#include "catch.hpp"

#include <osmium/io/detail/string_util.hpp>

#include <iterator>
#include <locale>
#include <stdexcept>
#include <string>

// From C++20 we need to handle unicode literals differently
#ifdef __cpp_char8_t
static const char* u8cast(const char8_t *s) noexcept {
    return reinterpret_cast<const char*>(s);
}
#else
static const char* u8cast(const char *s) noexcept {
    return s;
}
#endif

TEST_CASE("output formatted with small results") {
    std::string out;
    osmium::io::detail::append_printf_formatted_string(out, "%d", 17);
    REQUIRE(out == "17");
}

TEST_CASE("output formatted with several parameters in format string") {
    std::string out;
    osmium::io::detail::append_printf_formatted_string(out, "%d %s", 17, "foo");
    REQUIRE(out == "17 foo");
}

TEST_CASE("output formatted into string already containing something") {
    std::string out{"foo"};
    osmium::io::detail::append_printf_formatted_string(out, " %d", 23);
    REQUIRE(out == "foo 23");
}

TEST_CASE("output formatted with large results") {
    const char* str =
        "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
        "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
        "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
        "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX"
        "XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX";

    std::string out;
#pragma GCC diagnostic push
#if __GNUC__ >= 8
#pragma GCC diagnostic ignored "-Wformat-truncation"
#endif
    osmium::io::detail::append_printf_formatted_string(out, "%s", str);
#pragma GCC diagnostic pop
    REQUIRE(out == str);
}

TEST_CASE("UTF8 encoding: append to string") {
    std::string out{"1234"};
    osmium::io::detail::append_utf8_encoded_string(out, "abc");
    REQUIRE(out == "1234abc");
}

TEST_CASE("UTF8 encoding: don't encode alphabetic characters") {
    const char* s = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ";
    std::string out;
    osmium::io::detail::append_utf8_encoded_string(out, s);
    REQUIRE(out == s);
}

TEST_CASE("UTF8 encoding: don't encode numeric characters") {
    const char* s = "0123456789";
    std::string out;
    osmium::io::detail::append_utf8_encoded_string(out, s);
    REQUIRE(out == s);
}

TEST_CASE("UTF8 encoding: don't encode lots of often used characters characters") {
    const char* s = ".-;:_#+";
    std::string out;
    osmium::io::detail::append_utf8_encoded_string(out, s);
    REQUIRE(out == s);
}

TEST_CASE("UTF8 encoding: encode characters that are special in OPL") {
    std::string out;
    osmium::io::detail::append_utf8_encoded_string(out, " \n,=@");
    REQUIRE(out == "%20%%0a%%2c%%3d%%40%");
}

TEST_CASE("UTF8 encoding: encode multibyte character") {
    std::string out;
    osmium::io::detail::append_utf8_encoded_string(out, u8cast(u8"\u30dc_\U0001d11e_\U0001f6eb"));
    REQUIRE(out == "%30dc%_%1d11e%_%1f6eb%");
}

TEST_CASE("html encoding does not encode normal characters") {
    const char* s = "abc123,.-";
    std::string out;
    osmium::io::detail::append_xml_encoded_string(out, s);
    REQUIRE(out == s);
}

TEST_CASE("html encoding encodes special XML characters") {
    const char* s = "& \" \' < > \n \r \t";
    std::string out;
    osmium::io::detail::append_xml_encoded_string(out, s);
    REQUIRE(out == "&amp; &quot; &apos; &lt; &gt; &#xA; &#xD; &#x9;");
}

TEST_CASE("debug encoding does not encode normal characters") {
    const char* s = "abc123,.-";
    std::string out;
    osmium::io::detail::append_debug_encoded_string(out, s, "[", "]");
    REQUIRE(out == s);
}

TEST_CASE("debug encoding encodes some unicode characters") {
    const char* s = u8cast(u8"\n_\u30dc_\U0001d11e_\U0001f6eb");
    std::string out;
    osmium::io::detail::append_debug_encoded_string(out, s, "[", "]");
    REQUIRE(out == "[<U+000A>]_[<U+30DC>]_[<U+1D11E>]_[<U+1F6EB>]");
}

TEST_CASE("utf8 encoding of non-printable characters in the first 127 characters") {
    const std::locale cloc{"C"};
    char s[] = "a\0";

    for (char c = 1; c < 0x7f; ++c) {
        std::string out;
        s[0] = c;
        osmium::io::detail::append_utf8_encoded_string(out, s);
        if (!std::isprint(c, cloc)) {
            REQUIRE(out[0] == '%');
        }
    }
}

