First version

This commit is contained in:
2016-11-09 03:32:47 +01:00
parent e11f753caf
commit 5baeb631d5
44 changed files with 8453 additions and 0 deletions
+167
View File
@@ -0,0 +1,167 @@
//
// Copyright (c) 2012 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_ARGS_HPP_INCLUDED
#define NOWIDE_ARGS_HPP_INCLUDED
#include <nowide/config.hpp>
#include <nowide/stackstring.hpp>
#include <vector>
#ifdef NOWIDE_WINDOWS
#include <nowide/windows.hpp>
#endif
namespace nowide {
#if !defined(NOWIDE_WINDOWS) && !defined(NOWIDE_DOXYGEN)
class args {
public:
args(int &,char **&) {}
args(int &,char **&,char **&){}
~args() {}
};
#else
///
/// \brief args is a class that fixes standard main() function arguments and changes them to UTF-8 under
/// Microsoft Windows.
///
/// The class uses \c GetCommandLineW(), \c CommandLineToArgvW() and \c GetEnvironmentStringsW()
/// in order to obtain the information. It does not relates to actual values of argc,argv and env
/// under Windows.
///
/// It restores the original values in its destructor
///
/// \note the class owns the memory of the newly allocated strings
///
class args {
public:
///
/// Fix command line agruments
///
args(int &argc,char **&argv) :
old_argc_(argc),
old_argv_(argv),
old_env_(0),
old_argc_ptr_(&argc),
old_argv_ptr_(&argv),
old_env_ptr_(0)
{
fix_args(argc,argv);
}
///
/// Fix command line agruments and environment
///
args(int &argc,char **&argv,char **&en) :
old_argc_(argc),
old_argv_(argv),
old_env_(en),
old_argc_ptr_(&argc),
old_argv_ptr_(&argv),
old_env_ptr_(&en)
{
fix_args(argc,argv);
fix_env(en);
}
///
/// Restore original argc,argv,env values, if changed
///
~args()
{
if(old_argc_ptr_)
*old_argc_ptr_ = old_argc_;
if(old_argv_ptr_)
*old_argv_ptr_ = old_argv_;
if(old_env_ptr_)
*old_env_ptr_ = old_env_;
}
private:
void fix_args(int &argc,char **&argv)
{
int wargc;
wchar_t **wargv = CommandLineToArgvW(GetCommandLineW(),&wargc);
if(!wargv) {
argc = 0;
static char *dummy = 0;
argv = &dummy;
return;
}
try{
args_.resize(wargc+1,0);
arg_values_.resize(wargc);
for(int i=0;i<wargc;i++) {
if(!arg_values_[i].convert(wargv[i])) {
wargc = i;
break;
}
args_[i] = arg_values_[i].c_str();
}
argc = wargc;
argv = &args_[0];
}
catch(...) {
LocalFree(wargv);
throw;
}
LocalFree(wargv);
}
void fix_env(char **&en)
{
static char *dummy = 0;
en = &dummy;
wchar_t *wstrings = GetEnvironmentStringsW();
if(!wstrings)
return;
try {
wchar_t *wstrings_end = 0;
int count = 0;
for(wstrings_end = wstrings;*wstrings_end;wstrings_end+=wcslen(wstrings_end)+1)
count++;
if(env_.convert(wstrings,wstrings_end)) {
envp_.resize(count+1,0);
char *p=env_.c_str();
int pos = 0;
for(int i=0;i<count;i++) {
if(*p!='=')
envp_[pos++] = p;
p+=strlen(p)+1;
}
en = &envp_[0];
}
}
catch(...) {
FreeEnvironmentStringsW(wstrings);
throw;
}
FreeEnvironmentStringsW(wstrings);
}
std::vector<char *> args_;
std::vector<short_stackstring> arg_values_;
stackstring env_;
std::vector<char *> envp_;
int old_argc_;
char **old_argv_;
char **old_env_;
int *old_argc_ptr_;
char ***old_argv_ptr_;
char ***old_env_ptr_;
};
#endif
} // nowide
#endif
///
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+126
View File
@@ -0,0 +1,126 @@
//
// Copyright (c) 2012 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_CENV_H_INCLUDED
#define NOWIDE_CENV_H_INCLUDED
#include <string>
#include <stdexcept>
#include <stdlib.h>
#include <nowide/config.hpp>
#include <nowide/stackstring.hpp>
#include <vector>
#ifdef NOWIDE_WINDOWS
#include <nowide/windows.hpp>
#endif
namespace nowide {
#if !defined(NOWIDE_WINDOWS) && !defined(NOWIDE_DOXYGEN)
using ::getenv;
using ::setenv;
using ::unsetenv;
using ::putenv;
#else
///
/// \brief UTF-8 aware getenv. Returns 0 if the variable is not set.
///
/// This function is not thread safe or reenterable as defined by the standard library
///
inline char *getenv(char const *key)
{
static stackstring value;
wshort_stackstring name;
if(!name.convert(key))
return 0;
static const size_t buf_size = 64;
wchar_t buf[buf_size];
std::vector<wchar_t> tmp;
wchar_t *ptr = buf;
size_t n = GetEnvironmentVariableW(name.c_str(),buf,buf_size);
if(n == 0 && GetLastError() == 203) // ERROR_ENVVAR_NOT_FOUND
return 0;
if(n >= buf_size) {
tmp.resize(n+1,L'\0');
n = GetEnvironmentVariableW(name.c_str(),&tmp[0],static_cast<unsigned>(tmp.size() - 1));
// The size may have changed
if(n >= tmp.size() - 1)
return 0;
ptr = &tmp[0];
}
if(!value.convert(ptr))
return 0;
return value.c_str();
}
///
/// \brief UTF-8 aware setenv, \a key - the variable name, \a value is a new UTF-8 value,
///
/// if override is not 0, that the old value is always overridded, otherwise,
/// if the variable exists it remains unchanged
///
inline int setenv(char const *key,char const *value,int override)
{
wshort_stackstring name;
if(!name.convert(key))
return -1;
if(!override) {
wchar_t unused[2];
if(!(GetEnvironmentVariableW(name.c_str(),unused,2)==0 && GetLastError() == 203)) // ERROR_ENVVAR_NOT_FOUND
return 0;
}
wstackstring wval;
if(!wval.convert(value))
return -1;
if(SetEnvironmentVariableW(name.c_str(),wval.c_str()))
return 0;
return -1;
}
///
/// \brief Remove enviroment variable \a key
///
inline int unsetenv(char const *key)
{
wshort_stackstring name;
if(!name.convert(key))
return -1;
if(SetEnvironmentVariableW(name.c_str(),0))
return 0;
return -1;
}
///
/// \brief UTF-8 aware putenv implementation, expects string in format KEY=VALUE
///
inline int putenv(char *string)
{
char const *key = string;
char const *key_end = string;
while(*key_end!='=' && key_end!='\0')
key_end++;
if(*key_end == '\0')
return -1;
wshort_stackstring wkey;
if(!wkey.convert(key,key_end))
return -1;
wstackstring wvalue;
if(!wvalue.convert(key_end+1))
return -1;
if(SetEnvironmentVariableW(wkey.c_str(),wvalue.c_str()))
return 0;
return -1;
}
#endif
} // nowide
#endif
///
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+37
View File
@@ -0,0 +1,37 @@
//
// Copyright (c) 2012 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_CONFIG_H_INCLUDED
#define NOWIDE_CONFIG_H_INCLUDED
#if (defined(__WIN32) || defined(_WIN32) || defined(WIN32)) && !defined(__CYGWIN__)
#define NOWIDE_WINDOWS
#endif
#ifdef _MSC_VER
#define NOWIDE_MSVC
#endif
#ifdef NOWIDE_WINDOWS
# if defined(DLL_EXPORT) || defined(NOWIDE_EXPORT)
# ifdef NOWIDE_SOURCE
# define NOWIDE_DECL __declspec(dllexport)
# else
# define NOWIDE_DECL __declspec(dllimport)
# endif //NOWIDE_SOURCE
# endif // DYN_LINK
#endif
#ifndef NOWIDE_DECL
# define NOWIDE_DECL
#endif
#endif
///
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+154
View File
@@ -0,0 +1,154 @@
//
// Copyright (c) 2012 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_CONVERT_H_INCLUDED
#define NOWIDE_CONVERT_H_INCLUDED
#include <string>
#include <nowide/encoding_utf.hpp>
namespace nowide {
///
/// \brief Template function that converts a buffer of UTF sequence in range [source_begin,source_end)
/// to the output \a buffer of size \a buffer_size.
