From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: by sourceware.org (Postfix, from userid 2181) id EA12C387103D; Fri, 29 Sep 2023 15:01:41 +0000 (GMT) DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org EA12C387103D DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org; s=default; t=1695999701; bh=sMapC0aiqtG5svsgzSzuCXTUM038Y6TB035X5romDpU=; h=From:To:Subject:Date:From; b=nHhdcA88HA4SyxQnQRoa39mmT8UJDvsXE9YL41gvUJFqqMRfaX0BQQ35/Ow+HSXyM 6GsX4TfUavqOtU4bxgNGSPYSqs8h2i2t+PBJa8BLPK6HpR7Ba/SIMGewjH8b2Nfjub cq80b55Xfz0+xPOaM9DiwP/c/AiohiC/IPvR/5aM= MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset="utf-8" From: Jonathan Wakely To: gcc-cvs@gcc.gnu.org, libstdc++-cvs@gcc.gnu.org Subject: [gcc r14-4335] libstdc++: Fix handling of surrogate CP in codecvt [PR108976] X-Act-Checkin: gcc X-Git-Author: Dimitrij Mijoski X-Git-Refname: refs/heads/master X-Git-Oldrev: 28adad7a32ed92a3c0906b38411c4ed686bbacc5 X-Git-Newrev: a8b9c32da787ea0bfbfc9118ac816fa7be4b1bc8 Message-Id: <20230929150141.EA12C387103D@sourceware.org> Date: Fri, 29 Sep 2023 15:01:41 +0000 (GMT) List-Id: https://gcc.gnu.org/g:a8b9c32da787ea0bfbfc9118ac816fa7be4b1bc8 commit r14-4335-ga8b9c32da787ea0bfbfc9118ac816fa7be4b1bc8 Author: Dimitrij Mijoski Date: Thu Sep 28 21:38:11 2023 +0200 libstdc++: Fix handling of surrogate CP in codecvt [PR108976] This patch fixes the handling of surrogate code points in all standard facets for transcoding Unicode that are based on std::codecvt. Surrogate code points should always be treated as error. On the other hand surrogate code units can only appear in UTF-16 and only when they come in a proper pair. Additionally, it fixes a bug in std::codecvt_utf16::in() when odd number of bytes were given in the range [from, from_end), error was returned always. The last byte in such range does not form a full UTF-16 code unit and we can not make any decisions for error, instead partial should be returned. The testsuite for testing these facets was updated in the following order: 1. All functions that test codecvts that work with UTF-8 were refactored and made more generic so they accept codecvt that works with the char type char8_t. 2. The same functions were updated with new test cases for transcoding errors and now additionally test for surrogates, overlong UTF-8 sequences, code points out of the Unicode range, and more tests for missing leading and trailing code units. 3. New tests were added to test codecvt_utf16 in both of its variants, UTF-16 <-> UTF-32/UCS-4 and UTF-16 <-> UCS-2. libstdc++-v3/ChangeLog: PR libstdc++/108976 * src/c++11/codecvt.cc (read_utf8_code_point): Fix handing of surrogates in UTF-8. (ucs4_out): Fix handling of surrogates in UCS-4 -> UTF-8. (ucs4_in): Fix handling of range with odd number of bytes. (ucs4_out): Fix handling of surrogates in UCS-4 -> UTF-16. (ucs2_out): Fix handling of surrogates in UCS-2 -> UTF-16. (ucs2_in): Fix handling of range with odd number of bytes. (__codecvt_utf16_base::do_in): Likewise. (__codecvt_utf16_base::do_in): Likewise. (__codecvt_utf16_base::do_in): Likewise. * testsuite/22_locale/codecvt/codecvt_unicode.cc: Renames, add tests for codecvt_utf16 and codecvt_utf16. * testsuite/22_locale/codecvt/codecvt_unicode.h: Refactor UTF-8 testing functions for char8_t, add more test cases for errors, add testing functions for codecvt_utf16. * testsuite/22_locale/codecvt/codecvt_unicode_wchar_t.cc: Renames, add tests for codecvt_utf16. * testsuite/22_locale/codecvt/codecvt_utf16/79980.cc (test06): Fix test. * testsuite/22_locale/codecvt/codecvt_unicode_char8_t.cc: New test. Diff: --- libstdc++-v3/src/c++11/codecvt.cc | 18 +- .../testsuite/22_locale/codecvt/codecvt_unicode.cc | 38 +- .../testsuite/22_locale/codecvt/codecvt_unicode.h | 1799 +++++++++++++++----- .../22_locale/codecvt/codecvt_unicode_char8_t.cc | 53 + .../22_locale/codecvt/codecvt_unicode_wchar_t.cc | 32 +- .../22_locale/codecvt/codecvt_utf16/79980.cc | 2 +- 6 files changed, 1493 insertions(+), 449 deletions(-) diff --git a/libstdc++-v3/src/c++11/codecvt.cc b/libstdc++-v3/src/c++11/codecvt.cc index 02f05752de8..2cc812cfc34 100644 --- a/libstdc++-v3/src/c++11/codecvt.cc +++ b/libstdc++-v3/src/c++11/codecvt.cc @@ -284,6 +284,8 @@ namespace return invalid_mb_sequence; if (c1 == 0xE0 && c2 < 0xA0) [[unlikely]] // overlong return invalid_mb_sequence; + if (c1 == 0xED && c2 >= 0xA0) [[unlikely]] // surrogate + return invalid_mb_sequence; if (avail < 3) [[unlikely]] return incomplete_mb_character; char32_t c3 = (unsigned char) from[2]; @@ -484,6 +486,8 @@ namespace while (from.size()) { const char32_t c = from[0]; + if (0xD800 <= c && c <= 0xDFFF) [[unlikely]] + return codecvt_base::error; if (c > maxcode) [[unlikely]] return codecvt_base::error; if (!write_utf8_code_point(to, c)) [[unlikely]] @@ -508,7 +512,7 @@ namespace return codecvt_base::error; to = codepoint; } - return from.size() ? codecvt_base::partial : codecvt_base::ok; + return from.nbytes() ? codecvt_base::partial : codecvt_base::ok; } // ucs4 -> utf16 @@ -521,6 +525,8 @@ namespace while (from.size()) { const char32_t c = from[0]; + if (0xD800 <= c && c <= 0xDFFF) [[unlikely]] + return codecvt_base::error; if (c > maxcode) [[unlikely]] return codecvt_base::error; if (!write_utf16_code_point(to, c, mode)) [[unlikely]] @@ -653,7 +659,7 @@ namespace while (from.size() && to.size()) { char16_t c = from[0]; - if (is_high_surrogate(c)) + if (0xD800 <= c && c <= 0xDFFF) return codecvt_base::error; if (c > maxcode) return codecvt_base::error; @@ -680,7 +686,7 @@ namespace return codecvt_base::error; to = c; } - return from.size() == 0 ? codecvt_base::ok : codecvt_base::partial; + return from.nbytes() == 0 ? codecvt_base::ok : codecvt_base::partial; } const char16_t* @@ -1344,8 +1350,6 @@ do_in(state_type&, const extern_type* __from, const extern_type* __from_end, auto res = ucs2_in(from, to, _M_maxcode, _M_mode); __from_next = reinterpret_cast(from.next); __to_next = to.next; - if (res == codecvt_base::ok && __from_next != __from_end) - res = codecvt_base::error; return res; } @@ -1419,8 +1423,6 @@ do_in(state_type&, const extern_type* __from, const extern_type* __from_end, auto res = ucs4_in(from, to, _M_maxcode, _M_mode); __from_next = reinterpret_cast(from.next); __to_next = to.next; - if (res == codecvt_base::ok && __from_next != __from_end) - res = codecvt_base::error; return res; } @@ -1521,8 +1523,6 @@ do_in(state_type&, const extern_type* __from, const extern_type* __from_end, #endif __from_next = reinterpret_cast(from.next); __to_next = reinterpret_cast(to.next); - if (res == codecvt_base::ok && __from_next != __from_end) - res = codecvt_base::error; return res; } diff --git a/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode.cc b/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode.cc index df1a2b4cc51..c563781ca72 100644 --- a/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode.cc +++ b/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode.cc @@ -27,38 +27,58 @@ void test_utf8_utf32_codecvts () { using codecvt_c32 = codecvt; - auto loc_c = locale::classic (); + auto &loc_c = locale::classic (); VERIFY (has_facet (loc_c)); auto &cvt = use_facet (loc_c); - test_utf8_utf32_codecvts (cvt); + test_utf8_utf32_cvt (cvt); codecvt_utf8 cvt2; - test_utf8_utf32_codecvts (cvt2); + test_utf8_utf32_cvt (cvt2); } void test_utf8_utf16_codecvts () { using codecvt_c16 = codecvt; - auto loc_c = locale::classic (); + auto &loc_c = locale::classic (); VERIFY (has_facet (loc_c)); auto &cvt = use_facet (loc_c); - test_utf8_utf16_cvts (cvt); + test_utf8_utf16_cvt (cvt); codecvt_utf8_utf16 cvt2; - test_utf8_utf16_cvts (cvt2); + test_utf8_utf16_cvt (cvt2); codecvt_utf8_utf16 cvt3; - test_utf8_utf16_cvts (cvt3); + test_utf8_utf16_cvt (cvt3); } void test_utf8_ucs2_codecvts () { codecvt_utf8 cvt; - test_utf8_ucs2_cvts (cvt); + test_utf8_ucs2_cvt (cvt); +} + +void +test_utf16_utf32_codecvts () +{ + codecvt_utf16 cvt; + test_utf16_utf32_cvt (cvt, utf16_big_endian); + + codecvt_utf16 cvt2; + test_utf16_utf32_cvt (cvt2, utf16_little_endian); +} + +void +test_utf16_ucs2_codecvts () +{ + codecvt_utf16 cvt; + test_utf16_ucs2_cvt (cvt, utf16_big_endian); + + codecvt_utf16 cvt2; + test_utf16_ucs2_cvt (cvt2, utf16_little_endian); } int @@ -67,4 +87,6 @@ main () test_utf8_utf32_codecvts (); test_utf8_utf16_codecvts (); test_utf8_ucs2_codecvts (); + test_utf16_utf32_codecvts (); + test_utf16_ucs2_codecvts (); } diff --git a/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode.h b/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode.h index fbdc7a35b28..d3ae42fac13 100644 --- a/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode.h +++ b/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode.h @@ -42,33 +42,33 @@ auto constexpr array_size (const T (&)[N]) -> size_t return N; } -template +template void -utf8_to_utf32_in_ok (const std::codecvt &cvt) +utf8_to_utf32_in_ok (const std::codecvt &cvt) { using namespace std; // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP and 4-byte CP - const char in[] = "bш\uAAAA\U0010AAAA"; - const char32_t exp_literal[] = U"bш\uAAAA\U0010AAAA"; - CharT exp[array_size (exp_literal)] = {}; - std::copy (begin (exp_literal), end (exp_literal), begin (exp)); - - static_assert (array_size (in) == 11, ""); - static_assert (array_size (exp_literal) == 5, ""); - static_assert (array_size (exp) == 5, ""); - VERIFY (char_traits::length (in) == 10); - VERIFY (char_traits::length (exp_literal) == 4); - VERIFY (char_traits::length (exp) == 4); + const unsigned char input[] = "b\u0448\uAAAA\U0010AAAA"; + const char32_t expected[] = U"b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 11, ""); + static_assert (array_size (expected) == 5, ""); + + ExternT in[array_size (input)]; + InternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 10); + VERIFY (char_traits::length (exp) == 4); test_offsets_ok offsets[] = {{0, 0}, {1, 1}, {3, 2}, {6, 3}, {10, 4}}; for (auto t : offsets) { - CharT out[array_size (exp) - 1] = {}; + InternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); auto state = mbstate_t{}; - auto in_next = (const char *) nullptr; - auto out_next = (CharT *) nullptr; + auto in_next = (const ExternT *) nullptr; + auto out_next = (InternT *) nullptr; auto res = codecvt_base::result (); res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -76,19 +76,19 @@ utf8_to_utf32_in_ok (const std::codecvt &cvt) VERIFY (res == cvt.ok); VERIFY (in_next == in + t.in_size); VERIFY (out_next == out + t.out_size); - VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); if (t.out_size < array_size (out)) VERIFY (out[t.out_size] == 0); } for (auto t : offsets) { - CharT out[array_size (exp)] = {}; + InternT out[array_size (exp)] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); auto state = mbstate_t{}; - auto in_next = (const char *) nullptr; - auto out_next = (CharT *) nullptr; + auto in_next = (const ExternT *) nullptr; + auto out_next = (InternT *) nullptr; auto res = codecvt_base::result (); res @@ -96,29 +96,29 @@ utf8_to_utf32_in_ok (const std::codecvt &cvt) VERIFY (res == cvt.ok); VERIFY (in_next == in + t.in_size); VERIFY (out_next == out + t.out_size); - VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); if (t.out_size < array_size (out)) VERIFY (out[t.out_size] == 0); } } -template +template void -utf8_to_utf32_in_partial (const std::codecvt &cvt) +utf8_to_utf32_in_partial (const std::codecvt &cvt) { using namespace std; // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP and 4-byte CP - const char in[] = "bш\uAAAA\U0010AAAA"; - const char32_t exp_literal[] = U"bш\uAAAA\U0010AAAA"; - CharT exp[array_size (exp_literal)] = {}; - std::copy (begin (exp_literal), end (exp_literal), begin (exp)); - - static_assert (array_size (in) == 11, ""); - static_assert (array_size (exp_literal) == 5, ""); - static_assert (array_size (exp) == 5, ""); - VERIFY (char_traits::length (in) == 10); - VERIFY (char_traits::length (exp_literal) == 4); - VERIFY (char_traits::length (exp) == 4); + const unsigned char input[] = "b\u0448\uAAAA\U0010AAAA"; + const char32_t expected[] = U"b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 11, ""); + static_assert (array_size (expected) == 5, ""); + + ExternT in[array_size (input)]; + InternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 10); + VERIFY (char_traits::length (exp) == 4); test_offsets_partial offsets[] = { {1, 0, 0, 0}, // no space for first CP @@ -144,14 +144,14 @@ utf8_to_utf32_in_partial (const std::codecvt &cvt) for (auto t : offsets) { - CharT out[array_size (exp) - 1] = {}; + InternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); VERIFY (t.expected_in_next <= t.in_size); VERIFY (t.expected_out_next <= t.out_size); auto state = mbstate_t{}; - auto in_next = (const char *) nullptr; - auto out_next = (CharT *) nullptr; + auto in_next = (const ExternT *) nullptr; + auto out_next = (InternT *) nullptr; auto res = codecvt_base::result (); res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -159,37 +159,58 @@ utf8_to_utf32_in_partial (const std::codecvt &cvt) VERIFY (res == cvt.partial); VERIFY (in_next == in + t.expected_in_next); VERIFY (out_next == out + t.expected_out_next); - VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); if (t.expected_out_next < array_size (out)) VERIFY (out[t.expected_out_next] == 0); } } -template +template void -utf8_to_utf32_in_error (const std::codecvt &cvt) +utf8_to_utf32_in_error (const std::codecvt &cvt) { using namespace std; - // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP and 4-byte CP - const char valid_in[] = "bш\uAAAA\U0010AAAA"; - const char32_t exp_literal[] = U"bш\uAAAA\U0010AAAA"; - CharT exp[array_size (exp_literal)] = {}; - std::copy (begin (exp_literal), end (exp_literal), begin (exp)); + // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP, 4-byte CP + const unsigned char input[] = "b\u0448\uD700\U0010AAAA"; + const char32_t expected[] = U"b\u0448\uD700\U0010AAAA"; + static_assert (array_size (input) == 11, ""); + static_assert (array_size (expected) == 5, ""); + + ExternT in[array_size (input)]; + InternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 10); + VERIFY (char_traits::length (exp) == 4); + + // There are 5 classes of errors in UTF-8 decoding + // 1. Missing leading byte + // 2. Missing trailing byte + // 3. Surrogate CP + // 4. Overlong sequence + // 5. CP out of Unicode range + test_offsets_error offsets[] = { + + // 1. Missing leading byte. We will replace the leading byte with + // non-leading byte, such as a byte that is always invalid or a trailing + // byte. - static_assert (array_size (valid_in) == 11, ""); - static_assert (array_size (exp_literal) == 5, ""); - static_assert (array_size (exp) == 5, ""); - VERIFY (char_traits::length (valid_in) == 10); - VERIFY (char_traits::length (exp_literal) == 4); - VERIFY (char_traits::length (exp) == 4); + // replace leading byte with invalid byte + {1, 4, 0, 0, 0xFF, 0}, + {3, 4, 1, 1, 0xFF, 1}, + {6, 4, 3, 2, 0xFF, 3}, + {10, 4, 6, 3, 0xFF, 6}, - test_offsets_error offsets[] = { + // replace leading byte with trailing byte + {1, 4, 0, 0, 0b10101010, 0}, + {3, 4, 1, 1, 0b10101010, 1}, + {6, 4, 3, 2, 0b10101010, 3}, + {10, 4, 6, 3, 0b10101010, 6}, - // replace leading byte with invalid byte - {1, 4, 0, 0, '\xFF', 0}, - {3, 4, 1, 1, '\xFF', 1}, - {6, 4, 3, 2, '\xFF', 3}, - {10, 4, 6, 3, '\xFF', 6}, + // 2. Missing trailing byte. We will replace the trailing byte with + // non-trailing byte, such as a byte that is always invalid or a leading + // byte (simple ASCII byte in our case). // replace first trailing byte with ASCII byte {3, 4, 1, 1, 'z', 2}, @@ -197,21 +218,27 @@ utf8_to_utf32_in_error (const std::codecvt &cvt) {10, 4, 6, 3, 'z', 7}, // replace first trailing byte with invalid byte - {3, 4, 1, 1, '\xFF', 2}, - {6, 4, 3, 2, '\xFF', 4}, - {10, 4, 6, 3, '\xFF', 7}, + {3, 4, 1, 1, 0xFF, 2}, + {6, 4, 3, 2, 0xFF, 4}, + {10, 4, 6, 3, 0xFF, 7}, // replace second trailing byte with ASCII byte {6, 4, 3, 2, 'z', 5}, {10, 4, 6, 3, 'z', 8}, // replace second trailing byte with invalid byte - {6, 4, 3, 2, '\xFF', 5}, - {10, 4, 6, 3, '\xFF', 8}, + {6, 4, 3, 2, 0xFF, 5}, + {10, 4, 6, 3, 0xFF, 8}, // replace third trailing byte {10, 4, 6, 3, 'z', 9}, - {10, 4, 6, 3, '\xFF', 9}, + {10, 4, 6, 3, 0xFF, 9}, + + // 2.1 The following test-cases raise doubt whether error or partial should + // be returned. For example, we have 4-byte sequence with valid leading + // byte. If we hide the last byte we need to return partial. But, if the + // second or third byte, which are visible to the call to codecvt, are + // malformed then error should be returned. // replace first trailing byte with ASCII byte, also incomplete at end {5, 4, 3, 2, 'z', 4}, @@ -219,30 +246,51 @@ utf8_to_utf32_in_error (const std::codecvt &cvt) {9, 4, 6, 3, 'z', 7}, // replace first trailing byte with invalid byte, also incomplete at end - {5, 4, 3, 2, '\xFF', 4}, - {8, 4, 6, 3, '\xFF', 7}, - {9, 4, 6, 3, '\xFF', 7}, + {5, 4, 3, 2, 0xFF, 4}, + {8, 4, 6, 3, 0xFF, 7}, + {9, 4, 6, 3, 0xFF, 7}, // replace second trailing byte with ASCII byte, also incomplete at end {9, 4, 6, 3, 'z', 8}, // replace second trailing byte with invalid byte, also incomplete at end - {9, 4, 6, 3, '\xFF', 8}, + {9, 4, 6, 3, 0xFF, 8}, + + // 3. Surrogate CP. We modify the second byte (first trailing) of the 3-byte + // CP U+D700 + {6, 4, 3, 2, 0b10100000, 4}, // turn U+D700 into U+D800 + {6, 4, 3, 2, 0b10101100, 4}, // turn U+D700 into U+DB00 + {6, 4, 3, 2, 0b10110000, 4}, // turn U+D700 into U+DC00 + {6, 4, 3, 2, 0b10111100, 4}, // turn U+D700 into U+DF00 + + // 4. Overlong sequence. The CPs in the input are chosen such as modifying + // just the leading byte is enough to make them overlong, i.e. for the + // 3-byte and 4-byte CP the second byte (first trailing) has enough leading + // zeroes. + {3, 4, 1, 1, 0b11000000, 1}, // make the 2-byte CP overlong + {3, 4, 1, 1, 0b11000001, 1}, // make the 2-byte CP overlong + {6, 4, 3, 2, 0b11100000, 3}, // make the 3-byte CP overlong + {10, 4, 6, 3, 0b11110000, 6}, // make the 4-byte CP overlong + + // 5. CP above range + // turn U+10AAAA into U+14AAAA by changing its leading byte + {10, 4, 6, 3, 0b11110101, 6}, + // turn U+10AAAA into U+11AAAA by changing its 2nd byte + {10, 4, 6, 3, 0b10011010, 7}, }; for (auto t : offsets) { - char in[array_size (valid_in)] = {}; - CharT out[array_size (exp) - 1] = {}; + InternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); VERIFY (t.expected_in_next <= t.in_size); VERIFY (t.expected_out_next <= t.out_size); - char_traits::copy (in, valid_in, array_size (valid_in)); + auto old_char = in[t.replace_pos]; in[t.replace_pos] = t.replace_char; auto state = mbstate_t{}; - auto in_next = (const char *) nullptr; - auto out_next = (CharT *) nullptr; + auto in_next = (const ExternT *) nullptr; + auto out_next = (InternT *) nullptr; auto res = codecvt_base::result (); res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -250,48 +298,51 @@ utf8_to_utf32_in_error (const std::codecvt &cvt) VERIFY (res == cvt.error); VERIFY (in_next == in + t.expected_in_next); VERIFY (out_next == out + t.expected_out_next); - VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); if (t.expected_out_next < array_size (out)) VERIFY (out[t.expected_out_next] == 0); + + in[t.replace_pos] = old_char; } } -template +template void -utf8_to_utf32_in (const std::codecvt &cvt) +utf8_to_utf32_in (const std::codecvt &cvt) { utf8_to_utf32_in_ok (cvt); utf8_to_utf32_in_partial (cvt); utf8_to_utf32_in_error (cvt); } -template +template void -utf32_to_utf8_out_ok (const std::codecvt &cvt) +utf32_to_utf8_out_ok (const std::codecvt &cvt) { using namespace std; // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP and 4-byte CP - const char32_t in_literal[] = U"bш\uAAAA\U0010AAAA"; - const char exp[] = "bш\uAAAA\U0010AAAA"; - CharT in[array_size (in_literal)] = {}; - copy (begin (in_literal), end (in_literal), begin (in)); - - static_assert (array_size (in_literal) == 5, ""); - static_assert (array_size (in) == 5, ""); - static_assert (array_size (exp) == 11, ""); - VERIFY (char_traits::length (in_literal) == 4); - VERIFY (char_traits::length (in) == 4); - VERIFY (char_traits::length (exp) == 10); + const char32_t input[] = U"b\u0448\uAAAA\U0010AAAA"; + const unsigned char expected[] = "b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 5, ""); + static_assert (array_size (expected) == 11, ""); + + InternT in[array_size (input)]; + ExternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 4); + VERIFY (char_traits::length (exp) == 10); const test_offsets_ok offsets[] = {{0, 0}, {1, 1}, {2, 3}, {3, 6}, {4, 10}}; for (auto t : offsets) { - char out[array_size (exp) - 1] = {}; + ExternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); auto state = mbstate_t{}; - auto in_next = (const CharT *) nullptr; - auto out_next = (char *) nullptr; + auto in_next = (const InternT *) nullptr; + auto out_next = (ExternT *) nullptr; auto res = codecvt_base::result (); res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -299,29 +350,29 @@ utf32_to_utf8_out_ok (const std::codecvt &cvt) VERIFY (res == cvt.ok); VERIFY (in_next == in + t.in_size); VERIFY (out_next == out + t.out_size); - VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); if (t.out_size < array_size (out)) VERIFY (out[t.out_size] == 0); } } -template +template void -utf32_to_utf8_out_partial (const std::codecvt &cvt) +utf32_to_utf8_out_partial (const std::codecvt &cvt) { using namespace std; // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP and 4-byte CP - const char32_t in_literal[] = U"bш\uAAAA\U0010AAAA"; - const char exp[] = "bш\uAAAA\U0010AAAA"; - CharT in[array_size (in_literal)] = {}; - copy (begin (in_literal), end (in_literal), begin (in)); - - static_assert (array_size (in_literal) == 5, ""); - static_assert (array_size (in) == 5, ""); - static_assert (array_size (exp) == 11, ""); - VERIFY (char_traits::length (in_literal) == 4); - VERIFY (char_traits::length (in) == 4); - VERIFY (char_traits::length (exp) == 10); + const char32_t input[] = U"b\u0448\uAAAA\U0010AAAA"; + const unsigned char expected[] = "b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 5, ""); + static_assert (array_size (expected) == 11, ""); + + InternT in[array_size (input)]; + ExternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 4); + VERIFY (char_traits::length (exp) == 10); const test_offsets_partial offsets[] = { {1, 0, 0, 0}, // no space for first CP @@ -340,14 +391,14 @@ utf32_to_utf8_out_partial (const std::codecvt &cvt) }; for (auto t : offsets) { - char out[array_size (exp) - 1] = {}; + ExternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); VERIFY (t.expected_in_next <= t.in_size); VERIFY (t.expected_out_next <= t.out_size); auto state = mbstate_t{}; - auto in_next = (const CharT *) nullptr; - auto out_next = (char *) nullptr; + auto in_next = (const InternT *) nullptr; + auto out_next = (ExternT *) nullptr; auto res = codecvt_base::result (); res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -355,44 +406,58 @@ utf32_to_utf8_out_partial (const std::codecvt &cvt) VERIFY (res == cvt.partial); VERIFY (in_next == in + t.expected_in_next); VERIFY (out_next == out + t.expected_out_next); - VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); if (t.expected_out_next < array_size (out)) VERIFY (out[t.expected_out_next] == 0); } } -template +template void -utf32_to_utf8_out_error (const std::codecvt &cvt) +utf32_to_utf8_out_error (const std::codecvt &cvt) { using namespace std; - const char32_t valid_in[] = U"bш\uAAAA\U0010AAAA"; - const char exp[] = "bш\uAAAA\U0010AAAA"; - - static_assert (array_size (valid_in) == 5, ""); - static_assert (array_size (exp) == 11, ""); - VERIFY (char_traits::length (valid_in) == 4); - VERIFY (char_traits::length (exp) == 10); - - test_offsets_error offsets[] = {{4, 10, 0, 0, 0x00110000, 0}, - {4, 10, 1, 1, 0x00110000, 1}, - {4, 10, 2, 3, 0x00110000, 2}, - {4, 10, 3, 6, 0x00110000, 3}}; + // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP and 4-byte CP + const char32_t input[] = U"b\u0448\uAAAA\U0010AAAA"; + const unsigned char expected[] = "b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 5, ""); + static_assert (array_size (expected) == 11, ""); + + InternT in[array_size (input)]; + ExternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 4); + VERIFY (char_traits::length (exp) == 10); + + test_offsets_error offsets[] = { + + // Surrogate CP + {4, 10, 0, 0, 0xD800, 0}, + {4, 10, 1, 1, 0xDBFF, 1}, + {4, 10, 2, 3, 0xDC00, 2}, + {4, 10, 3, 6, 0xDFFF, 3}, + + // CP out of range + {4, 10, 0, 0, 0x00110000, 0}, + {4, 10, 1, 1, 0x00110000, 1}, + {4, 10, 2, 3, 0x00110000, 2}, + {4, 10, 3, 6, 0x00110000, 3}}; for (auto t : offsets) { - CharT in[array_size (valid_in)] = {}; - char out[array_size (exp) - 1] = {}; + ExternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); VERIFY (t.expected_in_next <= t.in_size); VERIFY (t.expected_out_next <= t.out_size); - copy (begin (valid_in), end (valid_in), begin (in)); + auto old_char = in[t.replace_pos]; in[t.replace_pos] = t.replace_char; auto state = mbstate_t{}; - auto in_next = (const CharT *) nullptr; - auto out_next = (char *) nullptr; + auto in_next = (const InternT *) nullptr; + auto out_next = (ExternT *) nullptr; auto res = codecvt_base::result (); res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -400,56 +465,59 @@ utf32_to_utf8_out_error (const std::codecvt &cvt) VERIFY (res == cvt.error); VERIFY (in_next == in + t.expected_in_next); VERIFY (out_next == out + t.expected_out_next); - VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); if (t.expected_out_next < array_size (out)) VERIFY (out[t.expected_out_next] == 0); + + in[t.replace_pos] = old_char; } } -template +template void -utf32_to_utf8_out (const std::codecvt &cvt) +utf32_to_utf8_out (const std::codecvt &cvt) { utf32_to_utf8_out_ok (cvt); utf32_to_utf8_out_partial (cvt); utf32_to_utf8_out_error (cvt); } -template +template void -test_utf8_utf32_codecvts (const std::codecvt &cvt) +test_utf8_utf32_cvt (const std::codecvt &cvt) { utf8_to_utf32_in (cvt); utf32_to_utf8_out (cvt); } -template +template void -utf8_to_utf16_in_ok (const std::codecvt &cvt) +utf8_to_utf16_in_ok (const std::codecvt &cvt) { using namespace std; // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP and 4-byte CP - const char in[] = "bш\uAAAA\U0010AAAA"; - const char16_t exp_literal[] = u"bш\uAAAA\U0010AAAA"; - CharT exp[array_size (exp_literal)] = {}; - copy (begin (exp_literal), end (exp_literal), begin (exp)); - - static_assert (array_size (in) == 11, ""); - static_assert (array_size (exp_literal) == 6, ""); - static_assert (array_size (exp) == 6, ""); - VERIFY (char_traits::length (in) == 10); - VERIFY (char_traits::length (exp_literal) == 5); - VERIFY (char_traits::length (exp) == 5); + const unsigned char input[] = "b\u0448\uAAAA\U0010AAAA"; + const char16_t expected[] = u"b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 11, ""); + static_assert (array_size (expected) == 6, ""); + + ExternT in[array_size (input)]; + InternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 10); + VERIFY (char_traits::length (exp) == 5); test_offsets_ok offsets[] = {{0, 0}, {1, 1}, {3, 2}, {6, 3}, {10, 5}}; for (auto t : offsets) { - CharT out[array_size (exp) - 1] = {}; + InternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); auto state = mbstate_t{}; - auto in_next = (const char *) nullptr; - auto out_next = (CharT *) nullptr; + auto in_next = (const ExternT *) nullptr; + auto out_next = (InternT *) nullptr; auto res = codecvt_base::result (); res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -457,19 +525,19 @@ utf8_to_utf16_in_ok (const std::codecvt &cvt) VERIFY (res == cvt.ok); VERIFY (in_next == in + t.in_size); VERIFY (out_next == out + t.out_size); - VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); if (t.out_size < array_size (out)) VERIFY (out[t.out_size] == 0); } for (auto t : offsets) { - CharT out[array_size (exp)] = {}; + InternT out[array_size (exp)] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); auto state = mbstate_t{}; - auto in_next = (const char *) nullptr; - auto out_next = (CharT *) nullptr; + auto in_next = (const ExternT *) nullptr; + auto out_next = (InternT *) nullptr; auto res = codecvt_base::result (); res @@ -477,29 +545,29 @@ utf8_to_utf16_in_ok (const std::codecvt &cvt) VERIFY (res == cvt.ok); VERIFY (in_next == in + t.in_size); VERIFY (out_next == out + t.out_size); - VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); if (t.out_size < array_size (out)) VERIFY (out[t.out_size] == 0); } } -template +template void -utf8_to_utf16_in_partial (const std::codecvt &cvt) +utf8_to_utf16_in_partial (const std::codecvt &cvt) { using namespace std; // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP and 4-byte CP - const char in[] = "bш\uAAAA\U0010AAAA"; - const char16_t exp_literal[] = u"bш\uAAAA\U0010AAAA"; - CharT exp[array_size (exp_literal)] = {}; - copy (begin (exp_literal), end (exp_literal), begin (exp)); - - static_assert (array_size (in) == 11, ""); - static_assert (array_size (exp_literal) == 6, ""); - static_assert (array_size (exp) == 6, ""); - VERIFY (char_traits::length (in) == 10); - VERIFY (char_traits::length (exp_literal) == 5); - VERIFY (char_traits::length (exp) == 5); + const unsigned char input[] = "b\u0448\uAAAA\U0010AAAA"; + const char16_t expected[] = u"b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 11, ""); + static_assert (array_size (expected) == 6, ""); + + ExternT in[array_size (input)]; + InternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 10); + VERIFY (char_traits::length (exp) == 5); test_offsets_partial offsets[] = { {1, 0, 0, 0}, // no space for first CP @@ -530,14 +598,14 @@ utf8_to_utf16_in_partial (const std::codecvt &cvt) for (auto t : offsets) { - CharT out[array_size (exp) - 1] = {}; + InternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); VERIFY (t.expected_in_next <= t.in_size); VERIFY (t.expected_out_next <= t.out_size); auto state = mbstate_t{}; - auto in_next = (const char *) nullptr; - auto out_next = (CharT *) nullptr; + auto in_next = (const ExternT *) nullptr; + auto out_next = (InternT *) nullptr; auto res = codecvt_base::result (); res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -545,36 +613,58 @@ utf8_to_utf16_in_partial (const std::codecvt &cvt) VERIFY (res == cvt.partial); VERIFY (in_next == in + t.expected_in_next); VERIFY (out_next == out + t.expected_out_next); - VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); if (t.expected_out_next < array_size (out)) VERIFY (out[t.expected_out_next] == 0); } } -template +template void -utf8_to_utf16_in_error (const std::codecvt &cvt) +utf8_to_utf16_in_error (const std::codecvt &cvt) { using namespace std; - const char valid_in[] = "bш\uAAAA\U0010AAAA"; - const char16_t exp_literal[] = u"bш\uAAAA\U0010AAAA"; - CharT exp[array_size (exp_literal)] = {}; - copy (begin (exp_literal), end (exp_literal), begin (exp)); + // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP, 4-byte CP + const unsigned char input[] = "b\u0448\uD700\U0010AAAA"; + const char16_t expected[] = u"b\u0448\uD700\U0010AAAA"; + static_assert (array_size (input) == 11, ""); + static_assert (array_size (expected) == 6, ""); + + ExternT in[array_size (input)]; + InternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 10); + VERIFY (char_traits::length (exp) == 5); + + // There are 5 classes of errors in UTF-8 decoding + // 1. Missing leading byte + // 2. Missing trailing byte + // 3. Surrogate CP + // 4. Overlong sequence + // 5. CP out of Unicode range + test_offsets_error offsets[] = { + + // 1. Missing leading byte. We will replace the leading byte with + // non-leading byte, such as a byte that is always invalid or a trailing + // byte. - static_assert (array_size (valid_in) == 11, ""); - static_assert (array_size (exp_literal) == 6, ""); - static_assert (array_size (exp) == 6, ""); - VERIFY (char_traits::length (valid_in) == 10); - VERIFY (char_traits::length (exp_literal) == 5); - VERIFY (char_traits::length (exp) == 5); + // replace leading byte with invalid byte + {1, 5, 0, 0, 0xFF, 0}, + {3, 5, 1, 1, 0xFF, 1}, + {6, 5, 3, 2, 0xFF, 3}, + {10, 5, 6, 3, 0xFF, 6}, - test_offsets_error offsets[] = { + // replace leading byte with trailing byte + {1, 5, 0, 0, 0b10101010, 0}, + {3, 5, 1, 1, 0b10101010, 1}, + {6, 5, 3, 2, 0b10101010, 3}, + {10, 5, 6, 3, 0b10101010, 6}, - // replace leading byte with invalid byte - {1, 5, 0, 0, '\xFF', 0}, - {3, 5, 1, 1, '\xFF', 1}, - {6, 5, 3, 2, '\xFF', 3}, - {10, 5, 6, 3, '\xFF', 6}, + // 2. Missing trailing byte. We will replace the trailing byte with + // non-trailing byte, such as a byte that is always invalid or a leading + // byte (simple ASCII byte in our case). // replace first trailing byte with ASCII byte {3, 5, 1, 1, 'z', 2}, @@ -582,21 +672,27 @@ utf8_to_utf16_in_error (const std::codecvt &cvt) {10, 5, 6, 3, 'z', 7}, // replace first trailing byte with invalid byte - {3, 5, 1, 1, '\xFF', 2}, - {6, 5, 3, 2, '\xFF', 4}, - {10, 5, 6, 3, '\xFF', 7}, + {3, 5, 1, 1, 0xFF, 2}, + {6, 5, 3, 2, 0xFF, 4}, + {10, 5, 6, 3, 0xFF, 7}, // replace second trailing byte with ASCII byte {6, 5, 3, 2, 'z', 5}, {10, 5, 6, 3, 'z', 8}, // replace second trailing byte with invalid byte - {6, 5, 3, 2, '\xFF', 5}, - {10, 5, 6, 3, '\xFF', 8}, + {6, 5, 3, 2, 0xFF, 5}, + {10, 5, 6, 3, 0xFF, 