dolphin/Source/Core/Common/StringUtil.cpp
Léo Lam 89b0ab2d22 StringUtil: Add IsPrintableCharacter and use it
Add a function that safely returns whether a character is printable
i.e. whether 0x20 <= c <= 0x7e is true.

This is done in several places in our codebase and it's easy to run
into undefined behaviour if the C version defined in <cctype>
is used instead of this one, since its behaviour is undefined
if the character is not representable as an unsigned char.

This fixes MemoryViewWidget.
2020-01-16 00:22:26 +01:00

604 lines
15 KiB
C++

// Copyright 2008 Dolphin Emulator Project
// Licensed under GPLv2+
// Refer to the license.txt file included.
#include "Common/StringUtil.h"
#include <algorithm>
#include <cstdarg>
#include <cstddef>
#include <cstdio>
#include <cstdlib>
#include <cstring>
#include <iomanip>
#include <istream>
#include <iterator>
#include <limits.h>
#include <locale>
#include <sstream>
#include <string>
#include <vector>
#include <fmt/format.h>
#include "Common/CommonFuncs.h"
#include "Common/CommonPaths.h"
#include "Common/CommonTypes.h"
#include "Common/Logging/Log.h"
#include "Common/Swap.h"
#ifdef _WIN32
#include <Windows.h>
constexpr u32 CODEPAGE_SHIFT_JIS = 932;
constexpr u32 CODEPAGE_WINDOWS_1252 = 1252;
#else
#include <codecvt>
#include <errno.h>
#include <iconv.h>
#include <locale.h>
#endif
#if !defined(_WIN32) && !defined(ANDROID) && !defined(__HAIKU__) && !defined(__OpenBSD__)
static locale_t GetCLocale()
{
static locale_t c_locale = newlocale(LC_ALL_MASK, "C", nullptr);
return c_locale;
}
#endif
std::string HexDump(const u8* data, size_t size)
{
constexpr size_t BYTES_PER_LINE = 16;
std::string out;
for (size_t row_start = 0; row_start < size; row_start += BYTES_PER_LINE)
{
out += fmt::format("{:06x}: ", row_start);
for (size_t i = 0; i < BYTES_PER_LINE; ++i)
{
if (row_start + i < size)
{
out += fmt::format("{:02x} ", data[row_start + i]);
}
else
{
out += " ";
}
}
out += " ";
for (size_t i = 0; i < BYTES_PER_LINE; ++i)
{
if (row_start + i < size)
{
char c = static_cast<char>(data[row_start + i]);
out += IsPrintableCharacter(c) ? c : '.';
}
}
out += "\n";
}
return out;
}
// faster than sscanf
bool AsciiToHex(const std::string& _szValue, u32& result)
{
// Set errno to a good state.
errno = 0;
char* endptr = nullptr;
const u32 value = strtoul(_szValue.c_str(), &endptr, 16);
if (!endptr || *endptr)
return false;
if (errno == ERANGE)
return false;
result = value;
return true;
}
bool CharArrayFromFormatV(char* out, int outsize, const char* format, va_list args)
{
int writtenCount;
#ifdef _WIN32
// You would think *printf are simple, right? Iterate on each character,
// if it's a format specifier handle it properly, etc.
//
// Nooooo. Not according to the C standard.
//
// According to the C99 standard (7.19.6.1 "The fprintf function")
// The format shall be a multibyte character sequence
//
// Because some character encodings might have '%' signs in the middle of
// a multibyte sequence (SJIS for example only specifies that the first
// byte of a 2 byte sequence is "high", the second byte can be anything),
// printf functions have to decode the multibyte sequences and try their
// best to not screw up.
//
// Unfortunately, on Windows, the locale for most languages is not UTF-8
// as we would need. Notably, for zh_TW, Windows chooses EUC-CN as the
// locale, and completely fails when trying to decode UTF-8 as EUC-CN.
//
// On the other hand, the fix is simple: because we use UTF-8, no such
// multibyte handling is required as we can simply assume that no '%' char
// will be present in the middle of a multibyte sequence.