///
/// In case of success a NUL terminated string returned (buffer), otherwise 0 is returned.
///
/// If there is not enough room in the buffer or the source sequence contains invalid UTF
/// 0 is returned, and the contend of the buffer is undefined.
///
template<typename CharOut,typename CharIn>
CharOut *basic_convert(CharOut *buffer,size_t buffer_size,CharIn const *source_begin,CharIn const *source_end)
{
CharOut *rv = buffer;
if(buffer_size == 0)
return 0;
buffer_size --;
while(source_begin!=source_end) {
using namespace nowide::utf;
code_point c = utf_traits<CharIn>::template decode<CharIn const *>(source_begin,source_end);
if(c==illegal || c==incomplete) {
rv = 0;
break;
}
size_t width = utf_traits<CharOut>::width(c);
if(buffer_size < width) {
rv=0;
break;
}
buffer = utf_traits<CharOut>::template encode<CharOut *>(c,buffer);
buffer_size -= width;
}
*buffer++ = 0;
return rv;
}
/// \cond INTERNAL
namespace details {
//
// wcslen defined only in C99... So we will not use it
//
template<typename Char>
Char const *basic_strend(Char const *s)
{
while(*s)
s++;
return s;
}
}
/// \endcond
///
/// Convert NUL terminated UTF source string to NUL terminated \a output string of size at
/// most output_size (including NUL)
///
/// In case of surcess output is returned, if the input sequence is illegal,
/// or there is not enough room NULL is returned
///
inline char *narrow(char *output,size_t output_size,wchar_t const *source)
{
return basic_convert(output,output_size,source,details::basic_strend(source));
}
///
/// Convert UTF text in range [begin,end) to NUL terminated \a output string of size at
/// most output_size (including NUL)
///
/// In case of surcess output is returned, if the input sequence is illegal,
/// or there is not enough room NULL is returned
///
inline char *narrow(char *output,size_t output_size,wchar_t const *begin,wchar_t const *end)
{
return basic_convert(output,output_size,begin,end);
}
///
/// Convert NUL terminated UTF source string to NUL terminated \a output string of size at
/// most output_size (including NUL)
///
/// In case of surcess output is returned, if the input sequence is illegal,
/// or there is not enough room NULL is returned
///
inline wchar_t *widen(wchar_t *output,size_t output_size,char const *source)
{
return basic_convert(output,output_size,source,details::basic_strend(source));
}
///
/// Convert UTF text in range [begin,end) to NUL terminated \a output string of size at
/// most output_size (including NUL)
///
/// In case of surcess output is returned, if the input sequence is illegal,
/// or there is not enough room NULL is returned
///
inline wchar_t *widen(wchar_t *output,size_t output_size,char const *begin,char const *end)
{
return basic_convert(output,output_size,begin,end);
}
///
/// Convert between Wide - UTF-16/32 string and UTF-8 string.
///
/// nowide::conv::conversion_error is thrown in a case of a error
///
inline std::string narrow(wchar_t const *s)
{
return nowide::conv::utf_to_utf<char>(s);
}
///
/// Convert between UTF-8 and UTF-16 string, implemented only on Windows platform
///
/// nowide::conv::conversion_error is thrown in a case of a error
///
inline std::wstring widen(char const *s)
{
return nowide::conv::utf_to_utf<wchar_t>(s);
}
///
/// Convert between Wide - UTF-16/32 string and UTF-8 string
///
/// nowide::conv::conversion_error is thrown in a case of a error
///
inline std::string narrow(std::wstring const &s)
{
return nowide::conv::utf_to_utf<char>(s);
}
///
/// Convert between UTF-8 and UTF-16 string, implemented only on Windows platform
///
/// nowide::conv::conversion_error is thrown in a case of a error
///
inline std::wstring widen(std::string const &s)
{
return nowide::conv::utf_to_utf<wchar_t>(s);
}
} // nowide
#endif
///
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+101
View File
@@ -0,0 +1,101 @@
//
// Copyright (c) 2012 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_CSTDIO_H_INCLUDED
#define NOWIDE_CSTDIO_H_INCLUDED
#include <cstdio>
#include <stdio.h>
#include <nowide/config.hpp>
#include <nowide/convert.hpp>
#include <nowide/stackstring.hpp>
#include <errno.h>
#ifdef NOWIDE_MSVC
# pragma warning(push)
# pragma warning(disable : 4996)
#endif
namespace nowide {
#if !defined(NOWIDE_WINDOWS) && !defined(NOWIDE_DOXYGEN)
using std::fopen;
using std::freopen;
using std::remove;
using std::rename;
#else
///
/// \brief Same as freopen but file_name and mode are UTF-8 strings
///
/// In invalid UTF-8 given, NULL is returned and errno is set to EINVAL
///
inline FILE *freopen(char const *file_name,char const *mode,FILE *stream)
{
wstackstring wname;
wshort_stackstring wmode;
if(!wname.convert(file_name) || !wmode.convert(mode)) {
errno = EINVAL;
return 0;
}
return _wfreopen(wname.c_str(),wmode.c_str(),stream);
}
///
/// \brief Same as fopen but file_name and mode are UTF-8 strings
