我有一个指向const *char
缓冲区的指针及其长度,我正在尝试使用接受类型对象的API(在本例中为AWS S3 C++ upload request) :
std::basic_iostream <char, std::char_traits <char>>
是否有一种简单的标准C ++ 11方法将我的缓冲区转换为兼容的流,最好不要实际复制内存?
答案 0 :(得分:3)
感谢伊戈尔的评论,这似乎有效:
func(const * char buffer, std::size_t buffersize)
{
auto sstream = std::make_shared<std::stringstream>();
sstream->write(buffer, buffersize);
...
uploadRequest.SetBody(sstream);
....
答案 1 :(得分:2)
作为解决方案的一个相当明显的推论,您可以使用这样的代码创建一个空的basic_iostream。此示例创建一个0字节的伪目录S3密钥:
Aws::S3::Model::PutObjectRequest object_request;
// dirName ends in /.
object_request.WithBucket(bucketName).WithKey(dirName);
// Create an empty input stream to create the 0-byte directory file.
auto empty_sstream = std::make_shared<std::stringstream>();
object_request.SetBody(empty_sstream);
auto put_object_outcome = s3_client->PutObject(object_request);
答案 2 :(得分:1)
如果您不想要复制数据,并假设使用boost是一个选项,则可以使用boost中的basic_bufferstream:
#include <boost/interprocess/streams/bufferstream.hpp>
char* buf = nullptr; // get your buffer
size_t length = 0;
auto input = Aws::MakeShared<boost::interprocess::basic_bufferstream<char>> (
"PutObjectInputStream",
buf,
length);
然后您的s3客户端可以使用它:
Aws::S3::Model::PutObjectRequest req;
req.WithBucket(bucket).WithKey(key);
req.SetBody(input);
s3.PutObject(req);