1 //===----------------------------------------------------------------------===//
2 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
3 // See https://llvm.org/LICENSE.txt for license information.
4 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
5 //
6 //===----------------------------------------------------------------------===//
7 
8 // UNSUPPORTED: c++03, c++11, c++14, c++17
9 // UNSUPPORTED: libcpp-has-no-incomplete-format
10 
11 // This version runs the test when the platform has Unicode support.
12 // UNSUPPORTED: libcpp-has-no-unicode
13 
14 // <format>
15 
16 // Tests the Unicode width support of the standard format specifiers.
17 // It tests [format.string.std]/8 - 11:
18 // - Properly determining the estimated with of a unicode string.
19 // - Properly truncating to the wanted maximum width.
20 
21 // More specific extended grapheme cluster boundary rules are tested in
22 // test/libcxx/utilities/format/format.string/format.string.std/extended_grapheme_cluster.pass.cpp
23 // this test is based on test data provided by the Unicode Consortium.
24 
25 #include <format>
26 #include <cassert>
27 #include <vector>
28 
29 #include "make_string.h"
30 #include "test_macros.h"
31 #include "string_literal.h"
32 
33 #ifndef TEST_HAS_NO_LOCALIZATION
34 #  include <iostream>
35 #  include <type_traits>
36 #endif
37 
38 #define SV(S) MAKE_STRING_VIEW(CharT, S)
39 
40 auto check = []<string_literal fmt, class CharT, class... Args>(
41     std::basic_string_view<CharT> expected, const Args&... args) constexpr {
42   std::basic_string<CharT> out = std::format(fmt.template sv<CharT>(), args...);
43 #ifndef TEST_HAS_NO_LOCALIZATION
44   if constexpr (std::same_as<CharT, char>)
45     if (out != expected)
46       std::cerr << "\nFormat string   " << fmt.template sv<char>() << "\nExpected output " << expected
47                 << "\nActual output   " << out << '\n';
48 #endif
49   assert(out == expected);
50 };
51 
52 template <class CharT>
test_single_code_point_fill()53 static void test_single_code_point_fill() {
54   //*** 1-byte code points ***
55   check.template operator()<"{:*^3}">(SV("* *"), SV(" "));
56   check.template operator()<"{:*^3}">(SV("*~*"), SV("~"));
57 
58   //*** 2-byte code points ***
59   check.template operator()<"{:*^3}">(SV("*\u00a1*"), SV("\u00a1")); // INVERTED EXCLAMATION MARK
60   check.template operator()<"{:*^3}">(SV("*\u07ff*"), SV("\u07ff")); // NKO TAMAN SIGN
61 
62   //*** 3-byte code points ***
63   check.template operator()<"{:*^3}">(SV("*\u0800*"), SV("\u0800")); // SAMARITAN LETTER ALAF
64   check.template operator()<"{:*^3}">(SV("*\ufffd*"), SV("\ufffd")); // REPLACEMENT CHARACTER
65 
66   // 2 column ranges
67   check.template operator()<"{:*^4}">(SV("*\u1100*"), SV("\u1100")); // HANGUL CHOSEONG KIYEOK
68   check.template operator()<"{:*^4}">(SV("*\u115f*"), SV("\u115f")); // HANGUL CHOSEONG FILLER
69 
70   check.template operator()<"{:*^4}">(SV("*\u2329*"), SV("\u2329")); // LEFT-POINTING ANGLE BRACKET