///
/// In invalid UTF-8 given, NULL is returned and errno is set to EINVAL
///
inline FILE *fopen(char const *file_name,char const *mode)
{
wstackstring wname;
wshort_stackstring wmode;
if(!wname.convert(file_name) || !wmode.convert(mode)) {
errno = EINVAL;
return 0;
}
return _wfopen(wname.c_str(),wmode.c_str());
}
///
/// \brief Same as rename but old_name and new_name are UTF-8 strings
///
/// In invalid UTF-8 given, -1 is returned and errno is set to EINVAL
///
inline int rename(char const *old_name,char const *new_name)
{
wstackstring wold,wnew;
if(!wold.convert(old_name) || !wnew.convert(new_name)) {
errno = EINVAL;
return -1;
}
return _wrename(wold.c_str(),wnew.c_str());
}
///
/// \brief Same as rename but name is UTF-8 string
///
/// In invalid UTF-8 given, -1 is returned and errno is set to EINVAL
///
inline int remove(char const *name)
{
wstackstring wname;
if(!wname.convert(name)) {
errno = EINVAL;
return -1;
}
return _wremove(wname.c_str());
}
#endif
} // nowide
#ifdef NOWIDE_MSVC
#pragma warning(pop)
#endif
#endif
///
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+16
View File
@@ -0,0 +1,16 @@
//
// Copyright (c) 2012 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_CSTDLIB_HPP_INCLUDED
#define NOWIDE_CSTDLIB_HPP_INCLUDED
#include <nowide/cenv.hpp>
#include <nowide/system.hpp>
#endif
///
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+58
View File
@@ -0,0 +1,58 @@
//
// Copyright (c) 2009-2011 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_ENCODING_ERRORS_HPP_INCLUDED
#define NOWIDE_ENCODING_ERRORS_HPP_INCLUDED
#include <nowide/config.hpp>
#ifdef NOWIDE_MSVC
# pragma warning(push)
# pragma warning(disable : 4275 4251 4231 4660)
#endif
#include <stdexcept>
namespace nowide {
namespace conv {
///
/// \addtogroup codepage
///
/// @{
///
/// \brief The excepton that is thrown in case of conversion error
///
class conversion_error : public std::runtime_error {
public:
conversion_error() : std::runtime_error("Conversion failed") {}
};
///
/// enum that defines conversion policy
///
typedef enum {
skip = 0, ///< Skip illegal/unconvertable characters
stop = 1, ///< Stop conversion and throw conversion_error
default_method = skip ///< Default method - skip
} method_type;
/// @}
} // conv
} // nowide
#ifdef NOWIDE_MSVC
#pragma warning(pop)
#endif
#endif
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+84
View File
@@ -0,0 +1,84 @@
//
// Copyright (c) 2009-2011 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_ENCODING_UTF_HPP_INCLUDED
#define NOWIDE_ENCODING_UTF_HPP_INCLUDED
#include <nowide/utf.hpp>
#include <nowide/encoding_errors.hpp>
#include <string>
#include <iterator>
#ifdef NOWIDE_MSVC
# pragma warning(push)
# pragma warning(disable : 4275 4251 4231 4660)
#endif
namespace nowide{
namespace conv {
///
/// Convert a Unicode text in range [begin,end) to other Unicode encoding
///
template<typename CharOut,typename CharIn>
std::basic_string<CharOut>
utf_to_utf(CharIn const *begin,CharIn const *end,method_type how = default_method)
{
std::basic_string<CharOut> result;
result.reserve(end-begin);
typedef std::back_insert_iterator<std::basic_string<CharOut> > inserter_type;
inserter_type inserter(result);
utf::code_point c;
while(begin!=end) {
c=utf::utf_traits<CharIn>::template decode<CharIn const *>(begin,end);
if(c==utf::illegal || c==utf::incomplete) {
if(how==stop)
throw conversion_error();
}
else {
utf::utf_traits<CharOut>::template encode<inserter_type>(c,inserter);
}
}
return result;
}
///
/// Convert a Unicode NUL terminated string \a str other Unicode encoding
///
template<typename CharOut,typename CharIn>
std::basic_string<CharOut>
utf_to_utf(CharIn const *str,method_type how = default_method)
{
CharIn const *end = str;
while(*end)
end++;
return utf_to_utf<CharOut,CharIn>(str,end,how);
}
///
/// Convert a Unicode string \a str other Unicode encoding
///
template<typename CharOut,typename CharIn>
std::basic_string<CharOut>
utf_to_utf(std::basic_string<CharIn> const &str,method_type how = default_method)
{
return utf_to_utf<CharOut,CharIn>(str.c_str(),str.c_str()+str.size(),how);
}
} // conv
} // nowide
#ifdef NOWIDE_MSVC
#pragma warning(pop)
#endif
#endif
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+408
View File
@@ -0,0 +1,408 @@
//
// Copyright (c) 2012 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_FILEBUF_HPP
#define NOWIDE_FILEBUF_HPP
#include <iosfwd>
#include <nowide/config.hpp>
#include <nowide/stackstring.hpp>
#include <fstream>
#include <streambuf>
#include <stdio.h>
#ifdef NOWIDE_MSVC
# pragma warning(push)
# pragma warning(disable : 4996 4244)
#endif
namespace nowide {
#if !defined(NOWIDE_WINDOWS) && !defined(NOWIDE_FSTREAM_TESTS) && !defined(NOWIDE_DOXYGEN)
using std::basic_filebuf;
using std::filebuf;
#else // Windows
///
/// \brief This forward declaration defined the basic_filebuf type.
