uni_uax29_table_gen.php

<?php
declare(strict_types=1);

// Print data structures necessary for finding grapheme cluster breaks
// https://unicode.org/reports/tr29/

function print_prop_table_for_url($prefix, $url, $filter, $prop_strip) {
    $lines = explode("\n", trim(file_get_contents($url)));
    $ranges = [];
    foreach ($lines as $line) {
        $m = [];
        if (!preg_match('/^([0-9A-F]+)(?:\.\.([0-9A-F]+))?\s+;\s+([^#]+)/', $line, $m)) continue;
        if ($filter) if (!preg_match($filter, $line)) continue;
        $prop = $m[3];
        if ($prop_strip) $prop = preg_replace($prop_strip, ' ', $prop);
        $prop = preg_replace('/\W+/', ' ', $prop);
        $prop = trim(sprintf('UNI_PROP_%s %s', $prefix, $prop));
        $prop = preg_replace('/ +/', '_', strtoupper($prop));
        $ranges[] = [ hexdec($m[1]), hexdec($m[2] ?: $m[1]), $prop ];
    }

    $props = array_unique(array_map(fn ($e) => $e[2], $ranges));
    sort($props);
    echo "enum uni_prop_{$prefix} {\n";
    foreach ($props as $prop) {
        echo "    $prop,\n";
    }
    echo "};\n\n";

    printf("#define UNI_TABLE_%s_COUNT %d\n", strtoupper($prefix), count($ranges));
    usort($ranges, fn ($a, $b) => $a[0] <=> $b[0]);
    echo "static struct uni_table uni_table_{$prefix}[] = {\n";

    foreach ($ranges as $r) {
        printf("    { 0x%05x, 0x%05x, %s },\n", $r[0], $r[1], $r[2]);
    }
    echo "};\n\n";
}

print_prop_table_for_url(
    'grapheme_break',
    'http://www.unicode.org/Public/UCD/latest/ucd/auxiliary/GraphemeBreakProperty.txt',
    null,
    null
);

print_prop_table_for_url(
    'incb',
    'http://www.unicode.org/Public/UCD/latest/ucd/DerivedCoreProperties.txt',
    '/\bInCB\b/',
    '/InCB/'
);

print_prop_table_for_url(
    'ext_picto',
    'http://www.unicode.org/Public/UCD/latest/ucd/emoji/emoji-data.txt',
    '/\bExtended_Pictographic\b/',
    '/Extended_Pictographic/'
);

/*
struct uni_table {
    uint32_t range_start;
    uint32_t range_end;
    int prop;
};

int uni_prop_lookup(struct uni_table *table, int ntable, uint32_t c) {
    int lo = 0;
    int hi = ntable - 1;
    while (lo <= hi) {
        int i = (lo + hi) / 2;
        if (c < table[i].range_start) {
            hi = i - 1;
        } else if (c > table[i].range_end) {
            lo = i + 1;
        } else {
            return table[i].prop;
        }
    }
    return -1;
}
*/