Calculate width with ICU for cloud character (emoji variation selector)

main.cpp

#include <unicode/ubrk.h>
#include <unicode/brkiter.h>
#include <unicode/uchar.h>
#include <unicode/ustring.h>
#include <iostream>
#include <vector>

// Calculating the width of ☁️ using International Components for Unicode (ICU) in C++ with grapheme clusters.
// Compiles in C++11 or above, with ICU installed.
int main() {
    UErrorCode status = U_ZERO_ERROR;
    // Combine cloud character and variation selector into a single string.
    UChar text[] = {0x2601, 0xFE0F, 0};  // Null-terminated UTF-16 string for ☁️

    // Create a grapheme cluster boundary iterator.
    std::unique_ptr<icu::BreakIterator> iter(icu::BreakIterator::createCharacterInstance(icu::Locale::getDefault(), status));
    if (U_FAILURE(status)) {
        std::cerr << "Failed to create BreakIterator: " << u_errorName(status) << std::endl;
        return 1;
    }

    icu::UnicodeString str(text);
    iter->setText(str);

    // Iterate over grapheme clusters and count them.
    int32_t count = 0;
    int prev = iter->first();
    for (int32_t boundary = iter->next(); boundary != icu::BreakIterator::DONE; boundary = iter->next()) {
        // Returns 0-2 for ☁️, confirming the width is 2.
        std::cout << "Grapheme cluster from " << prev << " to " << boundary << std::endl;
        count++;
        prev = boundary;
    }
    // Returns 1 for cloud character (number of clusters).
    std::cout << "Number of grapheme clusters: " << count << std::endl;

    return 0;
}