///
/// it is implemented and specialized for CharType = char, it behaves
/// implements std::filebuf over standard C I/O
///
template<typename CharType,typename Traits = std::char_traits<CharType> >
class basic_filebuf;
///
/// \brief This is implementation of std::filebuf
///
/// it is implemented and specialized for CharType = char, it behaves
/// implements std::filebuf over standard C I/O
///
template<>
class basic_filebuf<char> : public std::basic_streambuf<char> {
public:
///
/// Creates new filebuf
///
basic_filebuf() :
buffer_size_(4),
buffer_(0),
file_(0),
own_(true),
mode_(std::ios::in | std::ios::out)
{
setg(0,0,0);
setp(0,0);
}
virtual ~basic_filebuf()
{
if(file_) {
::fclose(file_);
file_ = 0;
}
if(own_ && buffer_)
delete [] buffer_;
}
///
/// Same as std::filebuf::open but s is UTF-8 string
///
basic_filebuf *open(std::string const &s,std::ios_base::openmode mode)
{
return open(s.c_str(),mode);
}
///
/// Same as std::filebuf::open but s is UTF-8 string
///
basic_filebuf *open(char const *s,std::ios_base::openmode mode)
{
if(file_) {
sync();
::fclose(file_);
file_ = 0;
}
wchar_t const *smode = get_mode(mode);
if(!smode)
return 0;
wstackstring name;
if(!name.convert(s))
return 0;
#ifdef NOWIDE_FSTREAM_TESTS
FILE *f = ::fopen(s,nowide::convert(smode).c_str());
#else
FILE *f = ::_wfopen(name.c_str(),smode);
#endif
if(!f)
return 0;
file_ = f;
return this;
}
///
/// Same as std::filebuf::close()
///
basic_filebuf *close()
{
bool res = sync() == 0;
if(file_) {
if(::fclose(file_)!=0)
res = false;
file_ = 0;
}
return res ? this : 0;
}
///
/// Same as std::filebuf::is_open()
///
bool is_open() const
{
return file_ != 0;
}
private:
void make_buffer()
{
if(buffer_)
return;
if(buffer_size_ > 0) {
buffer_ = new char [buffer_size_];
own_ = true;
}
}
protected:
virtual std::streambuf *setbuf(char *s,std::streamsize n)
{
if(!buffer_ && n>=0) {
buffer_ = s;
buffer_size_ = n;
own_ = false;
}
return this;
}
#ifdef NOWIDE_DEBUG_FILEBUF
void print_buf(char *b,char *p,char *e)
{
std::cerr << "-- Is Null: " << (b==0) << std::endl;;
if(b==0)
return;
if(e != 0)
std::cerr << "-- Total: " << e - b <<" offset from start " << p - b << std::endl;
else
std::cerr << "-- Total: " << p - b << std::endl;
std::cerr << "-- [";
for(char *ptr = b;ptr<p;ptr++)
std::cerr << *ptr;
if(e!=0) {
std::cerr << "|";
for(char *ptr = p;ptr<e;ptr++)
std::cerr << *ptr;
}
std::cerr << "]" << std::endl;
}
void print_state()
{
std::cerr << "- Output:" << std::endl;
print_buf(pbase(),pptr(),0);
std::cerr << "- Input:" << std::endl;
print_buf(eback(),gptr(),egptr());
std::cerr << "- fpos: " << (file_ ? ftell(file_) : -1L) << std::endl;
}
struct print_guard
{
print_guard(basic_filebuf *p,char const *func)
{
self = p;
f=func;
std::cerr << "In: " << f << std::endl;
self->print_state();
}
~print_guard()
{
std::cerr << "Out: " << f << std::endl;
self->print_state();
}
basic_filebuf *self;
char const *f;
};
#else
#endif
int overflow(int c)
{
#ifdef NOWIDE_DEBUG_FILEBUF
print_guard g(this,__FUNCTION__);
#endif
if(!file_)
return EOF;
if(fixg() < 0)
return EOF;
size_t n = pptr() - pbase();
if(n > 0) {
if(::fwrite(pbase(),1,n,file_) < n)
return -1;
fflush(file_);
}
if(buffer_size_ > 0) {
make_buffer();
setp(buffer_,buffer_+buffer_size_);
if(c!=EOF)
sputc(c);
}
else if(c!=EOF) {
if(::fputc(c,file_)==EOF)
return EOF;
fflush(file_);
}
return 0;
}
int sync()
{
return overflow(EOF);
}
int underflow()
{
#ifdef NOWIDE_DEBUG_FILEBUF
print_guard g(this,__FUNCTION__);
#endif
if(!file_)
return EOF;
if(fixp() < 0)
return EOF;
if(buffer_size_ == 0) {
int c = ::fgetc(file_);
if(c==EOF) {
return EOF;
}
last_char_ = c;
setg(&last_char_,&last_char_,&last_char_ + 1);
return c;
}
make_buffer();
size_t n = ::fread(buffer_,1,buffer_size_,file_);
setg(buffer_,buffer_,buffer_+n);
if(n == 0)
return EOF;
return std::char_traits<char>::to_int_type(*gptr());
}
int pbackfail(int)
{
return pubseekoff(-1,std::ios::cur);
}
std::streampos seekoff(std::streamoff off,
std::ios_base::seekdir seekdir,
std::ios_base::openmode /*m*/)
{
#ifdef NOWIDE_DEBUG_FILEBUF
print_guard g(this,__FUNCTION__);
#endif
if(!file_)
return EOF;
if(fixp() < 0 || fixg() < 0)
return EOF;
if(seekdir == std::ios_base::cur) {
if( ::fseek(file_,off,SEEK_CUR) < 0)
return EOF;
}
else if(seekdir == std::ios_base::beg) {
if( ::fseek(file_,off,SEEK_SET) < 0)
return EOF;
}
else if(seekdir == std::ios_base::end) {
if( ::fseek(file_,off,SEEK_END) < 0)
return EOF;
}
else
return -1;
return ftell(file_);
}
std::streampos seekpos(std::streampos off,std::ios_base::openmode m)
{
return seekoff(std::streamoff(off),std::ios_base::beg,m);
}
private:
int fixg()
{
if(gptr()!=egptr()) {
std::streamsize off = gptr() - egptr();
setg(0,0,0);
if(fseek(file_,off,SEEK_CUR) != 0)
return -1;
}
setg(0,0,0);
return 0;
}
int fixp()
{
if(pptr()!=0) {
int r = sync();
setp(0,0);
return r;
}
return 0;
}
void reset(FILE *f = 0)
{
sync();
if(file_) {
fclose(file_);
file_ = 0;
}
file_ = f;
}
static wchar_t const *get_mode(std::ios_base::openmode mode)
{
//
// done according to n2914 table 106 27.9.1.4
//
// note can't use switch case as overload operator can't be used
// in constant expression
if(mode == (std::ios_base::out))
return L"w";
if(mode == (std::ios_base::out | std::ios_base::app))
return L"a";
if(mode == (std::ios_base::app))
return L"a";
if(mode == (std::ios_base::out | std::ios_base::trunc))
return L"w";
if(mode == (std::ios_base::in))
return L"r";
if(mode == (std::ios_base::in | std::ios_base::out))
return L"r+";
if(mode == (std::ios_base::in | std::ios_base::out | std::ios_base::trunc))
return L"w+";
if(mode == (std::ios_base::in | std::ios_base::out | std::ios_base::app))
return L"a+";
if(mode == (std::ios_base::in | std::ios_base::app))
return L"a+";
if(mode == (std::ios_base::binary | std::ios_base::out))
return L"wb";
if(mode == (std::ios_base::binary | std::ios_base::out | std::ios_base::app))
return L"ab";
if(mode == (std::ios_base::binary | std::ios_base::app))
return L"ab";
if(mode == (std::ios_base::binary | std::ios_base::out | std::ios_base::trunc))
return L"wb";
if(mode == (std::ios_base::binary | std::ios_base::in))
return L"rb";
if(mode == (std::ios_base::binary | std::ios_base::in | std::ios_base::out))
return L"r+b";
if(mode == (std::ios_base::binary | std::ios_base::in | std::ios_base::out | std::ios_base::trunc))
return L"w+b";
if(mode == (std::ios_base::binary | std::ios_base::in | std::ios_base::out | std::ios_base::app))
