaseEqualsASCII(const string16& a, const char* b) {
return DoLowerCaseEqualsASCII(a.begin(), a.end(), b);
}
#endif
bool LowerCaseEqualsASCII(std::string::const_iterator a_begin,
std::string::const_iterator a_end,
const char* b) {
return DoLowerCaseEqualsASCII(a_begin, a_end, b);
}
bool LowerCaseEqualsASCII(std::wstring::const_iterator a_begin,
std::wstring::const_iterator a_end,
const char* b) {
return DoLowerCaseEqualsASCII(a_begin, a_end, b);
}
#if !defined(WCHAR_T_IS_UTF16)
bool LowerCaseEqualsASCII(string16::const_iterator a_begin,
string16::const_iterator a_end,
const char* b) {
return DoLowerCaseEqualsASCII(a_begin, a_end, b);
}
#endif
#if !defined(ANDROID)
bool LowerCaseEqualsASCII(const char* a_begin,
const char* a_end,
const char* b) {
return DoLowerCaseEqualsASCII(a_begin, a_end, b);
}
#endif // !ANDROID
#if !defined(ANDROID)
bool LowerCaseEqualsASCII(const wchar_t* a_begin,
const wchar_t* a_end,
const char* b) {
return DoLowerCaseEqualsASCII(a_begin, a_end, b);
}
#endif // !ANDROID
#if !defined(WCHAR_T_IS_UTF16) && !defined(ANDROID)
bool LowerCaseEqualsASCII(const char16* a_begin,
const char16* a_end,
const char* b) {
return DoLowerCaseEqualsASCII(a_begin, a_end, b);
}
#endif
bool EqualsASCII(const string16& a, const base::StringPiece& b) {
if (a.length() != b.length())
return false;
return std::equal(b.begin(), b.end(), a.begin());
}
bool StartsWithASCII(const std::string& str,
const std::string& search,
bool case_sensitive) {
if (case_sensitive)
return str.compare(0, search.length(), search) == 0;
else
return base::strncasecmp(str.c_str(), search.c_str(), search.length()) == 0;
}
template <typename STR>
bool StartsWithT(const STR& str, const STR& search, bool case_sensitive) {
if (case_sensitive) {
return str.compare(0, search.length(), search) == 0;
} else {
if (search.size() > str.size())
return false;
return std::equal(search.begin(), search.end(), str.begin(),
base::CaseInsensitiveCompare<typename STR::value_type>());
}
}
bool StartsWith(const std::wstring& str, const std::wstring& search,
bool case_sensitive) {
return StartsWithT(str, search, case_sensitive);
}
#if !defined(WCHAR_T_IS_UTF16)
bool StartsWith(const string16& str, const string16& search,
bool case_sensitive) {
return StartsWithT(str, search, case_sensitive);
}
#endif
template <typename STR>
bool EndsWithT(const STR& str, const STR& search, bool case_sensitive) {
typename STR::size_type str_length = str.length();
typename STR::size_type search_length = search.length();
if (search_length > str_length)
return false;
if (case_sensitive) {
return str.compare(str_length - search_length, search_length, search) == 0;
} else {
return std::equal(search.begin(), search.end(),
str.begin() + (str_length - search_length),
base::CaseInsensitiveCompare<typename STR::value_type>());
}
}
bool EndsWith(const std::string& str, const std::string& search,
bool case_sensitive) {
return EndsWithT(str, search, case_sensitive);
}
bool EndsWith(const std::wstring& str, const std::wstring& search,
bool case_sensitive) {
return EndsWithT(str, search, case_sensitive);
}
#if !defined(WCHAR_T_IS_UTF16)
bool EndsWith(const string16& str, const string16& search,
bool case_sensitive) {
return EndsWithT(str, search, case_sensitive);
}
#endif
DataUnits GetByteDisplayUnits(int64 bytes) {
// The byte thresholds at which we display amounts. A byte count is displayed
// in unit U when kUnitThresholds[U] <= bytes < kUnitThresholds[U+1].