71   check.template operator()<"{:*^4}">(SV("*\u232a*"), SV("\u232a")); // RIGHT-POINTING ANGLE BRACKET
72 
73   check.template operator()<"{:*^4}">(SV("*\u2e80*"), SV("\u2e80")); // CJK RADICAL REPEAT
74   check.template operator()<"{:*^4}">(SV("*\u303e*"), SV("\u303e")); // IDEOGRAPHIC VARIATION INDICATOR
75 
76   check.template operator()<"{:*^4}">(SV("*\u3040*"), SV("\u3040")); // U+3041 HIRAGANA LETTER SMALL A
77   check.template operator()<"{:*^4}">(SV("*\ua4cf*"), SV("\ua4cf")); // U+A4D0 LISU LETTER BA
78 
79   check.template operator()<"{:*^4}">(SV("*\uac00*"), SV("\uac00")); // <Hangul Syllable, First>
80   check.template operator()<"{:*^4}">(SV("*\ud7a3*"), SV("\ud7a3")); // Hangul Syllable Hih
81 
82   check.template operator()<"{:*^4}">(SV("*\uf900*"), SV("\uf900")); // CJK COMPATIBILITY IDEOGRAPH-F900
83   check.template operator()<"{:*^4}">(SV("*\ufaff*"), SV("\ufaff")); // U+FB00 LATIN SMALL LIGATURE FF
84 
85   check.template operator()<"{:*^4}">(SV("*\ufe10*"), SV("\ufe10")); // PRESENTATION FORM FOR VERTICAL COMMA
86   check.template
87   operator()<"{:*^4}">(SV("*\ufe19*"), SV("\ufe19")); // PRESENTATION FORM FOR VERTICAL HORIZONTAL ELLIPSIS
88 
89   check.template operator()<"{:*^4}">(SV("*\ufe30*"), SV("\ufe30")); // PRESENTATION FORM FOR VERTICAL TWO DOT LEADER
90   check.template operator()<"{:*^4}">(SV("*\ufe6f*"), SV("\ufe6f")); // U+FE70 ARABIC FATHATAN ISOLATED FORM
91 
92   check.template operator()<"{:*^4}">(SV("*\uff00*"), SV("\uff00")); // U+FF01 FULLWIDTH EXCLAMATION MARK
93   check.template operator()<"{:*^4}">(SV("*\uff60*"), SV("\uff60")); // FULLWIDTH RIGHT WHITE PARENTHESIS
94 
95   check.template operator()<"{:*^4}">(SV("*\uffe0*"), SV("\uffe0")); // FULLWIDTH CENT SIGN
96   check.template operator()<"{:*^4}">(SV("*\uffe6*"), SV("\uffe6")); // FULLWIDTH WON SIGN
97 
98   //*** 4-byte code points ***
99   check.template operator()<"{:*^3}">(SV("*\U00010000*"), SV("\U00010000")); // LINEAR B SYLLABLE B008 A
100   check.template operator()<"{:*^3}">(SV("*\U0010FFFF*"), SV("\U0010FFFF")); // Undefined Character
101 
102   // 2 column ranges
103   check.template operator()<"{:*^4}">(SV("*\U0001f300*"), SV("\U0001f300")); // CYCLONE
104   check.template operator()<"{:*^4}">(SV("*\U0001f64f*"), SV("\U0001f64f")); // PERSON WITH FOLDED HANDS
105   check.template operator()<"{:*^4}">(SV("*\U0001f900*"), SV("\U0001f900")); // CIRCLED CROSS FORMEE WITH FOUR DOTS
106   check.template operator()<"{:*^4}">(SV("*\U0001f9ff*"), SV("\U0001f9ff")); // NAZAR AMULET
107   check.template operator()<"{:*^4}">(SV("*\U00020000*"), SV("\U00020000")); // <CJK Ideograph Extension B, First>
108   check.template operator()<"{:*^4}">(SV("*\U0002fffd*"), SV("\U0002fffd")); // Undefined Character
109   check.template operator()<"{:*^4}">(SV("*\U00030000*"), SV("\U00030000")); // <CJK Ideograph Extension G, First>
110   check.template operator()<"{:*^4}">(SV("*\U0003fffd*"), SV("\U0003fffd")); // Undefined Character
111 }
112 
113 // One column output is unaffected.
114 // Two column output is removed, thus the result is only the fill character.