return L"a+b";
if(mode == (std::ios_base::binary | std::ios_base::in | std::ios_base::app))
return L"a+b";
return 0;
}
size_t buffer_size_;
char *buffer_;
FILE *file_;
bool own_;
char last_char_;
std::ios::openmode mode_;
};
///
/// \brief Convinience typedef
///
typedef basic_filebuf<char> filebuf;
#endif // windows
} // nowide
#ifdef NOWIDE_MSVC
# pragma warning(pop)
#endif
#endif
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+248
View File
@@ -0,0 +1,248 @@
//
// Copyright (c) 2012 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_FSTREAM_INCLUDED_HPP
#define NOWIDE_FSTREAM_INCLUDED_HPP
#include <iosfwd>
#include <nowide/config.hpp>
#include <nowide/convert.hpp>
#include <nowide/scoped_ptr.hpp>
#include <fstream>
#include <memory>
#include <nowide/filebuf.hpp>
///
/// \brief This namespace includes implementation of the standard library functios
/// such that they accept UTF-8 strings on Windows. On other platforms it is just an alias
/// of std namespace (i.e. not on Windows)
///
namespace nowide {
#if !defined(NOWIDE_WINDOWS) && !defined(NOWIDE_FSTREAM_TESTS) && !defined(NOWIDE_DOXYGEN)
using std::basic_ifstream;
using std::basic_ofstream;
using std::basic_fstream;
using std::ifstream;
using std::ofstream;
using std::fstream;
#else
///
/// \brief Same as std::basic_ifstream<char> but accepts UTF-8 strings under Windows
///
template<typename CharType,typename Traits = std::char_traits<CharType> >
class basic_ifstream : public std::basic_istream<CharType,Traits>
{
public:
typedef basic_filebuf<CharType,Traits> internal_buffer_type;
typedef std::basic_istream<CharType,Traits> internal_stream_type;
basic_ifstream() :
internal_stream_type(0)
{
buf_.reset(new internal_buffer_type());
std::ios::rdbuf(buf_.get());
}
explicit basic_ifstream(char const *file_name,std::ios_base::openmode mode = std::ios_base::in) :
internal_stream_type(0)
{
buf_.reset(new internal_buffer_type());
std::ios::rdbuf(buf_.get());
open(file_name,mode);
}
void open(char const *file_name,std::ios_base::openmode mode = std::ios_base::in)
{
if(!buf_->open(file_name,mode | std::ios_base::in)) {
this->setstate(std::ios_base::failbit);
}
else {
this->clear();
}
}
bool is_open()
{
return buf_->is_open();
}
bool is_open() const
{
return buf_->is_open();
}
void close()
{
if(!buf_->close())
this->setstate(std::ios_base::failbit);
else
this->clear();
}
internal_buffer_type *rdbuf() const
{
return buf_.get();
}
~basic_ifstream()
{
buf_->close();
}
private:
nowide::scoped_ptr<internal_buffer_type> buf_;
};
///
/// \brief Same as std::basic_ofstream<char> but accepts UTF-8 strings under Windows
///
template<typename CharType,typename Traits = std::char_traits<CharType> >
class basic_ofstream : public std::basic_ostream<CharType,Traits>
{
public:
typedef basic_filebuf<CharType,Traits> internal_buffer_type;
typedef std::basic_ostream<CharType,Traits> internal_stream_type;
basic_ofstream() :
internal_stream_type(0)
{
buf_.reset(new internal_buffer_type());
std::ios::rdbuf(buf_.get());
}
explicit basic_ofstream(char const *file_name,std::ios_base::openmode mode = std::ios_base::out) :
internal_stream_type(0)
{
buf_.reset(new internal_buffer_type());
std::ios::rdbuf(buf_.get());
open(file_name,mode);
}
void open(char const *file_name,std::ios_base::openmode mode = std::ios_base::out)
{
if(!buf_->open(file_name,mode | std::ios_base::out)) {
this->setstate(std::ios_base::failbit);
}
else {
this->clear();
}
}
bool is_open()
{
return buf_->is_open();
}
bool is_open() const
{
return buf_->is_open();
}
void close()
{
if(!buf_->close())
this->setstate(std::ios_base::failbit);
else
this->clear();
}
internal_buffer_type *rdbuf() const
{
return buf_.get();
}
~basic_ofstream()
{
buf_->close();
}
private:
nowide::scoped_ptr<internal_buffer_type> buf_;
};
///
/// \brief Same as std::basic_fstream<char> but accepts UTF-8 strings under Windows
///
template<typename CharType,typename Traits = std::char_traits<CharType> >
class basic_fstream : public std::basic_iostream<CharType,Traits>
{
public:
typedef basic_filebuf<CharType,Traits> internal_buffer_type;
typedef std::basic_iostream<CharType,Traits> internal_stream_type;
basic_fstream() :
internal_stream_type(0)
{
buf_.reset(new internal_buffer_type());
std::ios::rdbuf(buf_.get());
}
explicit basic_fstream(char const *file_name,std::ios_base::openmode mode = std::ios_base::out | std::ios_base::in) :
internal_stream_type(0)
{
buf_.reset(new internal_buffer_type());
std::ios::rdbuf(buf_.get());
open(file_name,mode);
}
void open(char const *file_name,std::ios_base::openmode mode = std::ios_base::out | std::ios_base::out)
{
if(!buf_->open(file_name,mode)) {
this->setstate(std::ios_base::failbit);
}
else {
this->clear();
}
}
bool is_open()
{
return buf_->is_open();
}
bool is_open() const
{
return buf_->is_open();
}
void close()
{
if(!buf_->close())
this->setstate(std::ios_base::failbit);
else
this->clear();
}
internal_buffer_type *rdbuf() const
{
return buf_.get();
}
~basic_fstream()
{
buf_->close();
}
private:
nowide::scoped_ptr<internal_buffer_type> buf_;
};
///
/// \brief Same as std::filebuf but accepts UTF-8 strings under Windows
///
typedef basic_filebuf<char> filebuf;
///
/// Same as std::ifstream but accepts UTF-8 strings under Windows
///
typedef basic_ifstream<char> ifstream;
///
/// Same as std::ofstream but accepts UTF-8 strings under Windows
///
typedef basic_ofstream<char> ofstream;
///
/// Same as std::fstream but accepts UTF-8 strings under Windows
///
typedef basic_fstream<char> fstream;
#endif
} // nowide
#endif
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+99
View File
@@ -0,0 +1,99 @@
//
// Copyright (c) 2012 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_IOSTREAM_HPP_INCLUDED
#define NOWIDE_IOSTREAM_HPP_INCLUDED
#include <nowide/config.hpp>
#include <nowide/scoped_ptr.hpp>
#include <iostream>
#include <ostream>
#include <istream>
#ifdef NOWIDE_MSVC
# pragma warning(push)
# pragma warning(disable : 4251)
#endif
namespace nowide {
#if !defined(NOWIDE_WINDOWS) && !defined(NOWIDE_DOXYGEN)
using std::cout;
using std::cerr;
using std::cin;
using std::clog;
#else
/// \cond INTERNAL
namespace details {
class console_output_buffer;
class console_input_buffer;
class NOWIDE_DECL winconsole_ostream : public std::ostream {
winconsole_ostream(winconsole_ostream const &);
void operator=(winconsole_ostream const &);
