Ubuntu下十进制转Unicode转换器文件输出失败问题求助

Hey there! Let's figure out how to get your Unicode converter writing to files properly—console output working but files failing is such a common gotcha with Unicode handling, so I’ve got you covered.

First, let's break down the core issue #

Your extended fstream works for the console because it’s handling multi-byte Unicode (like the 3-byte UTF-8 sequence for €) correctly in that context. For files, though, we need to ensure we’re writing the right byte sequence or configuring the stream to handle Unicode encoding properly.

Option 1: Write raw UTF-8 bytes directly to file #

If your converter already turns decimal numbers into valid UTF-8 byte sequences (like 226 130 172 for €), this is the simplest approach. We’ll use a standard ofstream in binary mode to avoid corrupting the multi-byte data with automatic newline conversions.

#include <fstream>
#include <vector>

// Example: Your converter outputs this UTF-8 byte sequence for €
std::vector<unsigned char> utf8_bytes = {226, 130, 172};

// Open file in binary mode to preserve raw bytes
std::ofstream output_file("result.txt", std::ios::binary);
if (output_file.is_open()) {
    // Write the raw bytes to the file
    output_file.write(reinterpret_cast<const char*>(utf8_bytes.data()), utf8_bytes.size());
    output_file.close();
}

When you open result.txt with a text editor that supports UTF-8, you’ll see the correct € symbol.

Option 2: Use wide character streams (for your extended fstream) #

If your extended fstream relies on wide characters (wchar_t), you need to configure the stream to encode wide characters into UTF-8 before writing to the file. The default locale won’t do this automatically—we have to explicitly set a UTF-8-aware codecvt facet.

#include <fstream>
#include <locale>
#include <codecvt>
#include <string>

// Convert your wide character string to UTF-8
std::wstring_convert<std::codecvt_utf8<wchar_t>> utf8_converter;
std::wstring unicode_wstr = L"€"; // Example wide char from your converter
std::string utf8_str = utf8_converter.to_bytes(unicode_wstr);

// Write the UTF-8 string to file
std::ofstream output_file("result.txt");
if (output_file.is_open()) {
    output_file << utf8_str;
    output_file.close();
}

Alternatively, you can imbue the wide stream directly with a UTF-8 locale:

std::wofstream output_file("result.txt");
// Set locale to handle UTF-8 encoding for wide characters
output_file.imbue(std::locale(output_file.getloc(), new std::codecvt_utf8<wchar_t>));

if (output_file.is_open()) {
    output_file << L"€"; // Write your wide character directly
    output_file.close();
}

Note: codecvt_utf8 is marked deprecated in C17, but it’s still fully supported in GCC, Clang, and MSVC. For C20+, you could use <format> or newer Unicode utilities, but the above code will work for most use cases.

Key things to remember #

• Stick to one encoding: Always write files in a consistent encoding (UTF-8 is the safest choice) and make sure you open the file with that encoding in your text editor.
• Avoid text mode for raw bytes: When writing multi-byte sequences, use std::ios::binary to prevent the OS from altering newline characters, which can break Unicode sequences.
• Update your extended fstream: If you want to integrate this into your custom stream class, add a method that either writes raw UTF-8 bytes (like Option 1) or converts wide chars to UTF-8 before writing.

内容的提问来源于stack exchange，提问作者user9463791