Function validate_path_component

src/os/linux/pagecache.cpp:126–155

}

PathTrust validate_path_component(std::string_view comp) {
    if (comp.empty()) return {false, "empty path component"};
    if (comp == "." || comp == "..") return {false, "path traversal component"};
    const auto* p = reinterpret_cast<const unsigned char*>(comp.data());
    const std::size_t n = comp.size();
    std::size_t i = 0;
    while (i < n) {
        const unsigned char c = p[i];
        if (c < 0x80) {
            // ASCII byte: apply the display-/host-safety rules.
            if (c == 0) return {false, "embedded NUL in path component"};
            if (c == '/') return {false, "slash in path component"};
            if (c < 0x20 || c == 0x7f) return {false, "control byte in path component"};
            if (is_host_forbidden_path_char(c))
                return {false, "host-forbidden character in path component"};
            ++i;
        } else {
            // Non-ASCII: accept only well-formed UTF-8; reject binary garbage.
            const int len = utf8_sequence_len(p + i, n - i);
            if (len == 0) return {false, "invalid UTF-8 in path component"};
            // Reject C1 control codepoints U+0080–U+009F (0xC2 0x80–0x9F).
            if (len == 2 && c == 0xC2 && p[i + 1] <= 0x9F)
                return {false, "control codepoint in path component"};
            i += static_cast<std::size_t>(len);
        }
    }
    if (comp.back() == ' ' || comp.back() == '.')
        return {false, "component has trailing space or dot"};
    return {true, {}};
}

std::string escape_path_for_report(const std::string& s) {
    std::string out;