[PATCH v3 0/1] MR2288: kernel32:locale: Fix the French GetNumberFormatEx() test on Windows 11.

[Francois Gouget (＠fgouget)]

Windows 11 uses a narrow non-breaking space instead of a plain one when formatting numbers. This is probably more correct and maybe Wine should switch too. But marking the old result as broken() would require unbundling expect_wstr() which would be bothersome.

So this commit just accepts both.

--
  v3: kernel32/tests: Fix the French GetNumberFormatEx() test on Windows 11.

https://gitlab.winehq.org/wine/wine/-/merge_requests/2288

[PATCH v3 1/1] kernel32/tests: Fix the French GetNumberFormatEx() test on Windows 11.

[Francois Gouget]

From: Francois Gouget

Windows 11 uses a narrow non-breaking space instead of a plain one when
formatting numbers.

Wine-Bug: https://bugs.winehq.org/show_bug.cgi?id=54583
---
Using a narrow non-breaking space is probably more correct and maybe
Wine should switch too. But marking the old result as broken() would
require unbundling expect_wstr() which would be bothersome.
So this commit just accepts both.
---
 dlls/kernel32/tests/locale.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/dlls/kernel32/tests/locale.c b/dlls/kernel32/tests/locale.c
index b2c581dc518..975aeb84868 100644
--- a/dlls/kernel32/tests/locale.c
+++ b/dlls/kernel32/tests/locale.c
@@ -1816,8 +1816,12 @@ static void test_GetNumberFormatEx(void)
   /* Test French formatting */
   if (pIsValidLocaleName(frW))
   {
+    const WCHAR *expected;
     ret = pGetNumberFormatEx(frW, NUO, L"-12345", NULL, buffer, ARRAY_SIZE(buffer));
-    expect_wstr(ret, buffer, L"-12\xa0\x33\x34\x35,00"); /* Non breaking space */
+    expected = (ret && wcschr(buffer, 0x202f)) ?
+               L"-12\x202f\x33\x34\x35,00" : /* Same but narrow (win11) */
+               L"-12\xa0\x33\x34\x35,00"; /* Non breaking space */
+    expect_wstr(ret, buffer, expected);
   }
 }

[Francois Gouget (＠fgouget)]

v3: Start the summary with "kernel32/tests" as is customary instead of "kernel32:locale", although the latter would be nicer (more informative) for test fixes imho.

[Francois Gouget (＠fgouget)]

The dinput:device8 failure is a preexisting issue (see bug 54594).