// This must match the DataUnits enum.
static const int64 kUnitThresholds[] = {
0, // DATA_UNITS_BYTE,
3*1024, // DATA_UNITS_KIBIBYTE,
2*1024*1024, // DATA_UNITS_MEBIBYTE,
1024*1024*1024 // DATA_UNITS_GIBIBYTE,
};
if (bytes < 0) {
NOTREACHED() << "Negative bytes value";
return DATA_UNITS_BYTE;
}
int unit_index = arraysize(kUnitThresholds);
while (--unit_index > 0) {
if (bytes >= kUnitThresholds[unit_index])
break;
}
DCHECK(unit_index >= DATA_UNITS_BYTE && unit_index <= DATA_UNITS_GIBIBYTE);
return DataUnits(unit_index);
}
// TODO(mpcomplete): deal with locale
// Byte suffixes. This must match the DataUnits enum.
static const char* const kByteStrings[] = {
"B",
"kB",
"MB",
"GB"
};
static const char* const kSpeedStrings[] = {
"B/s",
"kB/s",
"MB/s",
"GB/s"
};
string16 FormatBytesInternal(int64 bytes,
DataUnits units,
bool show_units,
const char* const* suffix) {
if (bytes < 0) {
NOTREACHED() << "Negative bytes value";
return string16();
}
DCHECK(units >= DATA_UNITS_BYTE && units <= DATA_UNITS_GIBIBYTE);
// Put the quantity in the right units.
double unit_amount = static_cast<double>(bytes);
for (int i = 0; i < units; ++i)
unit_amount /= 1024.0;
char buf[64];
if (bytes != 0 && units != DATA_UNITS_BYTE && unit_amount < 100)
base::snprintf(buf, arraysize(buf), "%.1lf", unit_amount);
else
base::snprintf(buf, arraysize(buf), "%.0lf", unit_amount);
std::string ret(buf);
if (show_units) {
ret += " ";
ret += suffix[units];
}
return ASCIIToUTF16(ret);
}
string16 FormatBytes(int64 bytes, DataUnits units, bool show_units) {
return FormatBytesInternal(bytes, units, show_units, kByteStrings);
}
string16 FormatSpeed(int64 bytes, DataUnits units, bool show_units) {
return FormatBytesInternal(bytes, units, show_units, kSpeedStrings);
}
template<class StringType>
void DoReplaceSubstringsAfterOffset(StringType* str,
typename StringType::size_type start_offset,
const StringType& find_this,
const StringType& replace_with,
bool replace_all) {
if ((start_offset == StringType::npos) || (start_offset >= str->length()))
return;
DCHECK(!find_this.empty());
for (typename StringType::size_type offs(str->find(find_this, start_offset));
offs != StringType::npos; offs = str->find(find_this, offs)) {
str->replace(offs, find_this.length(), replace_with);
offs += replace_with.length();
if (!replace_all)
break;
}
}
void ReplaceFirstSubstringAfterOffset(string16* str,
string16::size_type start_offset,
const string16& find_this,
const string16& replace_with) {
DoReplaceSubstringsAfterOffset(str, start_offset, find_this, replace_with,
false); // replace first instance
}
void ReplaceFirstSubstringAfterOffset(std::string* str,
std::string::size_type start_offset,
const std::string& find_this,
const std::string& replace_with) {
DoReplaceSubstringsAfterOffset(str, start_offset, find_this, replace_with,
false); // replace first instance
}
void ReplaceSubstringsAfterOffset(string16* str,
string16::size_type start_offset,
const string16& find_this,
const string16& replace_with) {
DoReplaceSubstringsAfterOffset(str, start_offset, find_this, replace_with,
true); // replace all instances
}
void ReplaceSubstringsAfterOffset(std::string* str,
std::string::size_type start_offset,
const std::string& find_this,
const std::string& replace_with) {
DoReplaceSubstringsAfterOffset(str, start_offset, find_this, replace_with,
true); // replace all instances
}
template<typename STR>
static size_t TokenizeT(const STR& str,
const STR& delimiters,
std::vector<STR>* tokens) {
tokens->clear();
typename STR::size_type start = str.find_first_not_of(delimiters);
while (start != STR::npos) {
typename STR::size_type end = str.find_first_of(delimiters, start + 1);
if (end == STR::npos) {
tokens->push_back(str.substr(start));
break;
} else {
tokens->push_back(str.substr(start, end - start));
start = str.find_first_not_of(delimiters, end + 1);
}
}
return tokens->size();
}
size_t Tokenize(const std::wstring& str,
const std::wstring& delimiters,
std::vector<std::wstring>* tokens) {
return TokenizeT(str, delimiters, tokens);
}
#if !defined(WCHAR_T_IS_UTF16)
size_t Tokenize(const string16& str,
const string16& delimiters,
std::vector<string16>* tokens) {
return TokenizeT(str, delimiters, tokens);
}
#endif
size_t Tokenize(const std::string& str,
const std::string& delimiters,
std::vector<std::string>* tokens) {