115 template <class CharT>
test_single_code_point_truncate()116 static void test_single_code_point_truncate() {
117   //*** 1-byte code points ***
118   check.template operator()<"{:*^3.1}">(SV("* *"), SV(" "));
119   check.template operator()<"{:*^3.1}">(SV("*~*"), SV("~"));
120 
121   //*** 2-byte code points ***
122   check.template operator()<"{:*^3.1}">(SV("*\u00a1*"), SV("\u00a1")); // INVERTED EXCLAMATION MARK
123   check.template operator()<"{:*^3.1}">(SV("*\u07ff*"), SV("\u07ff")); // NKO TAMAN SIGN
124 
125   //*** 3.1-byte code points ***
126   check.template operator()<"{:*^3.1}">(SV("*\u0800*"), SV("\u0800")); // SAMARITAN LETTER ALAF
127   check.template operator()<"{:*^3.1}">(SV("*\ufffd*"), SV("\ufffd")); // REPLACEMENT CHARACTER
128 
129   // 2 column ranges
130   check.template operator()<"{:*^3.1}">(SV("***"), SV("\u1100")); // HANGUL CHOSEONG KIYEOK
131   check.template operator()<"{:*^3.1}">(SV("***"), SV("\u115f")); // HANGUL CHOSEONG FILLER
132 
133   check.template operator()<"{:*^3.1}">(SV("***"), SV("\u2329")); // LEFT-POINTING ANGLE BRACKET
134   check.template operator()<"{:*^3.1}">(SV("***"), SV("\u232a")); // RIGHT-POINTING ANGLE BRACKET
135 
136   check.template operator()<"{:*^3.1}">(SV("***"), SV("\u2e80")); // CJK RADICAL REPEAT
137   check.template operator()<"{:*^3.1}">(SV("***"), SV("\u303e")); // IDEOGRAPHIC VARIATION INDICATOR
138 
139   check.template operator()<"{:*^3.1}">(SV("***"), SV("\u3040")); // U+3041 HIRAGANA LETTER SMALL A
140   check.template operator()<"{:*^3.1}">(SV("***"), SV("\ua4cf")); // U+A4D0 LISU LETTER BA
141 
142   check.template operator()<"{:*^3.1}">(SV("***"), SV("\uac00")); // <Hangul Syllable, First>
143   check.template operator()<"{:*^3.1}">(SV("***"), SV("\ud7a3")); // Hangul Syllable Hih
144 
145   check.template operator()<"{:*^3.1}">(SV("***"), SV("\uf900")); // CJK COMPATIBILITY IDEOGRAPH-F900
146   check.template operator()<"{:*^3.1}">(SV("***"), SV("\ufaff")); // U+FB00 LATIN SMALL LIGATURE FF
147 
148   check.template operator()<"{:*^3.1}">(SV("***"), SV("\ufe10")); // PRESENTATION FORM FOR VERTICAL COMMA
149   check.template operator()<"{:*^3.1}">(SV("***"), SV("\ufe19")); // PRESENTATION FORM FOR VERTICAL HORIZONTAL ELLIPSIS
150 
151   check.template operator()<"{:*^3.1}">(SV("***"), SV("\ufe30")); // PRESENTATION FORM FOR VERTICAL TWO DOT LEADER
152   check.template operator()<"{:*^3.1}">(SV("***"), SV("\ufe6f")); // U+FE70 ARABIC FATHATAN ISOLATED FORM
153 
154   check.template operator()<"{:*^3.1}">(SV("***"), SV("\uff00")); // U+FF01 FULLWIDTH EXCLAMATION MARK
155   check.template operator()<"{:*^3.1}">(SV("***"), SV("\uff60")); // FULLWIDTH RIGHT WHITE PARENTHESIS
156 
157   check.template operator()<"{:*^3.1}">(SV("***"), SV("\uffe0")); // FULLWIDTH CENT SIGN
158   check.template operator()<"{:*^3.1}">(SV("***"), SV("\uffe6")); // FULLWIDTH WON SIGN
159 
160   //*** 3.1-byte code points ***
161   check.template operator()<"{:*^3.1}">(SV("*\U00010000*"), SV("\U00010000")); // LINEAR B SYLLABLE B008 A
162   check.template operator()<"{:*^3.1}">(SV("*\U0010FFFF*"), SV("\U0010FFFF")); // Undefined Character
163 
164   // 2 column ranges
165   check.template operator()<"{:*^3.1}">(SV("***"), SV("\U0001f300")); // CYCLONE
166   check.template operator()<"{:*^3.1}">(SV("***"), SV("\U0001f64f")); // PERSON WITH FOLDED HANDS
167   check.template operator()<"{:*^3.1}">(SV("***"), SV("\U0001f900")); // CIRCLED CROSS FORMEE WITH FOUR DOTS
168   check.template operator()<"{:*^3.1}">(SV("***"), SV("\U0001f9ff")); // NAZAR AMULET
169   check.template operator()<"{:*^3.1}">(SV("***"), SV("\U00020000")); // <CJK Ideograph Extension B, First>
170   check.template operator()<"{:*^3.1}">(SV("***"), SV("\U0002fffd")); // Undefined Character
171   check.template operator()<"{:*^3.1}">(SV("***"), SV("\U00030000")); // <CJK Ideograph Extension G, First>
172   check.template operator()<"{:*^3.1}">(SV("***"), SV("\U0003fffd")); // Undefined Character
173 }
174 
175 // The examples used in that paper.
