Can/How do you serialize an array using the cereal library.
可以/如何使用谷物库序列化数组。
I.e.
即。
void save(Archive & ar, const unsigned int version) const
{
unsigned int l = g1_size_bin(g,POINT_COMPRESS);
uint8_t data[l];
memset(data, 0, l);
g1_write_bin(data, l, g,POINT_COMPRESS);
ar(l);
ar(data); // what should be here
}
That doesn't work (nor would I expect it too). Nor does
这是行不通的(我也不希望如此)。也没有
ar(cereal::binary_data(data,l));
(which I would think would work, since it looks like the boost code one would use), which produces a compilation error :
(我认为这是可行的，因为它看起来像一个人会使用的boost代码)，这会产生一个编译错误:
/usr/local/include/cereal/cereal.hpp:79:17: note: candidate template ignored: substitution failure : variably modified type 'unsigned char (&)[l]' cannot be used as a template argument BinaryData binary_data( T && data, size_t size )
/usr/local/include/cereal/cereal.hpp:79:17:注意:候选模板被忽略:替换失败:变量修改的类型“unsigned char (&)[l]”不能用作模板参数BinaryData binary_data binary_data(T & data, size_t size)
Nor does
也没有
ar.saveBinaryValue(data,l);
Since that method only appears to be supported for XML/Json and I want a binary archive.
由于该方法似乎只支持XML/Json，所以我需要二进制归档。
1 个解决方案
#1
6
cereal::binary_data is the correct construct to use in this case assuming you want a binary representation of the POD array. This will only work for archives that support binary_data (binary and portable_binary). binary_data doesn't work for the text based archives because it is seen as an optimization for a more generic serialization method - see how vector is serialized for an example of this.
在本例中，如果您想要POD数组的二进制表示，那么binary_data是正确的结构。这将只适用于支持binary_data(二进制和portable_binary)的存档。binary_data不能用于基于文本的归档文件，因为它被视为对更通用的序列化方法的优化——请参见这个示例中的vector是如何序列化的。
Anyway, here's a working example of serializing a C style array:
总之，这里有一个序列化C样式数组的工作示例:
#include <cereal/archives/binary.hpp>
#include <iostream>
int main()
{
std::stringstream ss;
{
cereal::BinaryOutputArchive ar(ss);
std::uint8_t data[] = {1, 2, 3};
ar( cereal::binary_data( data, sizeof(std::uint8_t) * 3 ) );
}
{
cereal::BinaryInputArchive ar(ss);
std::uint8_t data[3];
ar( cereal::binary_data( data, sizeof(std::uint8_t) * 3 ) );
for( int i : data )
std::cout << i << " ";
}
return 0;
}
If you wanted to serialize a C style array to a text based archive, or if your array wasn't over POD types, you would need to iterate over each object and serialize it individually.
如果您希望将C样式数组序列化为基于文本的归档文件，或者如果您的数组没有超过POD类型，那么您将需要对每个对象进行迭代，并分别对其进行序列化。
#1
6
cereal::binary_data is the correct construct to use in this case assuming you want a binary representation of the POD array. This will only work for archives that support binary_data (binary and portable_binary). binary_data doesn't work for the text based archives because it is seen as an optimization for a more generic serialization method - see how vector is serialized for an example of this.
在本例中，如果您想要POD数组的二进制表示，那么binary_data是正确的结构。这将只适用于支持binary_data(二进制和portable_binary)的存档。binary_data不能用于基于文本的归档文件，因为它被视为对更通用的序列化方法的优化——请参见这个示例中的vector是如何序列化的。
Anyway, here's a working example of serializing a C style array:
总之，这里有一个序列化C样式数组的工作示例:
#include <cereal/archives/binary.hpp>
#include <iostream>
int main()
{
std::stringstream ss;
{
cereal::BinaryOutputArchive ar(ss);
std::uint8_t data[] = {1, 2, 3};
ar( cereal::binary_data( data, sizeof(std::uint8_t) * 3 ) );
}
{
cereal::BinaryInputArchive ar(ss);
std::uint8_t data[3];
ar( cereal::binary_data( data, sizeof(std::uint8_t) * 3 ) );
for( int i : data )
std::cout << i << " ";
}
return 0;
}
If you wanted to serialize a C style array to a text based archive, or if your array wasn't over POD types, you would need to iterate over each object and serialize it individually.
如果您希望将C样式数组序列化为基于文本的归档文件，或者如果您的数组没有超过POD类型，那么您将需要对每个对象进行迭代，并分别对其进行序列化。