8}, // replace third trailing byte {10, 5, 6, 3, 'z', 9}, - {10, 5, 6, 3, '\xFF', 9}, + {10, 5, 6, 3, 0xFF, 9}, + + // 2.1 The following test-cases raise doubt whether error or partial should + // be returned. For example, we have 4-byte sequence with valid leading + // byte. If we hide the last byte we need to return partial. But, if the + // second or third byte, which are visible to the call to codecvt, are + // malformed then error should be returned. // replace first trailing byte with ASCII byte, also incomplete at end {5, 5, 3, 2, 'z', 4}, @@ -604,30 +700,51 @@ utf8_to_utf16_in_error (const std::codecvt &cvt) {9, 5, 6, 3, 'z', 7}, // replace first trailing byte with invalid byte, also incomplete at end - {5, 5, 3, 2, '\xFF', 4}, - {8, 5, 6, 3, '\xFF', 7}, - {9, 5, 6, 3, '\xFF', 7}, + {5, 5, 3, 2, 0xFF, 4}, + {8, 5, 6, 3, 0xFF, 7}, + {9, 5, 6, 3, 0xFF, 7}, // replace second trailing byte with ASCII byte, also incomplete at end {9, 5, 6, 3, 'z', 8}, // replace second trailing byte with invalid byte, also incomplete at end - {9, 5, 6, 3, '\xFF', 8}, + {9, 5, 6, 3, 0xFF, 8}, + + // 3. Surrogate CP. We modify the second byte (first trailing) of the 3-byte + // CP U+D700 + {6, 5, 3, 2, 0b10100000, 4}, // turn U+D700 into U+D800 + {6, 5, 3, 2, 0b10101100, 4}, // turn U+D700 into U+DB00 + {6, 5, 3, 2, 0b10110000, 4}, // turn U+D700 into U+DC00 + {6, 5, 3, 2, 0b10111100, 4}, // turn U+D700 into U+DF00 + + // 4. Overlong sequence. The CPs in the input are chosen such as modifying + // just the leading byte is enough to make them overlong, i.e. for the + // 3-byte and 4-byte CP the second byte (first trailing) has enough leading + // zeroes. + {3, 5, 1, 1, 0b11000000, 1}, // make the 2-byte CP overlong + {3, 5, 1, 1, 0b11000001, 1}, // make the 2-byte CP overlong + {6, 5, 3, 2, 0b11100000, 3}, // make the 3-byte CP overlong + {10, 5, 6, 3, 0b11110000, 6}, // make the 4-byte CP overlong + + // 5. CP above range + // turn U+10AAAA into U+14AAAA by changing its leading byte + {10, 5, 6, 3, 0b11110101, 6}, + // turn U+10AAAA into U+11AAAA by changing its 2nd byte + {10, 5, 6, 3, 0b10011010, 7}, }; for (auto t : offsets) { - char in[array_size (valid_in)] = {}; - CharT out[array_size (exp) - 1] = {}; + InternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); VERIFY (t.expected_in_next <= t.in_size); VERIFY (t.expected_out_next <= t.out_size); - char_traits::copy (in, valid_in, array_size (valid_in)); + auto old_char = in[t.replace_pos]; in[t.replace_pos] = t.replace_char; auto state = mbstate_t{}; - auto in_next = (const char *) nullptr; - auto out_next = (CharT *) nullptr; + auto in_next = (const ExternT *) nullptr; + auto out_next = (InternT *) nullptr; auto res = codecvt_base::result (); res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -635,48 +752,51 @@ utf8_to_utf16_in_error (const std::codecvt &cvt) VERIFY (res == cvt.error); VERIFY (in_next == in + t.expected_in_next); VERIFY (out_next == out + t.expected_out_next); - VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); if (t.expected_out_next < array_size (out)) VERIFY (out[t.expected_out_next] == 0); + + in[t.replace_pos] = old_char; } } -template +template void -utf8_to_utf16_in (const std::codecvt &cvt) +utf8_to_utf16_in (const std::codecvt &cvt) { utf8_to_utf16_in_ok (cvt); utf8_to_utf16_in_partial (cvt); utf8_to_utf16_in_error (cvt); } -template +template void -utf16_to_utf8_out_ok (const std::codecvt &cvt) +utf16_to_utf8_out_ok (const std::codecvt &cvt) { using namespace std; // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP and 4-byte CP - const char16_t in_literal[] = u"bш\uAAAA\U0010AAAA"; - const char exp[] = "bш\uAAAA\U0010AAAA"; - CharT in[array_size (in_literal)]; - copy (begin (in_literal), end (in_literal), begin (in)); - - static_assert (array_size (in_literal) == 6, ""); - static_assert (array_size (exp) == 11, ""); - static_assert (array_size (in) == 6, ""); - VERIFY (char_traits::length (in_literal) == 5); - VERIFY (char_traits::length (exp) == 10); - VERIFY (char_traits::length (in) == 5); + const char16_t input[] = u"b\u0448\uAAAA\U0010AAAA"; + const unsigned char expected[] = "b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 6, ""); + static_assert (array_size (expected) == 11, ""); + + InternT in[array_size (input)]; + ExternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 5); + VERIFY (char_traits::length (exp) == 10); const test_offsets_ok offsets[] = {{0, 0}, {1, 1}, {2, 3}, {3, 6}, {5, 10}}; for (auto t : offsets) { - char out[array_size (exp) - 1] = {}; + ExternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); auto state = mbstate_t{}; - auto in_next = (const CharT *) nullptr; - auto out_next = (char *) nullptr; + auto in_next = (const InternT *) nullptr; + auto out_next = (ExternT *) nullptr; auto res = codecvt_base::result (); res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -684,29 +804,29 @@ utf16_to_utf8_out_ok (const std::codecvt &cvt) VERIFY (res == cvt.ok); VERIFY (in_next == in + t.in_size); VERIFY (out_next == out + t.out_size); - VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); if (t.out_size < array_size (out)) VERIFY (out[t.out_size] == 0); } } -template +template void -utf16_to_utf8_out_partial (const std::codecvt &cvt) +utf16_to_utf8_out_partial (const std::codecvt &cvt) { using namespace std; // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP and 4-byte CP - const char16_t in_literal[] = u"bш\uAAAA\U0010AAAA"; - const char exp[] = "bш\uAAAA\U0010AAAA"; - CharT in[array_size (in_literal)]; - copy (begin (in_literal), end (in_literal), begin (in)); - - static_assert (array_size (in_literal) == 6, ""); - static_assert (array_size (exp) == 11, ""); - static_assert (array_size (in) == 6, ""); - VERIFY (char_traits::length (in_literal) == 5); - VERIFY (char_traits::length (exp) == 10); - VERIFY (char_traits::length (in) == 5); + const char16_t input[] = u"b\u0448\uAAAA\U0010AAAA"; + const unsigned char expected[] = "b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 6, ""); + static_assert (array_size (expected) == 11, ""); + + InternT in[array_size (input)]; + ExternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 5); + VERIFY (char_traits::length (exp) == 10); const test_offsets_partial offsets[] = { {1, 0, 0, 0}, // no space for first CP @@ -732,14 +852,14 @@ utf16_to_utf8_out_partial (const std::codecvt &cvt) }; for (auto t : offsets) { - char out[array_size (exp) - 1] = {}; + ExternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); VERIFY (t.expected_in_next <= t.in_size); VERIFY (t.expected_out_next <= t.out_size); auto state = mbstate_t{}; - auto in_next = (const CharT *) nullptr; - auto out_next = (char *) nullptr; + auto in_next = (const InternT *) nullptr; + auto out_next = (ExternT *) nullptr; auto res = codecvt_base::result (); res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -747,26 +867,34 @@ utf16_to_utf8_out_partial (const std::codecvt &cvt) VERIFY (res == cvt.partial); VERIFY (in_next == in + t.expected_in_next); VERIFY (out_next == out + t.expected_out_next); - VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); if (t.expected_out_next < array_size (out)) VERIFY (out[t.expected_out_next] == 0); } } -template +template void -utf16_to_utf8_out_error (const std::codecvt &cvt) +utf16_to_utf8_out_error (const std::codecvt &cvt) { using namespace std; - const char16_t valid_in[] = u"bш\uAAAA\U0010AAAA"; - const char exp[] = "bш\uAAAA\U0010AAAA"; - - static_assert (array_size (valid_in) == 6, ""); - static_assert (array_size (exp) == 11, ""); - VERIFY (char_traits::length (valid_in) == 5); - VERIFY (char_traits::length (exp) == 10); - - test_offsets_error offsets[] = { + // UTF-8 string of 1-byte CP, 2-byte CP, 3-byte CP and 4-byte CP + const char16_t input[] = u"b\u0448\uAAAA\U0010AAAA"; + const unsigned char expected[] = "b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 6, ""); + static_assert (array_size (expected) == 11, ""); + + InternT in[array_size (input)]; + ExternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 5); + VERIFY (char_traits::length (exp) == 10); + + // The only possible error in UTF-16 is unpaired surrogate code units. + // So we replace valid code points (scalar values) with lone surrogate CU. + test_offsets_error offsets[] = { {5, 10, 0, 0, 0xD800, 0}, {5, 10, 0, 0, 0xDBFF, 0}, {5, 10, 0, 0, 0xDC00, 0}, @@ -796,18 +924,17 @@ utf16_to_utf8_out_error (const std::codecvt &cvt) for (auto t : offsets) { - CharT in[array_size (valid_in)] = {}; - char out[array_size (exp) - 1] = {}; + ExternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); VERIFY (t.expected_in_next <= t.in_size); VERIFY (t.expected_out_next <= t.out_size); - copy (begin (valid_in), end (valid_in), begin (in)); + auto old_char = in[t.replace_pos]; in[t.replace_pos] = t.replace_char; auto state = mbstate_t{}; - auto in_next = (const CharT *) nullptr; - auto out_next = (char *) nullptr; + auto in_next = (const InternT *) nullptr; + auto out_next = (ExternT *) nullptr; auto res = codecvt_base::result (); res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -815,56 +942,59 @@ utf16_to_utf8_out_error (const std::codecvt &cvt) VERIFY (res == cvt.error); VERIFY (in_next == in + t.expected_in_next); VERIFY (out_next == out + t.expected_out_next); - VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); if (t.expected_out_next < array_size (out)) VERIFY (out[t.expected_out_next] == 0); + + in[t.replace_pos] = old_char; } } -template +template void -utf16_to_utf8_out (const std::codecvt &cvt) +utf16_to_utf8_out (const std::codecvt &cvt) { utf16_to_utf8_out_ok (cvt); utf16_to_utf8_out_partial (cvt); utf16_to_utf8_out_error (cvt); } -template +template void -test_utf8_utf16_cvts (const std::codecvt &cvt) +test_utf8_utf16_cvt (const std::codecvt &cvt) { utf8_to_utf16_in (cvt); utf16_to_utf8_out (cvt); } -template +template void -utf8_to_ucs2_in_ok (const std::codecvt &cvt) +utf8_to_ucs2_in_ok (const std::codecvt &cvt) { using namespace std; // UTF-8 string of 1-byte CP, 2-byte CP and 3-byte CP - const char in[] = "bш\uAAAA"; - const char16_t exp_literal[] = u"bш\uAAAA"; - CharT exp[array_size (exp_literal)] = {}; - copy (begin (exp_literal), end (exp_literal), begin (exp)); - - static_assert (array_size (in) == 7, ""); - static_assert (array_size (exp_literal) == 4, ""); - static_assert (array_size (exp) == 4, ""); - VERIFY (char_traits::length (in) == 6); - VERIFY (char_traits::length (exp_literal) == 3); - VERIFY (char_traits::length (exp) == 3); + const unsigned char input[] = "b\u0448\uAAAA"; + const char16_t expected[] = u"b\u0448\uAAAA"; + static_assert (array_size (input) == 7, ""); + static_assert (array_size (expected) == 4, ""); + + ExternT in[array_size (input)]; + InternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 6); + VERIFY (char_traits::length (exp) == 3); test_offsets_ok offsets[] = {{0, 0}, {1, 1}, {3, 2}, {6, 3}}; for (auto t : offsets) { - CharT out[array_size (exp) - 1] = {}; + InternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); auto state = mbstate_t{}; - auto in_next = (const char *) nullptr; - auto out_next = (CharT *) nullptr; + auto in_next = (const ExternT *) nullptr; + auto out_next = (InternT *) nullptr; auto res = codecvt_base::result (); res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -872,19 +1002,19 @@ utf8_to_ucs2_in_ok (const std::codecvt &cvt) VERIFY (res == cvt.ok); VERIFY (in_next == in + t.in_size); VERIFY (out_next == out + t.out_size); - VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); if (t.out_size < array_size (out)) VERIFY (out[t.out_size] == 0); } for (auto t : offsets) { - CharT out[array_size (exp)] = {}; + InternT out[array_size (exp)] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); auto state = mbstate_t{}; - auto in_next = (const char *) nullptr; - auto out_next = (CharT *) nullptr; + auto in_next = (const ExternT *) nullptr; + auto out_next = (InternT *) nullptr; auto res = codecvt_base::result (); res @@ -892,29 +1022,29 @@ utf8_to_ucs2_in_ok (const std::codecvt &cvt) VERIFY (res == cvt.ok); VERIFY (in_next == in + t.in_size); VERIFY (out_next == out + t.out_size); - VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); if (t.out_size < array_size (out)) VERIFY (out[t.out_size] == 0); } } -template +template void -utf8_to_ucs2_in_partial (const std::codecvt &cvt) +utf8_to_ucs2_in_partial (const std::codecvt &cvt) { using namespace std; // UTF-8 string of 1-byte CP, 2-byte CP and 3-byte CP - const char in[] = "bш\uAAAA"; - const char16_t exp_literal[] = u"bш\uAAAA"; - CharT exp[array_size (exp_literal)] = {}; - copy (begin (exp_literal), end (exp_literal), begin (exp)); - - static_assert (array_size (in) == 7, ""); - static_assert (array_size (exp_literal) == 4, ""); - static_assert (array_size (exp) == 4, ""); - VERIFY (char_traits::length (in) == 6); - VERIFY (char_traits::length (exp_literal) == 3); - VERIFY (char_traits::length (exp) == 3); + const unsigned char input[] = "b\u0448\uAAAA"; + const char16_t expected[] = u"b\u0448\uAAAA"; + static_assert (array_size (input) == 7, ""); + static_assert (array_size (expected) == 4, ""); + + ExternT in[array_size (input)]; + InternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 6); + VERIFY (char_traits::length (exp) == 3); test_offsets_partial offsets[] = { {1, 0, 0, 0}, // no space for first CP @@ -932,14 +1062,14 @@ utf8_to_ucs2_in_partial (const std::codecvt &cvt) for (auto t : offsets) { - CharT out[array_size (exp) - 1] = {}; + InternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); VERIFY (t.expected_in_next <= t.in_size); VERIFY (t.expected_out_next <= t.out_size); auto state = mbstate_t{}; - auto in_next = (const char *) nullptr; - auto out_next = (CharT *) nullptr; + auto in_next = (const ExternT *) nullptr; + auto out_next = (InternT *) nullptr; auto res = codecvt_base::result (); res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -947,36 +1077,57 @@ utf8_to_ucs2_in_partial (const std::codecvt &cvt) VERIFY (res == cvt.partial); VERIFY (in_next == in + t.expected_in_next); VERIFY (out_next == out + t.expected_out_next); - VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); if (t.expected_out_next < array_size (out)) VERIFY (out[t.expected_out_next] == 0); } } -template +template void -utf8_to_ucs2_in_error (const std::codecvt &cvt) +utf8_to_ucs2_in_error (const std::codecvt &cvt) { using namespace std; - const char valid_in[] = "bш\uAAAA\U0010AAAA"; - const char16_t exp_literal[] = u"bш\uAAAA\U0010AAAA"; - CharT exp[array_size (exp_literal)] = {}; - copy (begin (exp_literal), end (exp_literal), begin (exp)); + const unsigned char input[] = "b\u0448\uD700\U0010AAAA"; + const char16_t expected[] = u"b\u0448\uD700\U0010AAAA"; + static_assert (array_size (input) == 11, ""); + static_assert (array_size (expected) == 6, ""); + + ExternT in[array_size (input)]; + InternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 10); + VERIFY (char_traits::length (exp) == 5); + + // There are 5 classes of errors in UTF-8 decoding + // 1. Missing leading byte + // 2. Missing trailing byte + // 3. Surrogate CP + // 4. Overlong sequence + // 5. CP out of Unicode range + test_offsets_error offsets[] = { + + // 1. Missing leading byte. We will replace the leading byte with + // non-leading byte, such as a byte that is always invalid or a trailing + // byte. - static_assert (array_size (valid_in) == 11, ""); - static_assert (array_size (exp_literal) == 6, ""); - static_assert (array_size (exp) == 6, ""); - VERIFY (char_traits::length (valid_in) == 10); - VERIFY (char_traits::length (exp_literal) == 5); - VERIFY (char_traits::length (exp) == 5); + // replace leading byte with invalid byte + {1, 5, 0, 0, 0xFF, 0}, + {3, 5, 1, 1, 0xFF, 1}, + {6, 5, 3, 2, 0xFF, 3}, + {10, 5, 6, 3, 0xFF, 6}, - test_offsets_error offsets[] = { + // replace leading byte with trailing byte + {1, 5, 0, 0, 0b10101010, 0}, + {3, 5, 1, 1, 0b10101010, 1}, + {6, 5, 3, 2, 0b10101010, 3}, + {10, 5, 6, 3, 0b10101010, 6}, - // replace leading byte with invalid byte - {1, 5, 0, 0, '\xFF', 0}, - {3, 5, 1, 1, '\xFF', 1}, - {6, 5, 3, 2, '\xFF', 3}, - {10, 5, 6, 3, '\xFF', 6}, + // 2. Missing trailing byte. We will replace the trailing byte with + // non-trailing byte, such as a byte that is always invalid or a leading + // byte (simple ASCII byte in our case). // replace first trailing byte with ASCII byte {3, 5, 1, 1, 'z', 2}, @@ -984,72 +1135,90 @@ utf8_to_ucs2_in_error (const std::codecvt &cvt) {10, 5, 6, 3, 'z', 7}, // replace first trailing byte with invalid byte - {3, 5, 1, 1, '\xFF', 2}, - {6, 5, 3, 2, '\xFF', 4}, - {10, 5, 6, 3, '\xFF', 7}, + {3, 5, 1, 1, 0xFF, 2}, + {6, 5, 3, 2, 0xFF, 4}, + {10, 5, 6, 3, 0xFF, 7}, // replace second trailing byte with ASCII byte {6, 5, 3, 2, 'z', 5}, {10, 5, 6, 3, 'z', 8}, // replace second trailing byte with invalid byte - {6, 5, 3, 2, '\xFF', 5}, - {10, 5, 6, 3, '\xFF', 8}, + {6, 5, 3, 2, 0xFF, 5}, + {10, 5, 6, 3, 0xFF, 8}, // replace third trailing byte {10, 5, 6, 3, 'z', 9}, - {10, 5, 6, 3, '\xFF', 9}, - - // When we see a leading byte of 4-byte CP, we should return error, no - // matter if it is incomplete at the end or has errors in the trailing - // bytes. - - // Don't replace anything, show full 4-byte CP - {10, 4, 6, 3, 'b', 0}, - {10, 5, 6, 3, 'b', 0}, + {10, 5, 6, 3, 0xFF, 9}, - // Don't replace anything, show incomplete 4-byte CP at the end - {7, 4, 6, 3, 'b', 0}, // incomplete fourth CP - {8, 4, 6, 3, 'b', 0}, // incomplete fourth CP - {9, 4, 6, 3, 'b', 0}, // incomplete fourth CP - {7, 5, 6, 3, 'b', 0}, // incomplete fourth CP - {8, 5, 6, 3, 'b', 0}, // incomplete fourth CP - {9, 5, 6, 3, 'b', 0}, // incomplete fourth CP + // 2.1 The following test-cases raise doubt whether error or partial should + // be returned. For example, we have 4-byte sequence with valid leading + // byte. If we hide the last byte we need to return partial. But, if the + // second or third byte, which are visible to the call to codecvt, are + // malformed then error should be returned. // replace first trailing byte with ASCII byte, also incomplete at end {5, 5, 3, 2, 'z', 4}, - - // replace first trailing byte with invalid byte, also incomplete at end - {5, 5, 3, 2, '\xFF', 4}, - - // replace first trailing byte with ASCII byte, also incomplete at end {8, 5, 6, 3, 'z', 7}, {9, 5, 6, 3, 'z', 7}, // replace first trailing byte with invalid byte, also incomplete at end - {8, 5, 6, 3, '\xFF', 7}, - {9, 5, 6, 3, '\xFF', 7}, + {5, 5, 3, 2, 0xFF, 4}, + {8, 5, 6, 3, 0xFF, 7}, + {9, 5, 6, 3, 0xFF, 7}, // replace second trailing byte with ASCII byte, also incomplete at end {9, 5, 6, 3, 'z', 8}, // replace second trailing byte with invalid byte, also incomplete at end - {9, 5, 6, 3, '\xFF', 8}, + {9, 5, 6, 3, 0xFF, 8}, + + // 3. Surrogate CP. We modify the second byte (first trailing) of the 3-byte + // CP U+D700 + {6, 5, 3, 2, 0b10100000, 4}, // turn U+D700 into U+D800 + {6, 5, 3, 2, 0b10101100, 4}, // turn U+D700 into U+DB00 + {6, 5, 3, 2, 0b10110000, 4}, // turn U+D700 into U+DC00 + {6, 5, 3, 2, 0b10111100, 4}, // turn U+D700 into U+DF00 + + // 4. Overlong sequence. The CPs in the input are chosen such as modifying + // just the leading byte is enough to make them overlong, i.e. for the + // 3-byte and 4-byte CP the second byte (first trailing) has enough leading + // zeroes. + {3, 5, 1, 1, 0b11000000, 1}, // make the 2-byte CP overlong + {3, 5, 1, 1, 0b11000001, 1}, // make the 2-byte CP overlong + {6, 5, 3, 2, 0b11100000, 3}, // make the 3-byte CP overlong + {10, 5, 6, 3, 0b11110000, 6}, // make the 4-byte CP overlong + + // 5. CP above range + // turn U+10AAAA into U+14AAAA by changing its leading byte + {10, 5, 6, 3, 0b11110101, 6}, + // turn U+10AAAA into U+11AAAA by changing its 2nd byte + {10, 5, 6, 3, 0b10011010, 7}, + // Don't replace anything, show full 4-byte CP U+10AAAA + {10, 4, 6, 3, 'b', 0}, + {10, 5, 6, 3, 'b', 0}, + // Don't replace anything, show incomplete 4-byte CP at the end. It's still + // out of UCS2 range just by seeing the first byte. + {7, 4, 6, 3, 'b', 0}, // incomplete fourth CP + {8, 4, 6, 3, 'b', 0}, // incomplete fourth CP + {9, 4, 6, 3, 'b', 0}, // incomplete fourth CP + {7, 5, 6, 3, 'b', 0}, // incomplete fourth CP + {8, 5, 6, 3, 'b', 0}, // incomplete fourth CP + {9, 5, 6, 3, 'b', 0}, // incomplete fourth CP }; for (auto t : offsets) { - char in[array_size (valid_in)] = {}; - CharT out[array_size (exp) - 1] = {}; + InternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); VERIFY (t.expected_in_next <= t.in_size); VERIFY (t.expected_out_next <= t.out_size); - char_traits::copy (in, valid_in, array_size (valid_in)); + auto old_char = in[t.replace_pos]; in[t.replace_pos] = t.replace_char; auto state = mbstate_t{}; - auto in_next = (const char *) nullptr; - auto out_next = (CharT *) nullptr; + auto in_next = (const ExternT *) nullptr; + auto out_next = (InternT *) nullptr; auto res = codecvt_base::result (); res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -1057,48 +1226,51 @@ utf8_to_ucs2_in_error (const std::codecvt &cvt) VERIFY (res == cvt.error); VERIFY (in_next == in + t.expected_in_next); VERIFY (out_next == out + t.expected_out_next); - VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); if (t.expected_out_next < array_size (out)) VERIFY (out[t.expected_out_next] == 0); + + in[t.replace_pos] = old_char; } } -template +template void -utf8_to_ucs2_in (const std::codecvt &cvt) +utf8_to_ucs2_in (const std::codecvt &cvt) { utf8_to_ucs2_in_ok (cvt); utf8_to_ucs2_in_partial (cvt); utf8_to_ucs2_in_error (cvt); } -template +template void -ucs2_to_utf8_out_ok (const std::codecvt &cvt) +ucs2_to_utf8_out_ok (const std::codecvt &cvt) { using namespace std; // UTF-8 string of 1-byte CP, 2-byte CP and 3-byte CP - const char16_t in_literal[] = u"bш\uAAAA"; - const char exp[] = "bш\uAAAA"; - CharT in[array_size (in_literal)] = {}; - copy (begin (in_literal), end (in_literal), begin (in)); - - static_assert (array_size (in_literal) == 4, ""); - static_assert (array_size (exp) == 7, ""); - static_assert (array_size (in) == 4, ""); - VERIFY (char_traits::length (in_literal) == 3); - VERIFY (char_traits::length (exp) == 6); - VERIFY (char_traits::length (in) == 3); + const char16_t input[] = u"b\u0448\uAAAA"; + const unsigned char expected[] = "b\u0448\uAAAA"; + static_assert (array_size (input) == 4, ""); + static_assert (array_size (expected) == 7, ""); + + InternT in[array_size (input)]; + ExternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 3); + VERIFY (char_traits::length (exp) == 6); const test_offsets_ok offsets[] = {{0, 0}, {1, 1}, {2, 3}, {3, 6}}; for (auto t : offsets) { - char out[array_size (exp) - 1] = {}; + ExternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); auto state = mbstate_t{}; - auto in_next = (const CharT *) nullptr; - auto out_next = (char *) nullptr; + auto in_next = (const InternT *) nullptr; + auto out_next = (ExternT *) nullptr; auto res = codecvt_base::result (); res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -1106,29 +1278,29 @@ ucs2_to_utf8_out_ok (const std::codecvt &cvt) VERIFY (res == cvt.ok); VERIFY (in_next == in + t.in_size); VERIFY (out_next == out + t.out_size); - VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); if (t.out_size < array_size (out)) VERIFY (out[t.out_size] == 0); } } -template +template void -ucs2_to_utf8_out_partial (const std::codecvt &cvt) +ucs2_to_utf8_out_partial (const std::codecvt &cvt) { using namespace std; // UTF-8 string of 1-byte CP, 2-byte CP and 3-byte CP - const char16_t in_literal[] = u"bш\uAAAA"; - const char exp[] = "bш\uAAAA"; - CharT in[array_size (in_literal)] = {}; - copy (begin (in_literal), end (in_literal), begin (in)); - - static_assert (array_size (in_literal) == 4, ""); - static_assert (array_size (exp) == 7, ""); - static_assert (array_size (in) == 4, ""); - VERIFY (char_traits::length (in_literal) == 3); - VERIFY (char_traits::length (exp) == 6); - VERIFY (char_traits::length (in) == 3); + const char16_t input[] = u"b\u0448\uAAAA"; + const unsigned char expected[] = "b\u0448\uAAAA"; + static_assert (array_size (input) == 4, ""); + static_assert (array_size (expected) == 7, ""); + + InternT in[array_size (input)]; + ExternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 3); + VERIFY (char_traits::length (exp) == 6); const test_offsets_partial offsets[] = { {1, 0, 0, 0}, // no space for first CP @@ -1142,14 +1314,14 @@ ucs2_to_utf8_out_partial (const std::codecvt &cvt) }; for (auto t : offsets) { - char out[array_size (exp) - 1] = {}; + ExternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); VERIFY (t.expected_in_next <= t.in_size); VERIFY (t.expected_out_next <= t.out_size); auto state = mbstate_t{}; - auto in_next = (const CharT *) nullptr; - auto out_next = (char *) nullptr; + auto in_next = (const InternT *) nullptr; + auto out_next = (ExternT *) nullptr; auto res = codecvt_base::result (); res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -1157,43 +1329,45 @@ ucs2_to_utf8_out_partial (const std::codecvt &cvt) VERIFY (res == cvt.partial); VERIFY (in_next == in + t.expected_in_next); VERIFY (out_next == out + t.expected_out_next); - VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); if (t.expected_out_next < array_size (out)) VERIFY (out[t.expected_out_next] == 0); } } -template +template void -ucs2_to_utf8_out_error (const std::codecvt &cvt) +ucs2_to_utf8_out_error (const std::codecvt &cvt) { using namespace std; - const char16_t valid_in[] = u"bш\uAAAA\U0010AAAA"; - const char exp[] = "bш\uAAAA\U0010AAAA"; - - static_assert (array_size (valid_in) == 6, ""); - static_assert (array_size (exp) == 11, ""); - VERIFY (char_traits::length (valid_in) == 5); - VERIFY (char_traits::length (exp) == 10); - - test_offsets_error offsets[] = { - {5, 10, 0, 0, 0xD800, 0}, - {5, 10, 0, 0, 0xDBFF, 0}, - {5, 10, 0, 0, 0xDC00, 0}, - {5, 10, 0, 0, 0xDFFF, 0}, - - {5, 10, 1, 1, 0xD800, 1}, - {5, 10, 1, 1, 0xDBFF, 1}, - {5, 10, 1, 1, 0xDC00, 1}, - {5, 10, 1, 1, 0xDFFF, 1}, - - {5, 10, 2, 3, 0xD800, 2}, - {5, 10, 2, 3, 0xDBFF, 2}, - {5, 10, 2, 3, 0xDC00, 2}, - {5, 10, 2, 3, 0xDFFF, 2}, - - // dont replace anything, just show the surrogate pair - {5, 10, 3, 6, u'b', 0}, + const char16_t input[] = u"b\u0448\uAAAA\U0010AAAA"; + const unsigned char expected[] = "b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 6, ""); + static_assert (array_size (expected) == 11, ""); + + InternT in[array_size (input)]; + ExternT exp[array_size (expected)]; + copy (begin (input), end (input), begin (in)); + copy (begin (expected), end (expected), begin (exp)); + VERIFY (char_traits::length (in) == 5); + VERIFY (char_traits::length (exp) == 10); + + test_offsets_error offsets[] = { + {3, 6, 0, 0, 0xD800, 0}, + {3, 6, 0, 0, 0xDBFF, 0}, + {3, 6, 0, 0, 0xDC00, 0}, + {3, 6, 0, 0, 0xDFFF, 0}, + + {3, 6, 1, 1, 0xD800, 1}, + {3, 6, 1, 1, 0xDBFF, 1}, + {3, 6, 1, 1, 0xDC00, 1}, + {3, 6, 1, 1, 0xDFFF, 1}, + + {3, 6, 2, 3, 0xD800, 2}, + {3, 6, 2, 3, 0xDBFF, 2}, + {3, 6, 2, 3, 0xDC00, 2}, + {3, 6, 2, 3, 0xDFFF, 2}, // make the leading surrogate a trailing one {5, 10, 3, 6, 0xDC00, 3}, @@ -1206,6 +1380,9 @@ ucs2_to_utf8_out_error (const std::codecvt &cvt) // make the trailing surrogate a BMP char {5, 10, 3, 6, u'z', 4}, + // don't replace anything in the test cases bellow, just show the surrogate + // pair (fourth CP) fully or partially + {5, 10, 3, 6, u'b', 0}, {5, 7, 3, 6, u'b', 0}, // no space for fourth CP {5, 8, 3, 6, u'b', 0}, // no space for fourth CP {5, 9, 3, 6, u'b', 0}, // no space for fourth CP @@ -1214,23 +1391,21 @@ ucs2_to_utf8_out_error (const std::codecvt &cvt) {4, 7, 3, 6, u'b', 0}, // incomplete fourth CP, and no space for it {4, 8, 3, 6, u'b', 0}, // incomplete fourth CP, and no space for it {4, 9, 3, 6, u'b', 0}, // incomplete fourth CP, and no space for it - }; for (auto t : offsets) { - CharT in[array_size (valid_in)] = {}; - char out[array_size (exp) - 1] = {}; + ExternT out[array_size (exp) - 1] = {}; VERIFY (t.in_size <= array_size (in)); VERIFY (t.out_size <= array_size (out)); VERIFY (t.expected_in_next <= t.in_size); VERIFY (t.expected_out_next <= t.out_size); - copy (begin (valid_in), end (valid_in), begin (in)); + auto old_char = in[t.replace_pos]; in[t.replace_pos] = t.replace_char; auto state = mbstate_t{}; - auto in_next = (const CharT *) nullptr; - auto out_next = (char *) nullptr; + auto in_next = (const InternT *) nullptr; + auto out_next = (ExternT *) nullptr; auto res = codecvt_base::result (); res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, @@ -1238,25 +1413,793 @@ ucs2_to_utf8_out_error (const std::codecvt &cvt) VERIFY (res == cvt.error); VERIFY (in_next == in + t.expected_in_next); VERIFY (out_next == out + t.expected_out_next); - VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); if (t.expected_out_next < array_size (out)) VERIFY (out[t.expected_out_next] == 0); + + in[t.replace_pos] = old_char; } } -template +template void -ucs2_to_utf8_out (const std::codecvt &cvt) +ucs2_to_utf8_out (const std::codecvt &cvt) { ucs2_to_utf8_out_ok (cvt); ucs2_to_utf8_out_partial (cvt); ucs2_to_utf8_out_error (cvt); } -template +template void -test_utf8_ucs2_cvts (const std::codecvt &cvt) +test_utf8_ucs2_cvt (const std::codecvt &cvt) { utf8_to_ucs2_in (cvt); ucs2_to_utf8_out (cvt); } + +enum utf16_endianess +{ + utf16_big_endian, + utf16_little_endian +}; + +template +Iter2 +utf16_to_bytes (Iter1 f, Iter1 l, Iter2 o, utf16_endianess e) +{ + if (e == utf16_big_endian) + for (; f != l; ++f) + { + *o++ = (*f >> 8) & 0xFF; + *o++ = *f & 0xFF; + } + else + for (; f != l; ++f) + { + *o++ = *f & 0xFF; + *o++ = (*f >> 8) & 0xFF; + } + return o; +} + +template +void +utf16_to_utf32_in_ok (const std::codecvt &cvt, + utf16_endianess endianess) +{ + using namespace std; + const char16_t input[] = u"b\u0448\uAAAA\U0010AAAA"; + const char32_t expected[] = U"b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 6, ""); + static_assert (array_size (expected) == 5, ""); + + char in[array_size (input) * 2]; + InternT exp[array_size (expected)]; + utf16_to_bytes (begin (input), end (input), begin (in), endianess); + copy (begin (expected), end (expected), begin (exp)); + + test_offsets_ok offsets[] = {{0, 0}, {2, 1}, {4, 2}, {6, 3}, {10, 4}}; + for (auto t : offsets) + { + InternT out[array_size (exp) - 1] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + auto state = mbstate_t{}; + auto in_next = (const char *) nullptr; + auto out_next = (InternT *) nullptr; + auto res = codecvt_base::result (); + + res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, + out_next); + VERIFY (res == cvt.ok); + VERIFY (in_next == in + t.in_size); + VERIFY (out_next == out + t.out_size); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + if (t.out_size < array_size (out)) + VERIFY (out[t.out_size] == 0); + } + + for (auto t : offsets) + { + InternT out[array_size (exp)] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + auto state = mbstate_t{}; + auto in_next = (const char *) nullptr; + auto out_next = (InternT *) nullptr; + auto res = codecvt_base::result (); + + res + = cvt.in (state, in, in + t.in_size, in_next, out, end (out), out_next); + VERIFY (res == cvt.ok); + VERIFY (in_next == in + t.in_size); + VERIFY (out_next == out + t.out_size); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + if (t.out_size < array_size (out)) + VERIFY (out[t.out_size] == 0); + } +} + +template +void +utf16_to_utf32_in_partial (const std::codecvt &cvt, + utf16_endianess endianess) +{ + using namespace std; + const char16_t input[] = u"b\u0448\uAAAA\U0010AAAA"; + const char32_t expected[] = U"b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 6, ""); + static_assert (array_size (expected) == 5, ""); + + char in[array_size (input) * 2]; + InternT exp[array_size (expected)]; + auto in_iter = begin (in); + utf16_to_bytes (begin (input), end (input), begin (in), endianess); + copy (begin (expected), end (expected), begin (exp)); + + test_offsets_partial offsets[] = { + {2, 0, 0, 0}, // no space for first CP + {1, 1, 0, 0}, // incomplete first CP + {1, 0, 0, 0}, // incomplete first CP, and no space for it + + {4, 1, 2, 1}, // no space for second CP + {3, 2, 2, 1}, // incomplete second CP + {3, 1, 2, 1}, // incomplete second CP, and no space for it + + {6, 2, 4, 2}, // no space for third CP + {5, 3, 4, 2}, // incomplete third CP + {5, 2, 4, 2}, // incomplete third CP, and no space for it + + {10, 3, 6, 3}, // no space for fourth CP + {7, 4, 6, 3}, // incomplete fourth CP + {8, 4, 6, 3}, // incomplete fourth CP + {9, 4, 6, 3}, // incomplete fourth CP + {7, 3, 6, 3}, // incomplete fourth CP, and no space for it + {8, 3, 6, 3}, // incomplete fourth CP, and no space for it + {9, 3, 6, 3}, // incomplete fourth CP, and no space for it + }; + + for (auto t : offsets) + { + InternT out[array_size (exp) - 1] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + VERIFY (t.expected_in_next <= t.in_size); + VERIFY (t.expected_out_next <= t.out_size); + auto state = mbstate_t{}; + auto in_next = (const char *) nullptr; + auto out_next = (InternT *) nullptr; + auto res = codecvt_base::result (); + + res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, + out_next); + VERIFY (res == cvt.partial); + VERIFY (in_next == in + t.expected_in_next); + VERIFY (out_next == out + t.expected_out_next); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); + if (t.expected_out_next < array_size (out)) + VERIFY (out[t.expected_out_next] == 