176 template <class CharT>
test_P1868()177 static void test_P1868() {
178   // Fill
179   check.template operator()<"{:*^3}">(SV("*\u0041*"), SV("\u0041")); // { LATIN CAPITAL LETTER A }
180   check.template operator()<"{:*^3}">(SV("*\u00c1*"), SV("\u00c1")); // { LATIN CAPITAL LETTER A WITH ACUTE }
181   check.template operator()<"{:*^3}">(
182       SV("*\u0041\u0301*"),
183       SV("\u0041\u0301")); // { LATIN CAPITAL LETTER A } { COMBINING ACUTE ACCENT }
184   check.template operator()<"{:*^3}">(SV("*\u0132*"), SV("\u0132")); // { LATIN CAPITAL LIGATURE IJ }
185   check.template operator()<"{:*^3}">(SV("*\u0394*"), SV("\u0394")); // { GREEK CAPITAL LETTER DELTA }
186 
187   check.template operator()<"{:*^3}">(SV("*\u0429*"), SV("\u0429"));         // { CYRILLIC CAPITAL LETTER SHCHA }
188   check.template operator()<"{:*^3}">(SV("*\u05d0*"), SV("\u05d0"));         // { HEBREW LETTER ALEF }
189   check.template operator()<"{:*^3}">(SV("*\u0634*"), SV("\u0634"));         // { ARABIC LETTER SHEEN }
190   check.template operator()<"{:*^4}">(SV("*\u3009*"), SV("\u3009"));         // { RIGHT-POINTING ANGLE BRACKET }
191   check.template operator()<"{:*^4}">(SV("*\u754c*"), SV("\u754c"));         // { CJK Unified Ideograph-754C }
192   check.template operator()<"{:*^4}">(SV("*\U0001f921*"), SV("\U0001f921")); // { UNICORN FACE }
193   check.template operator()<"{:*^4}">(
194       SV("*\U0001f468\u200d\U0001F469\u200d\U0001F467\u200d\U0001F466*"),
195       SV("\U0001f468\u200d\U0001F469\u200d\U0001F467\u200d\U0001F466")); // { Family: Man, Woman, Girl, Boy }
196 
197   // Truncate to 1 column: 1 column grapheme clusters are kept together.