//
// This is why we look up the default C locale here and use _vsnprintf_l.
static _locale_t c_locale = nullptr;
if (!c_locale)
c_locale = _create_locale(LC_ALL, "C");
writtenCount = _vsnprintf_l(out, outsize, format, c_locale, args);
#else
#if !defined(ANDROID) && !defined(__HAIKU__) && !defined(__OpenBSD__)
locale_t previousLocale = uselocale(GetCLocale());
#endif
writtenCount = vsnprintf(out, outsize, format, args);
#if !defined(ANDROID) && !defined(__HAIKU__) && !defined(__OpenBSD__)
uselocale(previousLocale);
#endif
#endif
if (writtenCount > 0 && writtenCount < outsize)
{
out[writtenCount] = '\0';
return true;
}
else
{
out[outsize - 1] = '\0';
return false;
}
}
std::string StringFromFormat(const char* format, ...)
{
va_list args;
va_start(args, format);
std::string res = StringFromFormatV(format, args);
va_end(args);
return res;
}
std::string StringFromFormatV(const char* format, va_list args)
{
char* buf = nullptr;
#ifdef _WIN32
int required = _vscprintf(format, args);
buf = new char[required + 1];
CharArrayFromFormatV(buf, required + 1, format, args);
std::string temp = buf;
delete[] buf;
#else
#if !defined(ANDROID) && !defined(__HAIKU__) && !defined(__OpenBSD__)
locale_t previousLocale = uselocale(GetCLocale());
#endif
if (vasprintf(&buf, format, args) < 0)
{
ERROR_LOG(COMMON, "Unable to allocate memory for string");
buf = nullptr;
}
#if !defined(ANDROID) && !defined(__HAIKU__) && !defined(__OpenBSD__)
uselocale(previousLocale);
#endif
std::string temp = buf;
free(buf);
#endif
return temp;
}
// For Debugging. Read out an u8 array.
std::string ArrayToString(const u8* data, u32 size, int line_len, bool spaces)
{
std::ostringstream oss;
oss << std::setfill('0') << std::hex;
for (int line = 0; size; ++data, --size)
{
oss << std::setw(2) << static_cast<int>(*data);
if (line_len == ++line)
{
oss << '\n';
line = 0;
}
else if (spaces)
oss << ' ';
}
return oss.str();
}
// Turns " hello " into "hello". Also handles tabs.
std::string_view StripSpaces(std::string_view str)
{
const size_t s = str.find_first_not_of(" \t\r\n");
if (str.npos != s)
return str.substr(s, str.find_last_not_of(" \t\r\n") - s + 1);
else
return "";
}
// "\"hello\"" is turned to "hello"
// This one assumes that the string has already been space stripped in both
// ends, as done by StripSpaces above, for example.
std::string_view StripQuotes(std::string_view s)
{
if (!s.empty() && '\"' == s[0] && '\"' == *s.rbegin())
return s.substr(1, s.size() - 2);
else
return s;
}
bool TryParse(const std::string& str, bool* const output)
{
float value;
const bool is_valid_float = TryParse(str, &value);
if ((is_valid_float && value == 1) || !strcasecmp("true", str.c_str()))
*output = true;
else if ((is_valid_float && value == 0) || !strcasecmp("false", str.c_str()))
*output = false;
else
return false;
return true;
}
std::string ValueToString(u16 value)
{
return fmt::format("0x{:04x}", value);
}
std::string ValueToString(u32 value)
{
return fmt::format("0x{:08x}", value);
}
std::string ValueToString(u64 value)
{
return fmt::format("0x{:016x}", value);
}
std::string ValueToString(float value)
{
return fmt::format("{:#.9g}", value);
}
std::string ValueToString(double value)
{
return fmt::format("{:#.17g}", value);
}
std::string ValueToString(int value)
{
return std::to_string(value);
}
std::string ValueToString(s64 value)
{
return std::to_string(value);
}
std::string ValueToString(bool value)
{
return value ? "True" : "False";
}
bool SplitPath(std::string_view full_path, std::string* path, std::string* filename,
std::string* extension)
{
