mirror of
https://github.com/yanyiwu/cppjieba.git
synced 2025-07-18 00:00:12 +08:00
update cppcommon for gbkToUtf8
This commit is contained in:
parent
c10d3df523
commit
dd48f14959
@ -28,7 +28,7 @@ file_functs.test: file_functs.cpp file_functs.h
|
||||
io_functs.test: io_functs.cpp io_functs.h
|
||||
g++ -o $@ $< -DTEST_IO_FUNCTS
|
||||
str_functs.ut: str_functs.cpp str_functs.h
|
||||
g++ -o $@ $< -DTEST_STR_FUNCTS
|
||||
g++ -o $@ $< -DTEST_STR_FUNCTS -liconv
|
||||
vec_functs.test: vec_functs.cpp vec_functs.h vec_functs.tcc
|
||||
g++ -o $@ $< -DTEST_VEC_FUNCTS
|
||||
|
||||
|
@ -307,6 +307,49 @@ namespace CPPCOMMON
|
||||
return res;
|
||||
}
|
||||
|
||||
//iconv
|
||||
int code_convert(const char *from_charset,const char *to_charset,char *inbuf,size_t inlen,char *outbuf,size_t outlen)
|
||||
{
|
||||
iconv_t cd;
|
||||
|
||||
char **pin = &inbuf;
|
||||
char **pout = &outbuf;
|
||||
|
||||
cd = iconv_open(to_charset,from_charset);
|
||||
if (cd==NULL)
|
||||
{
|
||||
//cout<<__FILE__<<__LINE__<<endl;
|
||||
return -1;
|
||||
}
|
||||
memset(outbuf,0,outlen);
|
||||
size_t ret = iconv(cd,pin,&inlen,pout,&outlen);
|
||||
if (ret == -1)
|
||||
{
|
||||
//cout<<__FILE__<<__LINE__<<endl;
|
||||
return -1;
|
||||
}
|
||||
iconv_close(cd);
|
||||
return 0;
|
||||
}
|
||||
|
||||
//gbk -> utf8
|
||||
string gbkToUtf8(const string& gbk)
|
||||
{
|
||||
//cout<<__FILE__<<__LINE__<<gbk<<endl;
|
||||
string res;
|
||||
size_t maxLen = gbk.size()*4;
|
||||
char * pUtf = new char[maxLen];
|
||||
int ret = code_convert("gbk", "utf-8", (char *)gbk.c_str(), gbk.size(), pUtf, maxLen);
|
||||
if(ret == -1)
|
||||
{
|
||||
delete [] pUtf;
|
||||
return "";
|
||||
}
|
||||
res = pUtf;
|
||||
delete [] pUtf;
|
||||
return res;
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
#ifdef TEST_STR_FUNCTS
|
||||
@ -351,22 +394,30 @@ int main()
|
||||
// //cout<<strlen(utf8str);
|
||||
// cout<<utf8str<<endl;
|
||||
//}
|
||||
cout<<string_format("hehe%s11asd%dasf","[here]",2);
|
||||
ifstream ifile("testdata/dict.txt");
|
||||
//cout<<string_format("hehe%s11asd%dasf","[here]",2);
|
||||
//ifstream ifile("testdata/dict.utf8");
|
||||
//string line;
|
||||
//while(getline(ifile, line))
|
||||
//{
|
||||
// cout<<line<<endl;
|
||||
// string uniStr = utf8ToUnicode(line);
|
||||
// //cout<<uniStr<<endl;
|
||||
// string utfStr = unicodeToUtf8(uniStr);
|
||||
// cout<<utfStr<<endl;
|
||||
//}
|
||||
//vector<string> tmp;
|
||||
//tmp.push_back("1");
|
||||
////tmp.push_back("2");
|
||||
////tmp.clear();
|
||||
//cout<<joinStr(tmp, ",")<<endl;
|
||||
ifstream ifile("testdata/dict.gbk");
|
||||
string line;
|
||||
while(getline(ifile, line))
|
||||
{
|
||||
cout<<line<<endl;
|
||||
string uniStr = utf8ToUnicode(line);
|
||||
//cout<<uniStr<<endl;
|
||||
string utfStr = unicodeToUtf8(uniStr);
|
||||
cout<<utfStr<<endl;
|
||||
//cout<<line<<endl;
|
||||
string s = gbkToUtf8(line);
|
||||
cout<<s<<endl;
|
||||
}
|
||||
vector<string> tmp;
|
||||
tmp.push_back("1");
|
||||
//tmp.push_back("2");
|
||||
//tmp.clear();
|
||||
cout<<joinStr(tmp, ",")<<endl;
|
||||
return 0;
|
||||
}
|
||||
#endif
|
||||
|
@ -10,6 +10,8 @@
|
||||
#include <stdint.h>
|
||||
#include <stdio.h>
|
||||
#include <stdarg.h>
|
||||
#include <iconv.h>
|
||||
#include <memory.h>
|
||||
#include "typedefs.h"
|
||||
namespace CPPCOMMON
|
||||
{
|
||||
@ -35,6 +37,8 @@ namespace CPPCOMMON
|
||||
string unicodeToUtf8(const string& uniStr);
|
||||
int utf8ToUnicode(const char* inutf8, int len, uint16_t* unicode);
|
||||
string utf8ToUnicode(const string& utfStr);
|
||||
int code_convert(const char *from_charset,const char *to_charset,char *inbuf,size_t inlen,char *outbuf,size_t outlen);
|
||||
string gbkToUtf8(const string& gbk);
|
||||
|
||||
inline uint16_t twocharToUint16(char high, char low)
|
||||
{
|
||||
|
Loading…
x
Reference in New Issue
Block a user