198   check.template operator()<"{:*^3.1}">(SV("*\u0041*"), SV("\u0041")); // { LATIN CAPITAL LETTER A }
199   check.template operator()<"{:*^3.1}">(SV("*\u00c1*"), SV("\u00c1")); // { LATIN CAPITAL LETTER A WITH ACUTE }
200   check.template operator()<"{:*^3.1}">(
201       SV("*\u0041\u0301*"),
202       SV("\u0041\u0301")); // { LATIN CAPITAL LETTER A } { COMBINING ACUTE ACCENT }
203   check.template operator()<"{:*^3.1}">(SV("*\u0132*"), SV("\u0132")); // { LATIN CAPITAL LIGATURE IJ }
204   check.template operator()<"{:*^3.1}">(SV("*\u0394*"), SV("\u0394")); // { GREEK CAPITAL LETTER DELTA }
205 
206   check.template operator()<"{:*^3.1}">(SV("*\u0429*"), SV("\u0429")); // { CYRILLIC CAPITAL LETTER SHCHA }
207   check.template operator()<"{:*^3.1}">(SV("*\u05d0*"), SV("\u05d0")); // { HEBREW LETTER ALEF }
208   check.template operator()<"{:*^3.1}">(SV("*\u0634*"), SV("\u0634")); // { ARABIC LETTER SHEEN }
209   check.template operator()<"{:*^3.1}">(SV("***"), SV("\u3009"));      // { RIGHT-POINTING ANGLE BRACKET }
210   check.template operator()<"{:*^3.1}">(SV("***"), SV("\u754c"));      // { CJK Unified Ideograph-754C }
211   check.template operator()<"{:*^3.1}">(SV("***"), SV("\U0001f921"));  // { UNICORN FACE }
212   check.template operator()<"{:*^3.1}">(
213       SV("***"),
214       SV("\U0001f468\u200d\U0001F469\u200d\U0001F467\u200d\U0001F466")); // { Family: Man, Woman, Girl, Boy }
215 
216   // Truncate to 2 column: 2 column grapheme clusters are kept together.
217   check.template operator()<"{:*^3.2}">(SV("*\u0041*"), SV("\u0041")); // { LATIN CAPITAL LETTER A }
218   check.template operator()<"{:*^3.2}">(SV("*\u00c1*"), SV("\u00c1")); // { LATIN CAPITAL LETTER A WITH ACUTE }
219   check.template operator()<"{:*^3.2}">(
220       SV("*\u0041\u0301*"),
221       SV("\u0041\u0301")); // { LATIN CAPITAL LETTER A } { COMBINING ACUTE ACCENT }
222   check.template operator()<"{:*^3.2}">(SV("*\u0132*"), SV("\u0132")); // { LATIN CAPITAL LIGATURE IJ }
223   check.template operator()<"{:*^3.2}">(SV("*\u0394*"), SV("\u0394")); // { GREEK CAPITAL LETTER DELTA }
224 
225   check.template operator()<"{:*^3.2}">(SV("*\u0429*"), SV("\u0429"));         // { CYRILLIC CAPITAL LETTER SHCHA }
226   check.template operator()<"{:*^3.2}">(SV("*\u05d0*"), SV("\u05d0"));         // { HEBREW LETTER ALEF }
227   check.template operator()<"{:*^3.2}">(SV("*\u0634*"), SV("\u0634"));         // { ARABIC LETTER SHEEN }
228   check.template operator()<"{:*^4.2}">(SV("*\u3009*"), SV("\u3009"));         // { RIGHT-POINTING ANGLE BRACKET }
229   check.template operator()<"{:*^4.2}">(SV("*\u754c*"), SV("\u754c"));         // { CJK Unified Ideograph-754C }
230   check.template operator()<"{:*^4.2}">(SV("*\U0001f921*"), SV("\U0001f921")); // { UNICORN FACE }
231   check.template operator()<"{:*^4.2}">(
232       SV("*\U0001f468\u200d\U0001F469\u200d\U0001F467\u200d\U0001F466*"),
233       SV("\U0001f468\u200d\U0001F469\u200d\U0001F467\u200d\U0001F466")); // { Family: Man, Woman, Girl, Boy }
234 }
235 
236 #ifdef _LIBCPP_VERSION
237 // Tests the libc++ specific behaviour for malformed UTF-sequences. The
238 // Standard doesn't specify how to handle this.
239 template <class CharT>
test_malformed_code_point()240 static void test_malformed_code_point() {
241   if constexpr (sizeof(CharT) == 1) {
242     // Malformed at end.