if (full_path.empty())
return false;
size_t dir_end = full_path.find_last_of("/"
// Windows needs the : included for something like just "C:" to be considered a directory
#ifdef _WIN32
":"
#endif
);
if (std::string::npos == dir_end)
dir_end = 0;
else
dir_end += 1;
size_t fname_end = full_path.rfind('.');
if (fname_end < dir_end || std::string::npos == fname_end)
fname_end = full_path.size();
if (path)
*path = full_path.substr(0, dir_end);
if (filename)
*filename = full_path.substr(dir_end, fname_end - dir_end);
if (extension)
*extension = full_path.substr(fname_end);
return true;
}
void BuildCompleteFilename(std::string& complete_filename, std::string_view path,
std::string_view filename)
{
complete_filename = path;
// check for seperator
if (DIR_SEP_CHR != *complete_filename.rbegin())
complete_filename += DIR_SEP_CHR;
// add the filename
complete_filename += filename;
}
std::vector<std::string> SplitString(const std::string& str, const char delim)
{
std::istringstream iss(str);
std::vector<std::string> output(1);
while (std::getline(iss, *output.rbegin(), delim))
output.push_back("");
output.pop_back();
return output;
}
std::string JoinStrings(const std::vector<std::string>& strings, const std::string& delimiter)
{
// Check if we can return early, just for speed
if (strings.empty())
return "";
std::ostringstream res;
std::copy(strings.begin(), strings.end(),
std::ostream_iterator<std::string>(res, delimiter.c_str()));
// Drop the trailing delimiter.
std::string joined = res.str();
return joined.substr(0, joined.length() - delimiter.length());
}
std::string TabsToSpaces(int tab_size, std::string str)
{
const std::string spaces(tab_size, ' ');
size_t i = 0;
while (str.npos != (i = str.find('\t')))
str.replace(i, 1, spaces);
return str;
}
std::string ReplaceAll(std::string result, std::string_view src, std::string_view dest)
{
size_t pos = 0;
if (src == dest)
return result;
while ((pos = result.find(src, pos)) != std::string::npos)
{
result.replace(pos, src.size(), dest);
pos += dest.length();
}
return result;
}
bool StringBeginsWith(std::string_view str, std::string_view begin)
{
return str.size() >= begin.size() && std::equal(begin.begin(), begin.end(), str.begin());
}
bool StringEndsWith(std::string_view str, std::string_view end)
{
return str.size() >= end.size() && std::equal(end.rbegin(), end.rend(), str.rbegin());
}
void StringPopBackIf(std::string* s, char c)
{
if (!s->empty() && s->back() == c)
s->pop_back();
}
#ifdef _WIN32
std::wstring CPToUTF16(u32 code_page, std::string_view input)
{
auto const size =
MultiByteToWideChar(code_page, 0, input.data(), static_cast<int>(input.size()), nullptr, 0);
std::wstring output;
output.resize(size);
if (size == 0 ||
size != MultiByteToWideChar(code_page, 0, input.data(), static_cast<int>(input.size()),
&output[0], static_cast<int>(output.size())))
{
output.clear();
}
return output;
}
std::string UTF16ToCP(u32 code_page, std::wstring_view input)
{
std::string output;
if (0 != input.size())
{
// "If cchWideChar [input buffer size] is set to 0, the function fails." -MSDN
auto const size = WideCharToMultiByte(
code_page, 0, input.data(), static_cast<int>(input.size()), nullptr, 0, nullptr, false);
output.resize(size);
if (size != WideCharToMultiByte(code_page, 0, input.data(), static_cast<int>(input.size()),
&output[0], static_cast<int>(output.size()), nullptr, nullptr))
{
const DWORD error_code = GetLastError();
ERROR_LOG(COMMON, "WideCharToMultiByte Error in String '%s': %lu",
std::wstring(input).c_str(), error_code);
output.clear();
}
}