return TokenizeT(str, delimiters, tokens);
}
size_t Tokenize(const base::StringPiece& str,
const base::StringPiece& delimiters,
std::vector<base::StringPiece>* tokens) {
return TokenizeT(str, delimiters, tokens);
}
template<typename STR>
static STR JoinStringT(const std::vector<STR>& parts,
typename STR::value_type sep) {
if (parts.empty())
return STR();
STR result(parts[0]);
typename std::vector<STR>::const_iterator iter = parts.begin();
++iter;
for (; iter != parts.end(); ++iter) {
result += sep;
result += *iter;
}
return result;
}
std::string JoinString(const std::vector<std::string>& parts, char sep) {
return JoinStringT(parts, sep);
}
string16 JoinString(const std::vector<string16>& parts, char16 sep) {
return JoinStringT(parts, sep);
}
template<class FormatStringType, class OutStringType>
OutStringType DoReplaceStringPlaceholders(const FormatStringType& format_string,
const std::vector<OutStringType>& subst, std::vector<size_t>* offsets) {
size_t substitutions = subst.size();
DCHECK(substitutions < 10);
size_t sub_length = 0;
for (typename std::vector<OutStringType>::const_iterator iter = subst.begin();
iter != subst.end(); ++iter) {
sub_length += iter->length();
}
OutStringType formatted;
formatted.reserve(format_string.length() + sub_length);
std::vector<ReplacementOffset> r_offsets;
for (typename FormatStringType::const_iterator i = format_string.begin();
i != format_string.end(); ++i) {
if ('$' == *i) {
if (i + 1 != format_string.end()) {
++i;
DCHECK('$' == *i || '1' <= *i) << "Invalid placeholder: " << *i;
if ('$' == *i) {
while (i != format_string.end() && '$' == *i) {
formatted.push_back('$');
++i;
}
--i;
} else {
uintptr_t index = *i - '1';
if (offsets) {
ReplacementOffset r_offset(index,
static_cast<int>(formatted.size()));
r_offsets.insert(std::lower_bound(r_offsets.begin(),
r_offsets.end(),
r_offset,
&CompareParameter),
r_offset);
}
if (index < substitutions)
formatted.append(subst.at(index));
}
}
} else {
formatted.push_back(*i);
}
}
if (offsets) {
for (std::vector<ReplacementOffset>::const_iterator i = r_offsets.begin();
i != r_offsets.end(); ++i) {
offsets->push_back(i->offset);
}
}
return formatted;
}
string16 ReplaceStringPlaceholders(const string16& format_string,
const std::vector<string16>& subst,
std::vector<size_t>* offsets) {
return DoReplaceStringPlaceholders(format_string, subst, offsets);
}
std::string ReplaceStringPlaceholders(const base::StringPiece& format_string,
const std::vector<std::string>& subst,
std::vector<size_t>* offsets) {
return DoReplaceStringPlaceholders(format_string, subst, offsets);
}
string16 ReplaceStringPlaceholders(const string16& format_string,
const string16& a,
size_t* offset) {
std::vector<size_t> offsets;
std::vector<string16> subst;
subst.push_back(a);
string16 result = ReplaceStringPlaceholders(format_string, subst, &offsets);
DCHECK(offsets.size() == 1);
if (offset) {
*offset = offsets[0];
}
return result;
}
static bool IsWildcard(base_icu::UChar32 character) {
return character == '*' || character == '?';
}
// Move the strings pointers to the point where they start to differ.
template <typename CHAR, typename NEXT>
static void EatSameChars(const CHAR** pattern, const CHAR* pattern_end,
const CHAR** string, const CHAR* string_end,
NEXT next) {
const CHAR* escape = NULL;
while (*pattern != pattern_end && *string != string_end) {
if (!escape && IsWildcard(**pattern)) {
// We don't want to match wildcard here, except if it's escaped.
return;
}
// Check if the escapement char is found. If so, skip it and move to the
// next character.
if (!escape && **pattern == '\\') {
escape = *pattern;
next(pattern, pattern_end);
continue;
}
// Check if the chars match, if so, increment the ptrs.
const CHAR* pattern_next = *pattern;
const CHAR* string_next = *string;
base_icu::UChar32 pattern_char = next(&pattern_next, pattern_end);
if (pattern_char == next(&string_next, string_end) &&
pattern_char != (base_icu::UChar32) CBU_SENTINEL) {
*pattern = pattern_next;
*string = string_next;
} else {
// Uh ho, it did not match, we are done. If the last char was an
// escapement, that means that it was an error to advance the ptr here,
// let's put it back where it was. This also mean that the MatchPattern
// function will return false because if we can't match an escape char
// here, then no one will.