TEST_CASE("debug encoding of non-printable characters in the first 127 characters") {
    const std::locale cloc{"C"};
    char s[] = "a\0";

    for (char c = 1; c < 0x7f; ++c) {
        std::string out;
        s[0] = c;
        osmium::io::detail::append_debug_encoded_string(out, s, "", "");
        if (!std::isprint(c, cloc)) {
            REQUIRE(out[0] == '<');
        }
    }
}

TEST_CASE("test codepoint to utf8 encoding") {
    const char* s = u8cast(u8"\n_\u01a2_\u30dc_\U0001d11e_\U0001f680");

    std::string out;
    osmium::io::detail::append_codepoint_as_utf8(0x0aU, std::back_inserter(out)); // 1 utf8 byte
    REQUIRE(out.size() == 1);
    osmium::io::detail::append_codepoint_as_utf8('_', std::back_inserter(out));
    REQUIRE(out.size() == 2);
    osmium::io::detail::append_codepoint_as_utf8(0x01a2U, std::back_inserter(out)); // 2 utf8 bytes
    REQUIRE(out.size() == 4);
    osmium::io::detail::append_codepoint_as_utf8('_', std::back_inserter(out));
    REQUIRE(out.size() == 5);
    osmium::io::detail::append_codepoint_as_utf8(0x30dcU, std::back_inserter(out)); // 3 utf8 bytes
    REQUIRE(out.size() == 8);
    osmium::io::detail::append_codepoint_as_utf8('_', std::back_inserter(out));
    REQUIRE(out.size() == 9);
    osmium::io::detail::append_codepoint_as_utf8(0x1d11eU, std::back_inserter(out)); // 4 utf8 bytes
    REQUIRE(out.size() == 13);
    osmium::io::detail::append_codepoint_as_utf8('_', std::back_inserter(out));
    REQUIRE(out.size() == 14);
    osmium::io::detail::append_codepoint_as_utf8(0x1f680U, std::back_inserter(out)); // 4 utf8 bytes
    REQUIRE(out.size() == 18);
    REQUIRE(out == s);
}

TEST_CASE("test utf8 to codepoint decoding") {
    const char* s = u8cast(u8"\n_\u01a2_\u30dc_\U0001d11e_\U0001f680");

    const char* it = s;
    const char* end = s + std::strlen(s) + 1;
    REQUIRE(osmium::io::detail::next_utf8_codepoint(&it, end) == 0x0aU);
    REQUIRE(osmium::io::detail::next_utf8_codepoint(&it, end) == '_');
    REQUIRE(osmium::io::detail::next_utf8_codepoint(&it, end) == 0x01a2U);
    REQUIRE(osmium::io::detail::next_utf8_codepoint(&it, end) == '_');
    REQUIRE(osmium::io::detail::next_utf8_codepoint(&it, end) == 0x30dcU);
    REQUIRE(osmium::io::detail::next_utf8_codepoint(&it, end) == '_');
    REQUIRE(osmium::io::detail::next_utf8_codepoint(&it, end) == 0x1d11eU);
    REQUIRE(osmium::io::detail::next_utf8_codepoint(&it, end) == '_');
    REQUIRE(osmium::io::detail::next_utf8_codepoint(&it, end) == 0x1f680U);

    REQUIRE(*it++ == '\0');
    REQUIRE(it == end);
}

TEST_CASE("Roundtrip unicode characters") {
    char s[4] = {0};

    const uint32_t max_code_point = 0x10ffffU;
    for (uint32_t cp = 0; cp <= max_code_point; ++cp) {
        const auto* const end = osmium::io::detail::append_codepoint_as_utf8(cp, s);
        const char* it = s;
        REQUIRE(osmium::io::detail::next_utf8_codepoint(&it, std::end(s)) == cp);
        REQUIRE(end == it);
    }
}

TEST_CASE("invalid codepoint") {
    const char s[] = {static_cast<char>(0xff), static_cast<char>(0xff), static_cast<char>(0xff), static_cast<char>(0xff)};
    const char* it = s;
    REQUIRE_THROWS_AS(osmium::io::detail::next_utf8_codepoint(&it, std::end(s)), std::runtime_error);
}

TEST_CASE("incomplete Unicode codepoint") {
    const char* s = u8cast(u8"\U0001f680");

    const auto* it = s;
    const char* end = s + std::strlen(s) + 1;
    REQUIRE(osmium::io::detail::next_utf8_codepoint(&it, end - 1) == 0x1f680U);
    REQUIRE(std::distance(s, it) == 4);

    for (const int i : {0, 1, 2, 3}) {
        it = s;
        REQUIRE_THROWS_AS(osmium::io::detail::next_utf8_codepoint(&it, s + i), std::out_of_range);
    }
}