BOM-sensitive payloads

examples/bom.cpp shows both sides of BOM handling: observing a stripped BOM while decoding and requesting a BOM while encoding.

Use this when your integration touches text formats where a leading BOM may appear or be required: CSV exports, XML files, office-system imports, UTF-16 files, or partner feeds that use BOMs for encoding detection.

#include <iostream>

#include <polycpp/iconv_lite/iconv_lite.hpp>

int main() {
    namespace iconv = polycpp::iconv_lite;

    auto bytes = iconv::Buffer::concat({
        iconv::Buffer::from({0xEF, 0xBB, 0xBF}),
        iconv::Buffer::from("hello"),
    });

    bool stripped = false;
    iconv::DecodeOptions decodeOptions;
    decodeOptions.onBOMStripped = [&] {
        stripped = true;
    };

    std::cout << iconv::decode(bytes, "utf8", decodeOptions) << "\n";
    std::cout << (stripped ? "BOM stripped" : "BOM kept") << "\n";

    iconv::EncodeOptions encodeOptions;
    encodeOptions.addBOM = true;
    std::cout << iconv::encode("hello", "utf8", encodeOptions).toString("hex") << "\n";
    return 0;
}

Expected output:

hello
BOM stripped
efbbbf68656c6c6f

The callback is an observer: it runs only when a leading BOM is actually removed. Set DecodeOptions::stripBOM to false if the U+FEFF character must remain part of the decoded text.