[libcxx] [test] Fix the monetary locale curr_symbol test on Windows, Apple and FreeBSD

[mstorsjo]

Summary

International currency symbols (like USD, EUR) are returned with a
trailing space, like "USD ", on previously supported Unix platforms.
On Windows, the locales return them without a trailing space.

Also adjust the test for expecting a different unicode sequence for
the national currency symbol for ru_RU.UTF-8 and zh_CN.UTF-8.

[• Quuxplusone]

Seems reasonable to me. Some pre-existing nitpicks.

libcxx/test/std/localization/locale.categories/category.monetary/locale.moneypunct.byname/curr_symbol.pass.cpp

9

Any chance this passes on Apple now? If not, do you mind adding a one-line comment here since presumably you've already dug into it a little bit?

162

Pre-existing: it sure would be nice if this could be changed to L" \u0440\u0443\u0431" for clarity. I'm like 50% sure those mean the same thing. :)

178–182

Could these be "\u00A5" and "\uFFE5" for clarity?

191–195

I think these should use \u not \x, too.

[mstorsjo]

libcxx/test/std/localization/locale.categories/category.monetary/locale.moneypunct.byname/curr_symbol.pass.cpp

9

Not quite yet (as CI would error out then), but I can have a look if it’s easily fixable at the same time - and if not I can add a comment.

162

Oh, yes, the inconsistency annoyed me too :-)

178–182

I think they could, but for UTF-8 strings, I’d actually prefer to stick consistently to a series of bytes, instead of \u escapes, as that makes the exact content a bit clearer (especially for cases when debugging the tests). But I guess I could add a comment with what Unicode sequence it corresponds to.

[mstorsjo]

Adjusted the test so that it passes on Windows, current macOS and FreeBSD (to avoid leaving any unused #else clause that isn't used by any target at all).

Also made the escape sequences consistent; use \u0000 form for all wchar strings, and \x00 form for all UTF-8 strings (with comments explaining what unicode chars it corresponds to).

[mstorsjo]

Rerun CI after unrelated failures on main are fixed.

[mstorsjo]

Anyone up for giving the full stamp of approval here? It has got a personal approval so far only.

[Mordante]

LGTM!

libcxx/test/std/localization/locale.categories/category.monetary/locale.moneypunct.byname/curr_symbol.pass.cpp

 //===----------------------------------------------------------------------===//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 // See https://llvm.org/LICENSE.txt for license information.
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
 //
 // XFAIL: darwin
 //
 // NetBSD does not support LC_MONETARY at the moment
 // XFAIL: netbsd
 
-// XFAIL: LIBCXX-WINDOWS-FIXME
-
 // REQUIRES: locale.en_US.UTF-8
 // REQUIRES: locale.fr_FR.UTF-8
 // REQUIRES: locale.ru_RU.UTF-8
 // REQUIRES: locale.zh_CN.UTF-8
 
 // <locale>
 
 // class moneypunct_byname<charT, International>
#ifndef TEST_HAS_NO_WIDE_CHARACTERS
         assert(f.curr_symbol() == std::wstring());
     }
     {
         Fwt f("C", 1);
         assert(f.curr_symbol() == std::wstring());
     }
 #endif
 
+#ifdef _WIN32
+    std::string curr_space = "";
+#else
+    std::string curr_space = " ";
+#endif
     {
         Fnf f(LOCALE_en_US_UTF_8, 1);
         assert(f.curr_symbol() == "$");
     }
     {
         Fnt f(LOCALE_en_US_UTF_8, 1);
-        assert(f.curr_symbol() == "USD ");
+        assert(f.curr_symbol() == "USD" + curr_space);
     }
 #ifndef TEST_HAS_NO_WIDE_CHARACTERS
+#ifdef _WIN32
+    std::wstring w_curr_space = L"";
+#else
+    std::wstring w_curr_space = L" ";
+#endif
     {
         Fwf f(LOCALE_en_US_UTF_8, 1);
         assert(f.curr_symbol() == L"$");
     }
     {
         Fwt f(LOCALE_en_US_UTF_8, 1);
-        assert(f.curr_symbol() == L"USD ");
+        assert(f.curr_symbol() == L"USD" + w_curr_space);
     }
 #endif
 
     {
         Fnf f(LOCALE_fr_FR_UTF_8, 1);
         assert(f.curr_symbol() == " \u20ac");
     }
     {

 #if defined(_CS_GNU_LIBC_VERSION)
         // GLIBC <= 2.23 uses currency_symbol="<U0440><U0443><U0431>"
         // GLIBC >= 2.24 uses currency_symbol="<U20BD>"
         // See also: http://www.fileformat.info/info/unicode/char/20bd/index.htm
         if (!glibc_version_less_than("2.24"))
           assert(f.curr_symbol() == " \u20BD");
         else
           assert(f.curr_symbol() == " \xD1\x80\xD1\x83\xD0\xB1");
+#elif defined(_WIN32)
+        assert(f.curr_symbol() == " \u20BD");
 #else
         assert(f.curr_symbol() == " \xD1\x80\xD1\x83\xD0\xB1");
 #endif
     }
     {
         Fnt f(LOCALE_ru_RU_UTF_8, 1);
         assert(f.curr_symbol() == " RUB");
     }
 #ifndef TEST_HAS_NO_WIDE_CHARACTERS
     {
         Fwf f(LOCALE_ru_RU_UTF_8, 1);
 #if defined(_CS_GNU_LIBC_VERSION)
         if (!glibc_version_less_than("2.24"))
           assert(f.curr_symbol() == L" \u20BD");
         else
           assert(f.curr_symbol() == L" \x440\x443\x431");
+#elif defined(_WIN32)
+        assert(f.curr_symbol() == L" \u20BD");
 #else
         assert(f.curr_symbol() == L" \x440\x443\x431");
 #endif
     }
 
     {
         Fwt f(LOCALE_ru_RU_UTF_8, 1);
         assert(f.curr_symbol() == L" RUB");
     }
 #endif // TEST_HAS_NO_WIDE_CHARACTERS
 
     {
         Fnf f(LOCALE_zh_CN_UTF_8, 1);
+#ifdef _WIN32
+        assert(f.curr_symbol() == "\xC2\xA5");
+#else
         assert(f.curr_symbol() == "\xEF\xBF\xA5");
+#endif
     }
     {
         Fnt f(LOCALE_zh_CN_UTF_8, 1);
-        assert(f.curr_symbol() == "CNY ");
+        assert(f.curr_symbol() == "CNY" + curr_space);
     }
 #ifndef TEST_HAS_NO_WIDE_CHARACTERS
     {
         Fwf f(LOCALE_zh_CN_UTF_8, 1);
+#ifdef _WIN32
+        assert(f.curr_symbol() == L"\x00A5");
+#else
         assert(f.curr_symbol() == L"\xFFE5");
+#endif
     }
     {
         Fwt f(LOCALE_zh_CN_UTF_8, 1);
-        assert(f.curr_symbol() == L"CNY ");
+        assert(f.curr_symbol() == L"CNY" + w_curr_space);
     }
 #endif // TEST_HAS_NO_WIDE_CHARACTERS
 
   return 0;
 }