a) 我需要使用filtering_streambuf 的filtering_stream 吗?我会
期望后者已经具有某种缓冲行为。
两者都可以。该流添加了文本和语言环境 功能,就像在标准库中一样。
b) 如何关闭过滤流(buf)并将其写入缓冲区?
您可以使用array_sink、back_inserter_device、内存映射等。请参阅https://www.boost.org/doc/libs/1_72_0/libs/iostreams/doc/(“模型”)。
c) 如何读取压缩数据的最终大小? .tellg() 是
没有为这些过滤流实现(如前所述
在其他地方)
从您的底层输出设备/流中检测它。在你做之前不要忘记刷新/关闭过滤层。
d) 你可以有多个来源,即我的三个缓冲区还是我需要
把它们结合起来? (见下文我的方法)。
你可以为所欲为。
给我看代码...
我会反转主动,并在写入时将过滤器压缩到输出缓冲区:
using RawBuffer = std::vector<char>;
using Device = bio::back_insert_device<RawBuffer>;
RawBuffer compressed_buffer; // optionally reserve some size
{
bio::filtering_ostream filter;
filter.push(bio::gzip_compressor());
filter.push(Device{ compressed_buffer });
filter.write(reinterpret_cast<char const*>(&frame.buf1),
sizeof(frame) - offsetof(Frame, buf1));
}
改为使用过滤流缓冲区:
{
bio::filtering_ostreambuf filter;
filter.push(bio::gzip_compressor());
filter.push(Device{ compressed_buffer });
std::copy_n(reinterpret_cast<char const*>(&frame.buf1),
sizeof(frame) - offsetof(Frame, buf1),
std::ostreambuf_iterator<char>(&filter));
}
现在你的问题的答案很突出:
const char *compressed = compressed_buffer.data();
int compressedSize = compressed_buffer.size();
我会将剩余的代码减少到:
{
std::ofstream ofs("ouput.data", std::ios::binary | std::ios::app);
ofs << "START";
ofs << "END " << compressed_buffer.size();
ofs.write(compressed_buffer.data(), compressed_buffer.size());
}
考虑不为每一帧重新打开输出流:)
现场演示
Live On Coliru
#include <boost/iostreams/filtering_streambuf.hpp>
#include <boost/iostreams/filter/gzip.hpp>
#include <boost/iostreams/device/back_inserter.hpp>
#include <iterator>
#include <fstream>
#include <vector>
namespace bio = boost::iostreams;
class Frame {
private:
/* other things */
public:
float buf1[3];
float buf2[3];
float buf3[4];
/* more things */
};
int main() {
Frame const frames[]{
{
{ 1, 2, 3 },
{ 4, 5, 6 },
{ 7, 8, 9, 10 },
},
{
{ 11, 12, 13 },
{ 14, 15, 16 },
{ 17, 18, 19, 20 },
},
{
{ 21, 22, 23 },
{ 24, 25, 26 },
{ 27, 28, 29, 30 },
},
};
// avoiding UB:
static_assert(std::is_trivial_v<Frame> &&
std::is_standard_layout_v<Frame>);
using RawBuffer = std::vector<char>;
using Device = bio::back_insert_device<RawBuffer>;
std::remove("output.data");
std::ofstream ofs("output.data", std::ios::binary | std::ios::app);
RawBuffer compressed_buffer; // optionally reserve some size
for (Frame const& frame : frames) {
compressed_buffer.clear(); // do not shrink_to_fit optimizing allocation
{
bio::filtering_ostreambuf filter;
filter.push(bio::gzip_compressor());
filter.push(Device{ compressed_buffer });
std::copy_n(reinterpret_cast<char const*>(&frame.buf1),
sizeof(frame) - offsetof(Frame, buf1),
std::ostreambuf_iterator<char>(&filter));
}
ofs << "START";
ofs << "END " << compressed_buffer.size();
ofs.write(compressed_buffer.data(), compressed_buffer.size());
}
}
确定性地生成 output.data:
00000000: 5354 4152 5445 4e44 2035 301f 8b08 0000 STARTEND 50.....
00000010: 0000 0000 ff63 6068 b067 6060 7000 2220 .....c`h.g``p."
00000020: 6e00 e205 407c 0088 1f00 3183 2303 8300 n...@|....1.#...
00000030: 102b 3802 0058 a049 af28 0000 0053 5441 .+8..X.I.(...STA
00000040: 5254 454e 4420 3438 1f8b 0800 0000 0000 RTEND 48........
00000050: 00ff 6360 3070 6460 7000 e200 204e 00e2 ..c`0pd`p... N..
00000060: 0220 6e00 e20e 209e 00c4 3380 7881 2300 . n... ...3.x.#.
00000070: 763b 7371 2800 0000 5354 4152 5445 4e44 v;sq(...STARTEND
00000080: 2034 391f 8b08 0000 0000 0000 ff63 6058 49..........c`X
00000090: e1c8 c0b0 0188 7700 f101 203e 01c4 1780 ......w... >....
000000a0: f806 103f 00e2 1740 fcc1 1100 dfb4 6cde ...?...@......l.
000000b0: 2800 0000 (...