我正在创建一个缓冲从std::cin
读取的二进制数据的类;我需要一些限制
能够通过流回溯。
更详细地说,当我从流中获得read
n
个字节时,我希望能够回放到
流中的n
个字节。
目前,我通过附加单个字符附加到缓冲区
到std::deque
,但我希望有更好的方法来复制
从std::istream
到std::deque
的数据块。
基本上有一种方法可以改进(删除?)while循环
下面的load
功能?
目前我有:
#include<algorithm>
#include<deque>
class BufferedStdIn{
public:
BufferedStdIn(): m_buffer(), m_cursor(0){
}
// copy data out from my internal buffer to the client's array
// format dictated by external interface requirement
void read(char* dest, size_t n){
load(n)
std::copy( m_buffer.begin(), m_buffer.begin()+n, dest );
m_cursor+=n;
}
// Go backwards in the stream; return the number of bytes moved.
// @retval < input arg n indicates that we can't rewind as far as asked for
size_t rewind(size_t n){
int final_pos=m_cursor-n;
if( final_pos < 0 ){
// can't go as far as the client wants, just go back to start of my buffer
size_t moved=m_cursor;
m_cursor=0;
return moved;
}
else{ // can do what client wants
m_cursor=final_pos;
return n;
}
}
void load(n){
m_buffer.erase( m_buffer.begin(), m_buffer.begin()+m_cursor);
m_cursor=0;
while( m_buffer.size()<n){
char c;
std::cin.get(c);
m_buffer.push_back(c);
}
}
// SNIP: other functions that let clients look at m_buffer
std::deque<char> m_buffer;
size_t m_cursor;
}
答案 0 :(得分:1)
可能效率不高,但更短:
void load(size_t n)
{
m_buffer.erase( m_buffer.begin(), m_buffer.begin()+m_cursor);
m_cursor=0;
if(m_buffer.size() < n)
std::copy_n(std::istream_iterator<char>(std::cin), n - m_buffer.size(), std::back_inserter(m_buffer));
}