public:
winconsole_ostream(int fd);
~winconsole_ostream();
private:
nowide::scoped_ptr<console_output_buffer> d;
};
class NOWIDE_DECL winconsole_istream : public std::istream {
winconsole_istream(winconsole_istream const &);
void operator=(winconsole_istream const &);
public:
winconsole_istream();
~winconsole_istream();
private:
struct data;
nowide::scoped_ptr<console_input_buffer> d;
};
} // details
/// \endcond
///
/// \brief Same as std::cin, but uses UTF-8
///
/// Note, the stream is not synchronized with stdio and not affected by std::ios::sync_with_stdio
///
extern NOWIDE_DECL details::winconsole_istream cin;
///
/// \brief Same as std::cout, but uses UTF-8
///
/// Note, the stream is not synchronized with stdio and not affected by std::ios::sync_with_stdio
///
extern NOWIDE_DECL details::winconsole_ostream cout;
///
/// \brief Same as std::cerr, but uses UTF-8
///
/// Note, the stream is not synchronized with stdio and not affected by std::ios::sync_with_stdio
///
extern NOWIDE_DECL details::winconsole_ostream cerr;
///
/// \brief Same as std::clog, but uses UTF-8
///
/// Note, the stream is not synchronized with stdio and not affected by std::ios::sync_with_stdio
///
extern NOWIDE_DECL details::winconsole_ostream clog;
#endif
} // nowide
#ifdef NOWIDE_MSVC
# pragma warning(pop)
#endif
#endif
///
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+93
View File
@@ -0,0 +1,93 @@
#ifndef NOWIDE_SCOPED_PTR_HPP
#define NOWIDE_SCOPED_PTR_HPP
// (C) Copyright Greg Colvin and Beman Dawes 1998, 1999.
// Copyright (c) 2001, 2002 Peter Dimov,
// Copyright (C) 2012 Artyom Beilis
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
// http://www.boost.org/libs/smart_ptr/scoped_ptr.htm
//
#include <assert.h>
namespace nowide
{
// scoped_ptr mimics a built-in pointer except that it guarantees deletion
// of the object pointed to, either on destruction of the scoped_ptr or via
// an explicit reset(). scoped_ptr is a simple solution for simple needs;
// use shared_ptr or std::auto_ptr if your needs are more complex.
template<class T> class scoped_ptr // noncopyable
{
private:
T * px;
scoped_ptr(scoped_ptr const &);
scoped_ptr & operator=(scoped_ptr const &);
typedef scoped_ptr<T> this_type;
void operator==( scoped_ptr const& ) const;
void operator!=( scoped_ptr const& ) const;
public:
typedef T element_type;
explicit scoped_ptr( T * p = 0 ): px( p ) // never throws
{
}
~scoped_ptr() // never throws
{
delete px;
}
void reset(T * p = 0) // never throws
{
assert( p == 0 || p != px ); // catch self-reset errors
this_type(p).swap(*this);
}
T & operator*() const // never throws
{
assert( px != 0 );
return *px;
}
T * operator->() const // never throws
{
assert( px != 0 );
return px;
}
T * get() const // never throws
{
return px;
}
operator bool() const
{
return px!=0;
}
void swap(scoped_ptr & b) // never throws
{
T * tmp = b.px;
b.px = px;
px = tmp;
}
};
} // namespace nowide
#endif
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+154
View File
@@ -0,0 +1,154 @@
//
// Copyright (c) 2012 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_DETAILS_WIDESTR_H_INCLUDED
#define NOWIDE_DETAILS_WIDESTR_H_INCLUDED
#include <nowide/convert.hpp>
#include <string.h>
#include <algorithm>
namespace nowide {
///
/// \brief A class that allows to create a temporary wide or narrow UTF strings from
/// wide or narrow UTF source.
///
/// It uses on stack buffer of the string is short enough
/// and allocated a buffer on the heap if the size of the buffer is too small
///
template<typename CharOut=wchar_t,typename CharIn = char,size_t BufferSize = 256>
class basic_stackstring {
public:
static const size_t buffer_size = BufferSize;
typedef CharOut output_char;
typedef CharIn input_char;
basic_stackstring(basic_stackstring const &other) :
mem_buffer_(0)
{
clear();
if(other.mem_buffer_) {
size_t len = 0;
while(other.mem_buffer_[len])
len ++;
mem_buffer_ = new output_char[len + 1];
memcpy(mem_buffer_,other.mem_buffer_,sizeof(output_char) * (len+1));
}
else {
memcpy(buffer_,other.buffer_,buffer_size * sizeof(output_char));
}
}
void swap(basic_stackstring &other)
{
std::swap(mem_buffer_,other.mem_buffer_);
for(size_t i=0;i<buffer_size;i++)
std::swap(buffer_[i],other.buffer_[i]);
}
basic_stackstring &operator=(basic_stackstring const &other)
{
if(this != &other) {
basic_stackstring tmp(other);
swap(tmp);
}
return *this;
}
basic_stackstring() : mem_buffer_(0)
{
}
bool convert(input_char const *input)
{
return convert(input,details::basic_strend(input));
}
bool convert(input_char const *begin,input_char const *end)
{
clear();
size_t space = get_space(sizeof(input_char),sizeof(output_char),end - begin) + 1;
if(space <= buffer_size) {
if(basic_convert(buffer_,buffer_size,begin,end))
return true;
clear();
return false;
}
else {
mem_buffer_ = new output_char[space];
if(!basic_convert(mem_buffer_,space,begin,end)) {
clear();
return false;
}
return true;
}
}
output_char *c_str()
{
if(mem_buffer_)
return mem_buffer_;
return buffer_;
}
output_char const *c_str() const
{
if(mem_buffer_)
return mem_buffer_;
return buffer_;
}
void clear()
{
if(mem_buffer_) {
delete [] mem_buffer_;
mem_buffer_=0;
}
buffer_[0] = 0;
}
~basic_stackstring()
{
clear();
}
private:
static size_t get_space(size_t insize,size_t outsize,size_t in)
{
if(insize <= outsize)
return in;
else if(insize == 2 && outsize == 1)
return 3 * in;
else if(insize == 4 && outsize == 1)
return 4 * in;
else // if(insize == 4 && outsize == 2)
return 2 * in;
}
output_char buffer_[buffer_size];
output_char *mem_buffer_;
}; //basic_stackstring
///
/// Convinience typedef
///
typedef basic_stackstring<wchar_t,char,256> wstackstring;
///
/// Convinience typedef
///
typedef basic_stackstring<char,wchar_t,256> stackstring;
///
/// Convinience typedef
///
typedef basic_stackstring<wchar_t,char,16> wshort_stackstring;
///
/// Convinience typedef
///
typedef basic_stackstring<char,wchar_t,16> short_stackstring;
} // nowide
#endif
///
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+46
View File
@@ -0,0 +1,46 @@
//
// Copyright (c) 2012 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_CSTDLIB_HPP
#define NOWIDE_CSTDLIB_HPP
#include <stdlib.h>
#include <errno.h>
#include <nowide/stackstring.hpp>
namespace nowide {
#if !defined(NOWIDE_WINDOWS) && !defined(NOWIDE_DOXYGEN)
using ::system;
#else // Windows
///
/// Same as std::system but cmd is UTF-8.