243     check.template operator()<"{:*^7}">(SV("*ZZZZ\x8f*"), SV("ZZZZ\x8f"));
244     check.template operator()<"{:*^7}">(SV("*ZZZZ\xcf*"), SV("ZZZZ\xcf"));
245     check.template operator()<"{:*^7}">(SV("*ZZZZ\xef*"), SV("ZZZZ\xef"));
246     check.template operator()<"{:*^7}">(SV("*ZZZZ\xff*"), SV("ZZZZ\xff"));
247 
248     // Malformed in middle, no continuation
249     check.template operator()<"{:*^8}">(SV("*ZZZZ\x8fZ*"), SV("ZZZZ\x8fZ"));
250     check.template operator()<"{:*^8}">(SV("*ZZZZ\xcfZ*"), SV("ZZZZ\xcfZ"));
251     check.template operator()<"{:*^8}">(SV("*ZZZZ\xefZ*"), SV("ZZZZ\xefZ"));
252     check.template operator()<"{:*^8}">(SV("*ZZZZ\xffZ*"), SV("ZZZZ\xffZ"));
253 
254     check.template operator()<"{:*^9}">(SV("*ZZZZ\x8fZZ*"), SV("ZZZZ\x8fZZ"));
255     check.template operator()<"{:*^9}">(SV("*ZZZZ\xcfZZ*"), SV("ZZZZ\xcfZZ"));
256     check.template operator()<"{:*^9}">(SV("*ZZZZ\xefZZ*"), SV("ZZZZ\xefZZ"));
257     check.template operator()<"{:*^9}">(SV("*ZZZZ\xffZZ*"), SV("ZZZZ\xffZZ"));
258 
259     check.template operator()<"{:*^10}">(SV("*ZZZZ\x8fZZZ*"), SV("ZZZZ\x8fZZZ"));
260     check.template operator()<"{:*^10}">(SV("*ZZZZ\xcfZZZ*"), SV("ZZZZ\xcfZZZ"));
261     check.template operator()<"{:*^10}">(SV("*ZZZZ\xefZZZ*"), SV("ZZZZ\xefZZZ"));
262     check.template operator()<"{:*^10}">(SV("*ZZZZ\xffZZZ*"), SV("ZZZZ\xffZZZ"));
263 
264     check.template operator()<"{:*^11}">(SV("*ZZZZ\x8fZZZZ*"), SV("ZZZZ\x8fZZZZ"));
265     check.template operator()<"{:*^11}">(SV("*ZZZZ\xcfZZZZ*"), SV("ZZZZ\xcfZZZZ"));
266     check.template operator()<"{:*^11}">(SV("*ZZZZ\xefZZZZ*"), SV("ZZZZ\xefZZZZ"));
267     check.template operator()<"{:*^11}">(SV("*ZZZZ\xffZZZZ*"), SV("ZZZZ\xffZZZZ"));
268 
269     // Premature end.
270     check.template operator()<"{:*^8}">(SV("*ZZZZ\xef\xf5*"), SV("ZZZZ\xef\xf5"));
271     check.template operator()<"{:*^12}">(SV("*ZZZZ\xef\xf5ZZZZ*"), SV("ZZZZ\xef\xf5ZZZZ"));
272     check.template operator()<"{:*^9}">(SV("*ZZZZ\xff\xf5\xf5*"), SV("ZZZZ\xff\xf5\xf5"));
273     check.template operator()<"{:*^13}">(SV("*ZZZZ\xff\xf5\xf5ZZZZ*"), SV("ZZZZ\xff\xf5\xf5ZZZZ"));
274 
275   } else if constexpr (sizeof(CharT) == 2) {
276     // TODO FMT Add these tests.
277   }
278   // UTF-32 doesn't combine characters, thus no corruption tests.
279 }
280 #endif
281 
282 template <class CharT>
test()283 static void test() {
284   test_single_code_point_fill<CharT>();
285   test_single_code_point_truncate<CharT>();
286   test_P1868<CharT>();
287 
288 #ifdef _LIBCPP_VERSION
289   test_malformed_code_point<CharT>();
290 #endif
291 }
292 
main(int,char **)293 int main(int, char**) {
294   test<char>();
295 
296 #ifndef TEST_HAS_NO_WIDE_CHARACTERS
297   test<wchar_t>();
298 #endif
299 
300   return 0;
301 }
302