if (escape) {
*pattern = escape;
}
return;
}
escape = NULL;
}
}
template <typename CHAR, typename NEXT>
static void EatWildcard(const CHAR** pattern, const CHAR* end, NEXT next) {
while (*pattern != end) {
if (!IsWildcard(**pattern))
return;
next(pattern, end);
}
}
template <typename CHAR, typename NEXT>
static bool MatchPatternT(const CHAR* eval, const CHAR* eval_end,
const CHAR* pattern, const CHAR* pattern_end,
int depth,
NEXT next) {
const int kMaxDepth = 16;
if (depth > kMaxDepth)
return false;
// Eat all the matching chars.
EatSameChars(&pattern, pattern_end, &eval, eval_end, next);
// If the string is empty, then the pattern must be empty too, or contains
// only wildcards.
if (eval == eval_end) {
EatWildcard(&pattern, pattern_end, next);
return pattern == pattern_end;
}
// Pattern is empty but not string, this is not a match.
if (pattern == pattern_end)
return false;
// If this is a question mark, then we need to compare the rest with
// the current string or the string with one character eaten.
const CHAR* next_pattern = pattern;
next(&next_pattern, pattern_end);
if (pattern[0] == '?') {
if (MatchPatternT(eval, eval_end, next_pattern, pattern_end,
depth + 1, next))
return true;
const CHAR* next_eval = eval;
next(&next_eval, eval_end);
if (MatchPatternT(next_eval, eval_end, next_pattern, pattern_end,
depth + 1, next))
return true;
}
// This is a *, try to match all the possible substrings with the remainder
// of the pattern.
if (pattern[0] == '*') {
// Collapse duplicate wild cards (********** into *) so that the
// method does not recurse unnecessarily. http://crbug.com/52839
EatWildcard(&next_pattern, pattern_end, next);
while (eval != eval_end) {
if (MatchPatternT(eval, eval_end, next_pattern, pattern_end,
depth + 1, next))
return true;
eval++;
}
// We reached the end of the string, let see if the pattern contains only
// wildcards.
if (eval == eval_end) {
EatWildcard(&pattern, pattern_end, next);
if (pattern != pattern_end)
return false;
return true;
}
}
return false;
}
struct NextCharUTF8 {
base_icu::UChar32 operator()(const char** p, const char* end) {
base_icu::UChar32 c;
int offset = 0;
CBU8_NEXT(*p, offset, end - *p, c);
*p += offset;
return c;
}
};
struct NextCharUTF16 {
base_icu::UChar32 operator()(const char16** p, const char16* end) {
base_icu::UChar32 c;
int offset = 0;
CBU16_NEXT(*p, offset, end - *p, c);
*p += offset;
return c;
}
};
bool MatchPattern(const base::StringPiece& eval,
const base::StringPiece& pattern) {
return MatchPatternT(eval.data(), eval.data() + eval.size(),
pattern.data(), pattern.data() + pattern.size(),
0, NextCharUTF8());
}
bool MatchPattern(const string16& eval, const string16& pattern) {
return MatchPatternT(eval.c_str(), eval.c_str() + eval.size(),
pattern.c_str(), pattern.c_str() + pattern.size(),
0, NextCharUTF16());
}
// The following code is compatible with the OpenBSD lcpy interface. See:
// http://www.gratisoft.us/todd/papers/strlcpy.html
// ftp://ftp.openbsd.org/pub/OpenBSD/src/lib/libc/string/{wcs,str}lcpy.c
namespace {
template <typename CHAR>
size_t lcpyT(CHAR* dst, const CHAR* src, size_t dst_size) {
for (size_t i = 0; i < dst_size; ++i) {
if ((dst[i] = src[i]) == 0) // We hit and copied the terminating NULL.
return i;
}
// We were left off at dst_size. We over copied 1 byte. Null terminate.
if (dst_size != 0)
dst[dst_size - 1] = 0;
// Count the rest of the |src|, and return it's length in characters.
while (src[dst_size]) ++dst_size;
return dst_size;
}
} // namespace
size_t base::strlcpy(char* dst, const char* src, size_t dst_size) {
return lcpyT<char>(dst, src, dst_size);
}
size_t base::wcslcpy(wchar_t* dst, const wchar_t* src, size_t dst_size) {
return lcpyT<wchar_t>(dst, src, dst_size);
}