///
/// If the input is not valid UTF-8, -1 returned and errno set to EINVAL
///
inline int system(char const *cmd)
{
if(!cmd)
return _wsystem(0);
wstackstring wcmd;
if(!wcmd.convert(cmd)) {
errno = EINVAL;
return -1;
}
return _wsystem(wcmd.c_str());
}
#endif
} // nowide
#endif
///
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+469
View File
@@ -0,0 +1,469 @@
//
// Copyright (c) 2009-2011 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_UTF_HPP_INCLUDED
#define NOWIDE_UTF_HPP_INCLUDED
#include <nowide/config.hpp>
#ifndef NOWIDE_MSVC
namespace nowide {
namespace utf {
typedef unsigned uint32_t;
typedef unsigned short uint16_t;
typedef unsigned char uint8_t;
}
}
#else
#include <stdint.h>
#endif
namespace nowide {
///
/// \brief Namespace that holds basic operations on UTF encoded sequences
///
/// All functions defined in this namespace do not require linking with Boost.Locale library
///
namespace utf {
/// \cond INTERNAL
#ifdef __GNUC__
# define NOWIDE_LIKELY(x) __builtin_expect((x),1)
# define NOWIDE_UNLIKELY(x) __builtin_expect((x),0)
#else
# define NOWIDE_LIKELY(x) (x)
# define NOWIDE_UNLIKELY(x) (x)
#endif
/// \endcond
///
/// \brief The integral type type that can hold a Unicode code point
///
typedef uint32_t code_point;
///
/// \brief Special constant that defines illegal code point
///
static const code_point illegal = 0xFFFFFFFFu;
///
/// \brief Special constant that defines incomplete code point
///
static const code_point incomplete = 0xFFFFFFFEu;
///
/// \brief the function checks if \a v is a valid code point
///
inline bool is_valid_codepoint(code_point v)
{
if(v>0x10FFFF)
return false;
if(0xD800 <=v && v<= 0xDFFF) // surragates
return false;
return true;
}
#ifdef NOWIDE_DOXYGEN
///
/// \brief UTF Traits class - functions to convert UTF sequences to and from Unicode code points
///
template<typename CharType,int size=sizeof(CharType)>
struct utf_traits {
///
/// The type of the character
///
typedef CharType char_type;
///
/// Read one code point from the range [p,e) and return it.
///
/// - If the sequence that was read is incomplete sequence returns \ref incomplete,
/// - If illegal sequence detected returns \ref illegal
///
/// Requirements
///
/// - Iterator is valid input iterator
///
/// Postconditions
///
/// - p points to the last consumed character
///
template<typename Iterator>
static code_point decode(Iterator &p,Iterator e);
///
/// Maximal width of valid sequence in the code units:
///
/// - UTF-8 - 4
/// - UTF-16 - 2
/// - UTF-32 - 1
///
static const int max_width;
///
/// The width of specific code point in the code units.
///
/// Requirement: value is a valid Unicode code point
/// Returns value in range [1..max_width]
///
static int width(code_point value);
///
/// Get the size of the trail part of variable length encoded sequence.
///
/// Returns -1 if C is not valid lead character
///
static int trail_length(char_type c);
///
/// Returns true if c is trail code unit, always false for UTF-32
///
static bool is_trail(char_type c);
///
/// Returns true if c is lead code unit, always true of UTF-32
///
static bool is_lead(char_type c);
///
/// Convert valid Unicode code point \a value to the UTF sequence.
///
/// Requirements:
///
/// - \a value is valid code point
/// - \a out is an output iterator should be able to accept at least width(value) units
///
/// Returns the iterator past the last written code unit.
///
template<typename Iterator>
static Iterator encode(code_point value,Iterator out);
///
/// Decodes valid UTF sequence that is pointed by p into code point.
///
/// If the sequence is invalid or points to end the behavior is undefined
///
template<typename Iterator>
static code_point decode_valid(Iterator &p);
};
#else
template<typename CharType,int size=sizeof(CharType)>
struct utf_traits;
template<typename CharType>
struct utf_traits<CharType,1> {
typedef CharType char_type;
static int trail_length(char_type ci)
{
unsigned char c = ci;
if(c < 128)
return 0;
if(NOWIDE_UNLIKELY(c < 194))
return -1;
if(c < 224)
return 1;
if(c < 240)
return 2;
if(NOWIDE_LIKELY(c <=244))
return 3;
return -1;
}
static const int max_width = 4;
static int width(code_point value)
{
if(value <=0x7F) {
return 1;
}
else if(value <=0x7FF) {
return 2;
}
else if(NOWIDE_LIKELY(value <=0xFFFF)) {
return 3;
}
else {
return 4;
}
}
static bool is_trail(char_type ci)
{
unsigned char c=ci;
return (c & 0xC0)==0x80;
}
static bool is_lead(char_type ci)
{
return !is_trail(ci);
}
template<typename Iterator>
static code_point decode(Iterator &p,Iterator e)
{
if(NOWIDE_UNLIKELY(p==e))
return incomplete;
unsigned char lead = *p++;
// First byte is fully validated here
int trail_size = trail_length(lead);
if(NOWIDE_UNLIKELY(trail_size < 0))
return illegal;
//
// Ok as only ASCII may be of size = 0
// also optimize for ASCII text
//
if(trail_size == 0)
return lead;
code_point c = lead & ((1<<(6-trail_size))-1);
// Read the rest
unsigned char tmp;
switch(trail_size) {
case 3:
if(NOWIDE_UNLIKELY(p==e))
return incomplete;
tmp = *p++;
if (!is_trail(tmp))
return illegal;
c = (c << 6) | ( tmp & 0x3F);
case 2:
if(NOWIDE_UNLIKELY(p==e))
return incomplete;
tmp = *p++;
if (!is_trail(tmp))
return illegal;
c = (c << 6) | ( tmp & 0x3F);
case 1:
if(NOWIDE_UNLIKELY(p==e))
return incomplete;
tmp = *p++;
if (!is_trail(tmp))
return illegal;
c = (c << 6) | ( tmp & 0x3F);
}
// Check code point validity: no surrogates and
