C / C ++-如何从网上按字节下载字节（如文件流）？

Question

I'm building a simple download and file transfer manager in C and C++, with support for pause and resume operations. I looked for a way to download byte by byte from a given URL, but all I could find were references to external libraries like libcurl. I don't want external libraries, I want to manually read and write the bytes in a similar way used to read and write bytes to files (like ifstream in C++, and fopen in C) using only the C/C++ native library. The pseudo-code for what I need would be:

void download(file, url) {
    file.open();
    url.open();
    set_active(true);
    while( ! pause() ) {
        bytes_read = url.read(byte_array);
        if(bytes_read < 1)
            break;
        file.write(byte_array, bytes_read);
    }
    url.close();
    file.close();
    set_active(false);
}

How would that pseudo-code work in both C and C++?

---

EDIT: I don't want external libraries for the simple interest on learning the capabilities and limits of both languages and their native libraries.

EDIT 2: Conclusion: Neither of the languages have native support for HTTP operations, therefore I have to choose between a using third-party library or writing a HTTP (and several other protocols) implementation from the scratch, which is unreliable. Thanks.

Answer 1

You should use libcurl or some other HTTP client library (eg libsoup if using GTK). If you need HTTP server support, you can find libraries for that, eg libonion .

You might also consider POCO or Qt5 frameworks, they both have HTTP things.

If you don't want to use any such HTTP 3 ^rd party libraries, you'll need to re-implement HTTP by yourself. That might be complex (and several months or years of work), since the HTTP specifications are several hundreds pages! Handling some simplistic GET requests is easy, but handing all the HTTP specification is hard.

^{You could try to implement your subset of HTTP which might work in some cases, but you'll easily find HTTP servers or clients needing more of HTTP, .... And your implementation would be much more brittle and buggy than existing HTTP libraries...}

Answer 2

Just noticed your edited requirement that you don't want external libraries. But here are extracts from a program I wrote in MSVC to download files from a Windows program.

// Windows Header Files:
#include <windows.h>
#include <wininet.h>

// Make internet connection.
HINTERNET hInternetSession = InternetOpen(
        "Microsoft Internet Explorer",          // agent
        INTERNET_OPEN_TYPE_PRECONFIG,           // access
        NULL, NULL, 0);                         // defaults
...

// Make connection to desired page.
HINTERNET hURL = InternetOpenUrl(
        hInternetSession,                       // session handle
        szSors,                                 // URL to access
        NULL, 0, 0, 0);                         // defaults
if (hURL == 0)
    return 0;
...

// Read data into memory buffer.
DWORD dwBytesRead = 0;
BOOL bResult = InternetReadFile(
        hURL,                                   // handle to URL
        (LPSTR)cBuffer,                         // pointer to buffer
        (DWORD)MYBSIZE,                         // size of buffer
        &dwBytesRead);                          // to hold return value
...

InternetCloseHandle(hInternetSession);

I never downloaded one byte as you ask, but I don't see why you could not.