0); + } +} + +template +void +utf16_to_utf32_in_error (const std::codecvt &cvt, + utf16_endianess endianess) +{ + using namespace std; + char16_t input[] = u"b\u0448\uAAAA\U0010AAAA"; + const char32_t expected[] = U"b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 6, ""); + static_assert (array_size (expected) == 5, ""); + + InternT exp[array_size (expected)]; + copy (begin (expected), end (expected), begin (exp)); + + // The only possible error in UTF-16 is unpaired surrogate code units. + // So we replace valid code points (scalar values) with lone surrogate CU. + test_offsets_error offsets[] = { + {10, 4, 0, 0, 0xD800, 0}, + {10, 4, 0, 0, 0xDBFF, 0}, + {10, 4, 0, 0, 0xDC00, 0}, + {10, 4, 0, 0, 0xDFFF, 0}, + + {10, 4, 2, 1, 0xD800, 1}, + {10, 4, 2, 1, 0xDBFF, 1}, + {10, 4, 2, 1, 0xDC00, 1}, + {10, 4, 2, 1, 0xDFFF, 1}, + + {10, 4, 4, 2, 0xD800, 2}, + {10, 4, 4, 2, 0xDBFF, 2}, + {10, 4, 4, 2, 0xDC00, 2}, + {10, 4, 4, 2, 0xDFFF, 2}, + + // make the leading surrogate a trailing one + {10, 4, 6, 3, 0xDC00, 3}, + {10, 4, 6, 3, 0xDFFF, 3}, + + // make the trailing surrogate a leading one + {10, 4, 6, 3, 0xD800, 4}, + {10, 4, 6, 3, 0xDBFF, 4}, + + // make the trailing surrogate a BMP char + {10, 4, 6, 3, u'z', 4}, + }; + + for (auto t : offsets) + { + char in[array_size (input) * 2]; + InternT out[array_size (exp) - 1] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + VERIFY (t.expected_in_next <= t.in_size); + VERIFY (t.expected_out_next <= t.out_size); + auto old_char = input[t.replace_pos]; + input[t.replace_pos] = t.replace_char; // replace in input, not in in + utf16_to_bytes (begin (input), end (input), begin (in), endianess); + + auto state = mbstate_t{}; + auto in_next = (const char *) nullptr; + auto out_next = (InternT *) nullptr; + auto res = codecvt_base::result (); + + res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, + out_next); + VERIFY (res == cvt.error); + VERIFY (in_next == in + t.expected_in_next); + VERIFY (out_next == out + t.expected_out_next); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); + if (t.expected_out_next < array_size (out)) + VERIFY (out[t.expected_out_next] == 0); + + input[t.replace_pos] = old_char; + } +} + +template +void +utf32_to_utf16_out_ok (const std::codecvt &cvt, + utf16_endianess endianess) +{ + using namespace std; + const char32_t input[] = U"b\u0448\uAAAA\U0010AAAA"; + const char16_t expected[] = u"b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 5, ""); + static_assert (array_size (expected) == 6, ""); + + InternT in[array_size (input)]; + char exp[array_size (expected) * 2]; + copy (begin (input), end (input), begin (in)); + utf16_to_bytes (begin (expected), end (expected), begin (exp), endianess); + + const test_offsets_ok offsets[] = {{0, 0}, {1, 2}, {2, 4}, {3, 6}, {4, 10}}; + for (auto t : offsets) + { + char out[array_size (exp) - 2] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + auto state = mbstate_t{}; + auto in_next = (const InternT *) nullptr; + auto out_next = (char *) nullptr; + auto res = codecvt_base::result (); + + res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, + out_next); + VERIFY (res == cvt.ok); + VERIFY (in_next == in + t.in_size); + VERIFY (out_next == out + t.out_size); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + if (t.out_size < array_size (out)) + VERIFY (out[t.out_size] == 0); + } +} + +template +void +utf32_to_utf16_out_partial (const std::codecvt &cvt, + utf16_endianess endianess) +{ + using namespace std; + const char32_t input[] = U"b\u0448\uAAAA\U0010AAAA"; + const char16_t expected[] = u"b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 5, ""); + static_assert (array_size (expected) == 6, ""); + + InternT in[array_size (input)]; + char exp[array_size (expected) * 2]; + copy (begin (input), end (input), begin (in)); + utf16_to_bytes (begin (expected), end (expected), begin (exp), endianess); + + const test_offsets_partial offsets[] = { + {1, 0, 0, 0}, // no space for first CP + {1, 1, 0, 0}, // no space for first CP + + {2, 2, 1, 2}, // no space for second CP + {2, 3, 1, 2}, // no space for second CP + + {3, 4, 2, 4}, // no space for third CP + {3, 5, 2, 4}, // no space for third CP + + {4, 6, 3, 6}, // no space for fourth CP + {4, 7, 3, 6}, // no space for fourth CP + {4, 8, 3, 6}, // no space for fourth CP + {4, 9, 3, 6}, // no space for fourth CP + }; + for (auto t : offsets) + { + char out[array_size (exp) - 2] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + VERIFY (t.expected_in_next <= t.in_size); + VERIFY (t.expected_out_next <= t.out_size); + auto state = mbstate_t{}; + auto in_next = (const InternT *) nullptr; + auto out_next = (char *) nullptr; + auto res = codecvt_base::result (); + + res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, + out_next); + VERIFY (res == cvt.partial); + VERIFY (in_next == in + t.expected_in_next); + VERIFY (out_next == out + t.expected_out_next); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + if (t.expected_out_next < array_size (out)) + VERIFY (out[t.expected_out_next] == 0); + } +} + +template +void +utf32_to_utf16_out_error (const std::codecvt &cvt, + utf16_endianess endianess) +{ + using namespace std; + const char32_t input[] = U"b\u0448\uAAAA\U0010AAAA"; + const char16_t expected[] = u"b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 5, ""); + static_assert (array_size (expected) == 6, ""); + + InternT in[array_size (input)]; + char exp[array_size (expected) * 2]; + copy (begin (input), end (input), begin (in)); + utf16_to_bytes (begin (expected), end (expected), begin (exp), endianess); + + test_offsets_error offsets[] = { + + // Surrogate CP + {4, 10, 0, 0, 0xD800, 0}, + {4, 10, 1, 2, 0xDBFF, 1}, + {4, 10, 2, 4, 0xDC00, 2}, + {4, 10, 3, 6, 0xDFFF, 3}, + + // CP out of range + {4, 10, 0, 0, 0x00110000, 0}, + {4, 10, 1, 2, 0x00110000, 1}, + {4, 10, 2, 4, 0x00110000, 2}, + {4, 10, 3, 6, 0x00110000, 3}}; + + for (auto t : offsets) + { + char out[array_size (exp) - 2] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + VERIFY (t.expected_in_next <= t.in_size); + VERIFY (t.expected_out_next <= t.out_size); + auto old_char = in[t.replace_pos]; + in[t.replace_pos] = t.replace_char; + + auto state = mbstate_t{}; + auto in_next = (const InternT *) nullptr; + auto out_next = (char *) nullptr; + auto res = codecvt_base::result (); + + res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, + out_next); + VERIFY (res == cvt.error); + VERIFY (in_next == in + t.expected_in_next); + VERIFY (out_next == out + t.expected_out_next); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + if (t.expected_out_next < array_size (out)) + VERIFY (out[t.expected_out_next] == 0); + + in[t.replace_pos] = old_char; + } +} + +template +void +test_utf16_utf32_cvt (const std::codecvt &cvt, + utf16_endianess endianess) +{ + utf16_to_utf32_in_ok (cvt, endianess); + utf16_to_utf32_in_partial (cvt, endianess); + utf16_to_utf32_in_error (cvt, endianess); + utf32_to_utf16_out_ok (cvt, endianess); + utf32_to_utf16_out_partial (cvt, endianess); + utf32_to_utf16_out_error (cvt, endianess); +} + +template +void +utf16_to_ucs2_in_ok (const std::codecvt &cvt, + utf16_endianess endianess) +{ + using namespace std; + const char16_t input[] = u"b\u0448\uAAAA"; + const char16_t expected[] = u"b\u0448\uAAAA"; + static_assert (array_size (input) == 4, ""); + static_assert (array_size (expected) == 4, ""); + + char in[array_size (input) * 2]; + InternT exp[array_size (expected)]; + utf16_to_bytes (begin (input), end (input), begin (in), endianess); + copy (begin (expected), end (expected), begin (exp)); + + test_offsets_ok offsets[] = {{0, 0}, {2, 1}, {4, 2}, {6, 3}}; + for (auto t : offsets) + { + InternT out[array_size (exp) - 1] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + auto state = mbstate_t{}; + auto in_next = (const char *) nullptr; + auto out_next = (InternT *) nullptr; + auto res = codecvt_base::result (); + + res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, + out_next); + VERIFY (res == cvt.ok); + VERIFY (in_next == in + t.in_size); + VERIFY (out_next == out + t.out_size); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + if (t.out_size < array_size (out)) + VERIFY (out[t.out_size] == 0); + } + + for (auto t : offsets) + { + InternT out[array_size (exp)] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + auto state = mbstate_t{}; + auto in_next = (const char *) nullptr; + auto out_next = (InternT *) nullptr; + auto res = codecvt_base::result (); + + res + = cvt.in (state, in, in + t.in_size, in_next, out, end (out), out_next); + VERIFY (res == cvt.ok); + VERIFY (in_next == in + t.in_size); + VERIFY (out_next == out + t.out_size); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + if (t.out_size < array_size (out)) + VERIFY (out[t.out_size] == 0); + } +} + +template +void +utf16_to_ucs2_in_partial (const std::codecvt &cvt, + utf16_endianess endianess) +{ + using namespace std; + const char16_t input[] = u"b\u0448\uAAAA"; + const char16_t expected[] = u"b\u0448\uAAAA"; + static_assert (array_size (input) == 4, ""); + static_assert (array_size (expected) == 4, ""); + + char in[array_size (input) * 2]; + InternT exp[array_size (expected)]; + auto in_iter = begin (in); + utf16_to_bytes (begin (input), end (input), begin (in), endianess); + copy (begin (expected), end (expected), begin (exp)); + + test_offsets_partial offsets[] = { + {2, 0, 0, 0}, // no space for first CP + {1, 1, 0, 0}, // incomplete first CP + {1, 0, 0, 0}, // incomplete first CP, and no space for it + + {4, 1, 2, 1}, // no space for second CP + {3, 2, 2, 1}, // incomplete second CP + {3, 1, 2, 1}, // incomplete second CP, and no space for it + + {6, 2, 4, 2}, // no space for third CP + {5, 3, 4, 2}, // incomplete third CP + {5, 2, 4, 2}, // incomplete third CP, and no space for it + }; + + for (auto t : offsets) + { + InternT out[array_size (exp) - 1] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + VERIFY (t.expected_in_next <= t.in_size); + VERIFY (t.expected_out_next <= t.out_size); + auto state = mbstate_t{}; + auto in_next = (const char *) nullptr; + auto out_next = (InternT *) nullptr; + auto res = codecvt_base::result (); + + res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, + out_next); + VERIFY (res == cvt.partial); + VERIFY (in_next == in + t.expected_in_next); + VERIFY (out_next == out + t.expected_out_next); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); + if (t.expected_out_next < array_size (out)) + VERIFY (out[t.expected_out_next] == 0); + } +} + +template +void +utf16_to_ucs2_in_error (const std::codecvt &cvt, + utf16_endianess endianess) +{ + using namespace std; + char16_t input[] = u"b\u0448\uAAAA\U0010AAAA"; + const char16_t expected[] = u"b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 6, ""); + static_assert (array_size (expected) == 6, ""); + + InternT exp[array_size (expected)]; + copy (begin (expected), end (expected), begin (exp)); + + // The only possible error in UTF-16 is unpaired surrogate code units. + // Additionally, because the target encoding is UCS-2, a proper pair of + // surrogates is also error. Simply, any surrogate