added ifdefs so we can build with ARABICA_NO_WCHAR_T

This commit is contained in:
jez_higgins 2003-10-03 05:56:26 +00:00
parent 5075dc5b7a
commit 84b93c75c9

View file

@ -1,5 +1,9 @@
// transcode.cpp
//
// Warning! Contains ifdef hackery
//
///////////////////////////////////
#include <iostream>
#include <Utils/convert_adaptor.h>
#include <Utils/utf8iso88591codecvt.h>
@ -13,21 +17,32 @@
using namespace Arabica::convert;
iconvert_adaptor<char> iByteConvertor(std::cin);
oconvert_adaptor<char> oByteConvertor(std::cout);
#ifndef ARABICA_NO_WCHAR_T
typedef iconvert_adaptor<wchar_t, std::char_traits<wchar_t>, char, std::char_traits<char> > Widener;
typedef oconvert_adaptor<wchar_t, std::char_traits<wchar_t>, char, std::char_traits<char> > Narrower;
iconvert_adaptor<char> iByteConvertor(std::cin);
Widener iCharAdaptor(iByteConvertor);
oconvert_adaptor<char> oByteConvertor(std::cout);
Narrower oCharAdaptor(oByteConvertor);
#endif
void transcode();
void imbueCodecvts(int argc, const char* argv[]);
int main(int argc, const char* argv[])
{
imbueCodecvts(argc, argv);
transcode();
return 0;
} // main
#ifndef ARABICA_NO_WCHAR_T
void transcode()
{
int count = 0;
wchar_t c = iCharAdaptor.get();
while(!iCharAdaptor.eof())
@ -44,14 +59,35 @@ int main(int argc, const char* argv[])
c = iCharAdaptor.get();
}
oCharAdaptor.flush();
oByteConvertor.flush();
} // transcode
#else
void transcode()
{
int count = 0;
char c = iByteConvertor.get();
while(!iByteConvertor.eof())
{
oByteConvertor << c;
if(count == 1024)
{
oByteConvertor.flush();
count = 0;
} // if ...
c = iByteConvertor.get();
} // while
oByteConvertor.flush();
return 0;
} // main
} // transcode
#endif
void imbueCodecvts(int argc, const char* argv[])
{
#ifndef ARABICA_NO_WCHAR_T
oCharAdaptor.imbue(std::locale(oCharAdaptor.getloc(), new utf16beucs2codecvt()));
#endif
for(int i = 1; i < argc; ++i)
{
@ -82,12 +118,14 @@ void imbueCodecvts(int argc, const char* argv[])
iByteConvertor.imbue(std::locale(iByteConvertor.getloc(), new rot13codecvt()));
else if(cvt == "base64")
iByteConvertor.imbue(std::locale(iByteConvertor.getloc(), new base64codecvt()));
#ifndef ARABICA_NO_WCHAR_T
else if(cvt == "utf8")
iCharAdaptor.imbue(std::locale(iCharAdaptor.getloc(), new utf8ucs2codecvt()));
else if(cvt == "utf16be")
iCharAdaptor.imbue(std::locale(iCharAdaptor.getloc(), new utf16beucs2codecvt()));
else if(cvt == "utf16le")
iCharAdaptor.imbue(std::locale(iCharAdaptor.getloc(), new utf16leucs2codecvt()));
#endif
else
{
std::cerr << cvt << " is not a valid input encoding." << std::endl;
@ -100,12 +138,14 @@ void imbueCodecvts(int argc, const char* argv[])
oByteConvertor.imbue(std::locale(oByteConvertor.getloc(), new rot13codecvt()));
else if(cvt == "base64")
oByteConvertor.imbue(std::locale(oByteConvertor.getloc(), new base64codecvt()));
#ifndef ARABICA_NO_WCHAR_T
else if(cvt == "utf8")
oCharAdaptor.imbue(std::locale(oCharAdaptor.getloc(), new utf8ucs2codecvt()));
else if(cvt == "utf16be")
oCharAdaptor.imbue(std::locale(oCharAdaptor.getloc(), new utf16beucs2codecvt()));
else if(cvt == "utf16le")
oCharAdaptor.imbue(std::locale(oCharAdaptor.getloc(), new utf16leucs2codecvt()));
#endif
else
{
std::cerr << cvt << " is not a valid output encoding." << std::endl;