return output;
}
std::wstring UTF8ToUTF16(std::string_view input)
{
return CPToUTF16(CP_UTF8, input);
}
std::string UTF16ToUTF8(std::wstring_view input)
{
return UTF16ToCP(CP_UTF8, input);
}
std::string SHIFTJISToUTF8(std::string_view input)
{
return UTF16ToUTF8(CPToUTF16(CODEPAGE_SHIFT_JIS, input));
}
std::string UTF8ToSHIFTJIS(std::string_view input)
{
return UTF16ToCP(CODEPAGE_SHIFT_JIS, UTF8ToUTF16(input));
}
std::string CP1252ToUTF8(std::string_view input)
{
return UTF16ToUTF8(CPToUTF16(CODEPAGE_WINDOWS_1252, input));
}
std::string UTF16BEToUTF8(const char16_t* str, size_t max_size)
{
const char16_t* str_end = std::find(str, str + max_size, '\0');
std::wstring result(static_cast<size_t>(str_end - str), '\0');
std::transform(str, str_end, result.begin(), static_cast<u16 (&)(u16)>(Common::swap16));
return UTF16ToUTF8(result);
}
#else
template <typename T>
std::string CodeTo(const char* tocode, const char* fromcode, std::basic_string_view<T> input)
{
std::string result;
iconv_t const conv_desc = iconv_open(tocode, fromcode);
if ((iconv_t)-1 == conv_desc)
{
ERROR_LOG(COMMON, "Iconv initialization failure [%s]: %s", fromcode, strerror(errno));
}
else
{
size_t const in_bytes = sizeof(T) * input.size();
size_t const out_buffer_size = 4 * in_bytes;
std::string out_buffer;
out_buffer.resize(out_buffer_size);
auto src_buffer = input.data();
size_t src_bytes = in_bytes;
auto dst_buffer = out_buffer.data();
size_t dst_bytes = out_buffer.size();
while (src_bytes != 0)
{
size_t const iconv_result =
iconv(conv_desc, (char**)(&src_buffer), &src_bytes, &dst_buffer, &dst_bytes);
if ((size_t)-1 == iconv_result)
{
if (EILSEQ == errno || EINVAL == errno)
{
// Try to skip the bad character
if (src_bytes != 0)
{
--src_bytes;
++src_buffer;
}
}
else
{
ERROR_LOG(COMMON, "iconv failure [%s]: %s", fromcode, strerror(errno));
break;
}
}
}
out_buffer.resize(out_buffer_size - dst_bytes);
out_buffer.swap(result);
iconv_close(conv_desc);
}
return result;
}
template <typename T>
std::string CodeToUTF8(const char* fromcode, std::basic_string_view<T> input)
{
return CodeTo("UTF-8", fromcode, input);
}
std::string CP1252ToUTF8(std::string_view input)
{
// return CodeToUTF8("CP1252//TRANSLIT", input);
// return CodeToUTF8("CP1252//IGNORE", input);
return CodeToUTF8("CP1252", input);
}
std::string SHIFTJISToUTF8(std::string_view input)
{
// return CodeToUTF8("CP932", input);
return CodeToUTF8("SJIS", input);
}
std::string UTF8ToSHIFTJIS(std::string_view input)
{
return CodeTo("SJIS", "UTF-8", input);
}
std::string UTF16ToUTF8(std::wstring_view input)
{
std::wstring_convert<std::codecvt_utf8<wchar_t>, wchar_t> converter;
return converter.to_bytes(input.data(), input.data() + input.size());
}
std::string UTF16BEToUTF8(const char16_t* str, size_t max_size)
{
const char16_t* str_end = std::find(str, str + max_size, '\0');
return CodeToUTF8("UTF-16BE", std::u16string_view(str, static_cast<size_t>(str_end - str)));
}
#endif
#ifdef HAS_STD_FILESYSTEM
// This is a replacement for path::u8path, which is deprecated starting with C++20.
std::filesystem::path StringToPath(std::string_view path)
{
#ifdef _MSC_VER
return std::filesystem::path(UTF8ToUTF16(path));
#else
return std::filesystem::path(path);
#endif
}
// This is a replacement for path::u8string that always has the return type std::string.
// path::u8string returns std::u8string starting with C++20, which is annoying to convert.
std::string PathToString(const std::filesystem::path& path)
{
#ifdef _MSC_VER
return UTF16ToUTF8(path.native());
#else
return path.native();
#endif
}
#endif