// valid range
if(NOWIDE_UNLIKELY(!is_valid_codepoint(c)))
return illegal;
// make sure it is the most compact representation
if(NOWIDE_UNLIKELY(width(c)!=trail_size + 1))
return illegal;
return c;
}
template<typename Iterator>
static code_point decode_valid(Iterator &p)
{
unsigned char lead = *p++;
if(lead < 192)
return lead;
int trail_size;
if(lead < 224)
trail_size = 1;
else if(NOWIDE_LIKELY(lead < 240)) // non-BMP rare
trail_size = 2;
else
trail_size = 3;
code_point c = lead & ((1<<(6-trail_size))-1);
switch(trail_size) {
case 3:
c = (c << 6) | ( static_cast<unsigned char>(*p++) & 0x3F);
case 2:
c = (c << 6) | ( static_cast<unsigned char>(*p++) & 0x3F);
case 1:
c = (c << 6) | ( static_cast<unsigned char>(*p++) & 0x3F);
}
return c;
}
template<typename Iterator>
static Iterator encode(code_point value,Iterator out)
{
if(value <= 0x7F) {
*out++ = static_cast<char_type>(value);
}
else if(value <= 0x7FF) {
*out++ = static_cast<char_type>((value >> 6) | 0xC0);
*out++ = static_cast<char_type>((value & 0x3F) | 0x80);
}
else if(NOWIDE_LIKELY(value <= 0xFFFF)) {
*out++ = static_cast<char_type>((value >> 12) | 0xE0);
*out++ = static_cast<char_type>(((value >> 6) & 0x3F) | 0x80);
*out++ = static_cast<char_type>((value & 0x3F) | 0x80);
}
else {
*out++ = static_cast<char_type>((value >> 18) | 0xF0);
*out++ = static_cast<char_type>(((value >> 12) & 0x3F) | 0x80);
*out++ = static_cast<char_type>(((value >> 6) & 0x3F) | 0x80);
*out++ = static_cast<char_type>((value & 0x3F) | 0x80);
}
return out;
}
}; // utf8
template<typename CharType>
struct utf_traits<CharType,2> {
typedef CharType char_type;
// See RFC 2781
static bool is_first_surrogate(uint16_t x)
{
return 0xD800 <=x && x<= 0xDBFF;
}
static bool is_second_surrogate(uint16_t x)
{
return 0xDC00 <=x && x<= 0xDFFF;
}
static code_point combine_surrogate(uint16_t w1,uint16_t w2)
{
return ((code_point(w1 & 0x3FF) << 10) | (w2 & 0x3FF)) + 0x10000;
}
static int trail_length(char_type c)
{
if(is_first_surrogate(c))
return 1;
if(is_second_surrogate(c))
return -1;
return 0;
}
///
/// Returns true if c is trail code unit, always false for UTF-32
///
static bool is_trail(char_type c)
{
return is_second_surrogate(c);
}
///
/// Returns true if c is lead code unit, always true of UTF-32
///
static bool is_lead(char_type c)
{
return !is_second_surrogate(c);
}
template<typename It>
static code_point decode(It &current,It last)
{
if(NOWIDE_UNLIKELY(current == last))
return incomplete;
uint16_t w1=*current++;
if(NOWIDE_LIKELY(w1 < 0xD800 || 0xDFFF < w1)) {
return w1;
}
if(w1 > 0xDBFF)
return illegal;
if(current==last)
return incomplete;
uint16_t w2=*current++;
if(w2 < 0xDC00 || 0xDFFF < w2)
return illegal;
return combine_surrogate(w1,w2);
}
template<typename It>
static code_point decode_valid(It &current)
{
uint16_t w1=*current++;
if(NOWIDE_LIKELY(w1 < 0xD800 || 0xDFFF < w1)) {
return w1;
}
uint16_t w2=*current++;
return combine_surrogate(w1,w2);
}
static const int max_width = 2;
static int width(code_point u)
{
return u>=0x10000 ? 2 : 1;
}
template<typename It>
static It encode(code_point u,It out)
{
if(NOWIDE_LIKELY(u<=0xFFFF)) {
*out++ = static_cast<char_type>(u);
}
else {
u -= 0x10000;
*out++ = static_cast<char_type>(0xD800 | (u>>10));
*out++ = static_cast<char_type>(0xDC00 | (u & 0x3FF));
}
return out;
}
}; // utf16;
template<typename CharType>
struct utf_traits<CharType,4> {
typedef CharType char_type;
static int trail_length(char_type c)
{
if(is_valid_codepoint(c))
return 0;
return -1;
}
static bool is_trail(char_type /*c*/)
{
return false;
}
static bool is_lead(char_type /*c*/)
{
return true;
}
template<typename It>
static code_point decode_valid(It &current)
{
return *current++;
}
template<typename It>
static code_point decode(It &current,It last)
{
if(NOWIDE_UNLIKELY(current == last))
return nowide::utf::incomplete;
code_point c=*current++;
if(NOWIDE_UNLIKELY(!is_valid_codepoint(c)))
return nowide::utf::illegal;
return c;
}
static const int max_width = 1;
static int width(code_point /*u*/)
{
return 1;
}
template<typename It>
static It encode(code_point u,It out)
{
*out++ = static_cast<char_type>(u);
return out;
}
}; // utf32
#endif
} // utf
} // nowide
#endif
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4
+39
View File
@@ -0,0 +1,39 @@
//
// Copyright (c) 2012 Artyom Beilis (Tonkikh)
//
// Distributed under the Boost Software License, Version 1.0. (See
// accompanying file LICENSE_1_0.txt or copy at
// http://www.boost.org/LICENSE_1_0.txt)
//
#ifndef NOWIDE_WINDOWS_HPP_INCLUDED
#define NOWIDE_WINDOWS_HPP_INCLUDED
#include <stddef.h>
#ifdef NOWIDE_USE_WINDOWS_H
#include <windows.h>
#else
//
// These are function prototypes... Allow to to include windows.h
//
extern "C" {
__declspec(dllimport) wchar_t* __stdcall GetEnvironmentStringsW(void);
__declspec(dllimport) int __stdcall FreeEnvironmentStringsW(wchar_t *);
__declspec(dllimport) wchar_t* __stdcall GetCommandLineW(void);
__declspec(dllimport) wchar_t** __stdcall CommandLineToArgvW(wchar_t const *,int *);
__declspec(dllimport) unsigned long __stdcall GetLastError();
__declspec(dllimport) void* __stdcall LocalFree(void *);
__declspec(dllimport) int __stdcall SetEnvironmentVariableW(wchar_t const *,wchar_t const *);
__declspec(dllimport) unsigned long __stdcall GetEnvironmentVariableW(wchar_t const *,wchar_t *,unsigned long);
}
#endif
#endif
///
// vim: tabstop=4 expandtab shiftwidth=4 softtabstop=4