[libcxx-commits] [clang] [libcxx] [llvm] [Clang] Add warnings when mixing different charN_t types (PR #138708)

[Tom Honermann via libcxx-commits]

================
@@ -0,0 +1,155 @@
+// RUN: %clang_cc1 -verify -fsyntax-only -std=c++20 -Wconversion %s
+
+void c8(char8_t);
+void c16(char16_t);
+void c32(char32_t);
+
+void test(char8_t u8, char16_t u16, char32_t u32) {
+    c8(u8);
+    c8(u16); // expected-warning {{implicit conversion from 'char16_t' to 'char8_t' may lose precision and change the meaning of the represented code unit}}
+    c8(u32); // expected-warning {{implicit conversion from 'char32_t' to 'char8_t' may lose precision and change the meaning of the represented code unit}}
+
+    c16(u8);  // expected-warning {{implicit conversion from 'char8_t' to 'char16_t' may change the meaning of the represented code unit}}
+    c16(u16);
+    c16(u32); // expected-warning {{implicit conversion from 'char32_t' to 'char16_t' may lose precision and change the meaning of the represented code unit}}
+
+    c32(u8);  // expected-warning {{implicit conversion from 'char8_t' to 'char32_t' may change the meaning of the represented code unit}}
+    c32(u16); // expected-warning {{implicit conversion from 'char16_t' to 'char32_t' may change the meaning of the represented code unit}}
+    c32(u32);
+
+
+    c8(char32_t(0x7f));
+    c8(char32_t(0x80));   // expected-warning {{implicit conversion from 'char32_t' to 'char8_t' changes the meaning of the codepoint '<U+0080>'}}
+
+    c8(char16_t(0x7f));
+    c8(char16_t(0x80));   // expected-warning {{implicit conversion from 'char16_t' to 'char8_t' changes the meaning of the codepoint '<U+0080>'}}
+    c8(char16_t(0xD800)); // expected-warning {{implicit conversion from 'char16_t' to 'char8_t' changes the meaning of the code unit '<0xD800>'}}
+    c8(char16_t(0xE000)); // expected-warning {{implicit conversion from 'char16_t' to 'char8_t' changes the meaning of the codepoint '<U+E000>'}}
+
+
+    c16(char32_t(0x7f));
+    c16(char32_t(0x80));
+    c16(char32_t(0xD7FF));
+    c16(char32_t(0xD800)); // expected-warning {{implicit conversion from 'char32_t' to 'char16_t' changes the meaning of the code unit '<0xD800>'}}
+    c16(char32_t(0xE000));
+    c16(char32_t(U'🐉')); // expected-warning {{implicit conversion from 'char32_t' to 'char16_t' changes the meaning of the codepoint '🐉'}}
+
+
+    c32(char8_t(0x7f));
+    c32(char8_t(0x80)); // expected-warning {{implicit conversion from 'char8_t' to 'char32_t' changes the meaning of the code unit '<0x80>'}}
+    c32(char8_t(0xFF)); // expected-warning {{implicit conversion from 'char8_t' to 'char32_t' changes the meaning of the code unit '<0xFF>'}}
+
+
+    c32(char16_t(0x7f));
+    c32(char16_t(0x80));
+
+    c32(char16_t(0xD7FF));
+    c32(char16_t(0xD800)); // expected-warning {{implicit conversion from 'char16_t' to 'char32_t' changes the meaning of the code unit '<0xD800>'}}
+    c32(char16_t(0xDFFF)); // expected-warning {{implicit conversion from 'char16_t' to 'char32_t' changes the meaning of the code unit '<0xDFFF>'}}
+    c32(char16_t(0xE000));
+    c32(char16_t(u'☕'));
+
+    (void)static_cast<char32_t>(char8_t(0x80)); // sanity check: no explicit conversion;
+
+    using Char8 = char8_t;
+    Char8 c81 = u16; // expected-warning {{implicit conversion from 'char16_t' to 'Char8' (aka 'char8_t') may lose precision and change the meaning of the represented code unit}}
+
+    [[maybe_unused]] char c = u16; // expected-warning {{implicit conversion loses integer precision: 'char16_t' to 'char'}}
+
+    // FIXME: We should apply the same logic to wchar
+    [[maybe_unused]] wchar_t wc = u16;
+    [[maybe_unused]] wchar_t wc2 = u8;
+}
+
+void test_comp(char8_t u8, char16_t u16, char32_t u32) {
+    (void)(u8 == u8' ');
+    (void)(u8 == u' '); // expected-warning{{comparing values of different Unicode code unit types 'char8_t' and 'char16_t' may compare different codepoints}}
+    (void)(u8 == U' '); // expected-warning{{comparing values of different Unicode code unit types 'char8_t' and 'char32_t' may compare different codepoints}}
----------------
tahonermann wrote:

I think these warnings should be suppressed since the RHS may be constant evaluated and designates a character that is representable with a single code unit in the type of the LHS.

https://github.com/llvm/llvm-project/pull/138708