CU is error. + test_offsets_error offsets[] = { + {6, 3, 0, 0, 0xD800, 0}, + {6, 3, 0, 0, 0xDBFF, 0}, + {6, 3, 0, 0, 0xDC00, 0}, + {6, 3, 0, 0, 0xDFFF, 0}, + + {6, 3, 2, 1, 0xD800, 1}, + {6, 3, 2, 1, 0xDBFF, 1}, + {6, 3, 2, 1, 0xDC00, 1}, + {6, 3, 2, 1, 0xDFFF, 1}, + + {6, 3, 4, 2, 0xD800, 2}, + {6, 3, 4, 2, 0xDBFF, 2}, + {6, 3, 4, 2, 0xDC00, 2}, + {6, 3, 4, 2, 0xDFFF, 2}, + + // make the leading surrogate a trailing one + {10, 5, 6, 3, 0xDC00, 3}, + {10, 5, 6, 3, 0xDFFF, 3}, + + // make the trailing surrogate a leading one + {10, 5, 6, 3, 0xD800, 4}, + {10, 5, 6, 3, 0xDBFF, 4}, + + // make the trailing surrogate a BMP char + {10, 5, 6, 3, u'z', 4}, + + // don't replace anything in the test cases bellow, just show the surrogate + // pair (fourth CP) fully or partially (just the first surrogate) + {10, 5, 6, 3, u'b', 0}, + {8, 5, 6, 3, u'b', 0}, + {9, 5, 6, 3, u'b', 0}, + + {10, 4, 6, 3, u'b', 0}, + {8, 4, 6, 3, u'b', 0}, + {9, 4, 6, 3, u'b', 0}, + }; + + for (auto t : offsets) + { + char in[array_size (input) * 2]; + InternT out[array_size (exp) - 1] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + VERIFY (t.expected_in_next <= t.in_size); + VERIFY (t.expected_out_next <= t.out_size); + auto old_char = input[t.replace_pos]; + input[t.replace_pos] = t.replace_char; // replace in input, not in in + utf16_to_bytes (begin (input), end (input), begin (in), endianess); + + auto state = mbstate_t{}; + auto in_next = (const char *) nullptr; + auto out_next = (InternT *) nullptr; + auto res = codecvt_base::result (); + + res = cvt.in (state, in, in + t.in_size, in_next, out, out + t.out_size, + out_next); + VERIFY (res == cvt.error); + VERIFY (in_next == in + t.expected_in_next); + VERIFY (out_next == out + t.expected_out_next); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) + == 0); + if (t.expected_out_next < array_size (out)) + VERIFY (out[t.expected_out_next] == 0); + + input[t.replace_pos] = old_char; + } +} + +template +void +ucs2_to_utf16_out_ok (const std::codecvt &cvt, + utf16_endianess endianess) +{ + using namespace std; + const char16_t input[] = u"b\u0448\uAAAA"; + const char16_t expected[] = u"b\u0448\uAAAA"; + static_assert (array_size (input) == 4, ""); + static_assert (array_size (expected) == 4, ""); + + InternT in[array_size (input)]; + char exp[array_size (expected) * 2]; + copy (begin (input), end (input), begin (in)); + utf16_to_bytes (begin (expected), end (expected), begin (exp), endianess); + + const test_offsets_ok offsets[] = {{0, 0}, {1, 2}, {2, 4}, {3, 6}}; + for (auto t : offsets) + { + char out[array_size (exp) - 2] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + auto state = mbstate_t{}; + auto in_next = (const InternT *) nullptr; + auto out_next = (char *) nullptr; + auto res = codecvt_base::result (); + + res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, + out_next); + VERIFY (res == cvt.ok); + VERIFY (in_next == in + t.in_size); + VERIFY (out_next == out + t.out_size); + VERIFY (char_traits::compare (out, exp, t.out_size) == 0); + if (t.out_size < array_size (out)) + VERIFY (out[t.out_size] == 0); + } +} + +template +void +ucs2_to_utf16_out_partial (const std::codecvt &cvt, + utf16_endianess endianess) +{ + using namespace std; + const char16_t input[] = u"b\u0448\uAAAA"; + const char16_t expected[] = u"b\u0448\uAAAA"; + static_assert (array_size (input) == 4, ""); + static_assert (array_size (expected) == 4, ""); + + InternT in[array_size (input)]; + char exp[array_size (expected) * 2]; + copy (begin (input), end (input), begin (in)); + utf16_to_bytes (begin (expected), end (expected), begin (exp), endianess); + + const test_offsets_partial offsets[] = { + {1, 0, 0, 0}, // no space for first CP + {1, 1, 0, 0}, // no space for first CP + + {2, 2, 1, 2}, // no space for second CP + {2, 3, 1, 2}, // no space for second CP + + {3, 4, 2, 4}, // no space for third CP + {3, 5, 2, 4}, // no space for third CP + }; + for (auto t : offsets) + { + char out[array_size (exp) - 2] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + VERIFY (t.expected_in_next <= t.in_size); + VERIFY (t.expected_out_next <= t.out_size); + auto state = mbstate_t{}; + auto in_next = (const InternT *) nullptr; + auto out_next = (char *) nullptr; + auto res = codecvt_base::result (); + + res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, + out_next); + VERIFY (res == cvt.partial); + VERIFY (in_next == in + t.expected_in_next); + VERIFY (out_next == out + t.expected_out_next); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + if (t.expected_out_next < array_size (out)) + VERIFY (out[t.expected_out_next] == 0); + } +} + +template +void +ucs2_to_utf16_out_error (const std::codecvt &cvt, + utf16_endianess endianess) +{ + using namespace std; + const char16_t input[] = u"b\u0448\uAAAA\U0010AAAA"; + const char16_t expected[] = u"b\u0448\uAAAA\U0010AAAA"; + static_assert (array_size (input) == 6, ""); + static_assert (array_size (expected) == 6, ""); + + InternT in[array_size (input)]; + char exp[array_size (expected) * 2]; + copy (begin (input), end (input), begin (in)); + utf16_to_bytes (begin (expected), end (expected), begin (exp), endianess); + + test_offsets_error offsets[] = { + {3, 6, 0, 0, 0xD800, 0}, + {3, 6, 0, 0, 0xDBFF, 0}, + {3, 6, 0, 0, 0xDC00, 0}, + {3, 6, 0, 0, 0xDFFF, 0}, + + {3, 6, 1, 2, 0xD800, 1}, + {3, 6, 1, 2, 0xDBFF, 1}, + {3, 6, 1, 2, 0xDC00, 1}, + {3, 6, 1, 2, 0xDFFF, 1}, + + {3, 6, 2, 4, 0xD800, 2}, + {3, 6, 2, 4, 0xDBFF, 2}, + {3, 6, 2, 4, 0xDC00, 2}, + {3, 6, 2, 4, 0xDFFF, 2}, + + // make the leading surrogate a trailing one + {5, 10, 3, 6, 0xDC00, 3}, + {5, 10, 3, 6, 0xDFFF, 3}, + + // make the trailing surrogate a leading one + {5, 10, 3, 6, 0xD800, 4}, + {5, 10, 3, 6, 0xDBFF, 4}, + + // make the trailing surrogate a BMP char + {5, 10, 3, 6, u'z', 4}, + + // don't replace anything in the test cases bellow, just show the surrogate + // pair (fourth CP) fully or partially (just the first surrogate) + {5, 10, 3, 6, u'b', 0}, + {5, 8, 3, 6, u'b', 0}, + {5, 9, 3, 6, u'b', 0}, + + {4, 10, 3, 6, u'b', 0}, + {4, 8, 3, 6, u'b', 0}, + {4, 9, 3, 6, u'b', 0}, + }; + + for (auto t : offsets) + { + char out[array_size (exp) - 2] = {}; + VERIFY (t.in_size <= array_size (in)); + VERIFY (t.out_size <= array_size (out)); + VERIFY (t.expected_in_next <= t.in_size); + VERIFY (t.expected_out_next <= t.out_size); + auto old_char = in[t.replace_pos]; + in[t.replace_pos] = t.replace_char; + + auto state = mbstate_t{}; + auto in_next = (const InternT *) nullptr; + auto out_next = (char *) nullptr; + auto res = codecvt_base::result (); + + res = cvt.out (state, in, in + t.in_size, in_next, out, out + t.out_size, + out_next); + VERIFY (res == cvt.error); + VERIFY (in_next == in + t.expected_in_next); + VERIFY (out_next == out + t.expected_out_next); + VERIFY (char_traits::compare (out, exp, t.expected_out_next) == 0); + if (t.expected_out_next < array_size (out)) + VERIFY (out[t.expected_out_next] == 0); + + in[t.replace_pos] = old_char; + } +} + +template +void +test_utf16_ucs2_cvt (const std::codecvt &cvt, + utf16_endianess endianess) +{ + utf16_to_ucs2_in_ok (cvt, endianess); + utf16_to_ucs2_in_partial (cvt, endianess); + utf16_to_ucs2_in_error (cvt, endianess); + ucs2_to_utf16_out_ok (cvt, endianess); + ucs2_to_utf16_out_partial (cvt, endianess); + ucs2_to_utf16_out_error (cvt, endianess); +} diff --git a/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode_char8_t.cc b/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode_char8_t.cc new file mode 100644 index 00000000000..8ab5ba79f42 --- /dev/null +++ b/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode_char8_t.cc @@ -0,0 +1,53 @@ +// Copyright (C) 2020-2023 Free Software Foundation, Inc. +// +// This file is part of the GNU ISO C++ Library. This library is free +// software; you can redistribute it and/or modify it under the +// terms of the GNU General Public License as published by the +// Free Software Foundation; either version 3, or (at your option) +// any later version. + +// This library is distributed in the hope that it will be useful, +// but WITHOUT ANY WARRANTY; without even the implied warranty of +// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +// GNU General Public License for more details. + +// You should have received a copy of the GNU General Public License along +// with this library; see the file COPYING3. If not see +// . + +// { dg-do run { target c++11 } } +// { dg-require-cstdint "" } +// { dg-options "-fchar8_t" } + +#include "codecvt_unicode.h" + +using namespace std; + +void +test_utf8_utf32_codecvts () +{ + using codecvt_c32_c8 = codecvt; + auto &loc_c = locale::classic (); + VERIFY (has_facet (loc_c)); + + auto &cvt = use_facet (loc_c); + test_utf8_utf32_cvt (cvt); +} + +void +test_utf8_utf16_codecvts () +{ + using codecvt_c16_c8 = codecvt; + auto &loc_c = locale::classic (); + VERIFY (has_facet (loc_c)); + + auto &cvt = use_facet (loc_c); + test_utf8_utf16_cvt (cvt); +} + +int +main () +{ + test_utf8_utf32_codecvts (); + test_utf8_utf16_codecvts (); +} diff --git a/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode_wchar_t.cc b/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode_wchar_t.cc index 4fd1bfec63a..d6e5b20e89b 100644 --- a/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode_wchar_t.cc +++ b/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_unicode_wchar_t.cc @@ -28,7 +28,7 @@ test_utf8_utf32_codecvts () { #if __SIZEOF_WCHAR_T__ == 4 codecvt_utf8 cvt; - test_utf8_utf32_codecvts (cvt); + test_utf8_utf32_cvt (cvt); #endif } @@ -37,7 +37,7 @@ test_utf8_utf16_codecvts () { #if __SIZEOF_WCHAR_T__ >= 2 codecvt_utf8_utf16 cvt; - test_utf8_utf16_cvts (cvt); + test_utf8_utf16_cvt (cvt); #endif } @@ -46,7 +46,31 @@ test_utf8_ucs2_codecvts () { #if __SIZEOF_WCHAR_T__ == 2 codecvt_utf8 cvt; - test_utf8_ucs2_cvts (cvt); + test_utf8_ucs2_cvt (cvt); +#endif +} + +void +test_utf16_utf32_codecvts () +{ +#if __SIZEOF_WCHAR_T__ == 4 + codecvt_utf16 cvt3; + test_utf16_utf32_cvt (cvt3, utf16_big_endian); + + codecvt_utf16 cvt4; + test_utf16_utf32_cvt (cvt4, utf16_little_endian); +#endif +} + +void +test_utf16_ucs2_codecvts () +{ +#if __SIZEOF_WCHAR_T__ == 2 + codecvt_utf16 cvt3; + test_utf16_ucs2_cvt (cvt3, utf16_big_endian); + + codecvt_utf16 cvt4; + test_utf16_ucs2_cvt (cvt4, utf16_little_endian); #endif } @@ -56,4 +80,6 @@ main () test_utf8_utf32_codecvts (); test_utf8_utf16_codecvts (); test_utf8_ucs2_codecvts (); + test_utf16_utf32_codecvts (); + test_utf16_ucs2_codecvts (); } diff --git a/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_utf16/79980.cc b/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_utf16/79980.cc index 970d2bc1113..f3f802f7c98 100644 --- a/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_utf16/79980.cc +++ b/libstdc++-v3/testsuite/22_locale/codecvt/codecvt_utf16/79980.cc @@ -83,7 +83,7 @@ test06() const char src[] = "\0\x61\xAB\xCD"; Conv conv("to_bytes failed", u"from_bytes failed"); std::u16string result = conv.from_bytes(src, src+3); // incomplete character - VERIFY( result == u"from_bytes failed" ); + VERIFY( result == u"\u0061" ); VERIFY( conv.converted() == 2 ); }