Piotr Caban (@piotr) commented about dlls/msvcp120/tests/msvcp120.c:
{ "Arabic", 1256, L"t\x062a\x0686.txt" },
{ "Japanese", 932, L"t\x30af\x30e4.txt" },
{ "Chinese", 936, L"t\x4e02\x9f6b.txt" },
- };
- static const struct {
const char *loc;
UINT cp;
const char *path;
- } tests_a[] = {
{ "German", 1252, "t\xe4\xcf\xf6\xdf.txt" },
{ "Polish", 1250, "t\xea\x9c\xe6.txt" },
{ "Turkish", 1254, "t\xd0\xf0\xdd\xde\xfd\xfe.txt" },
{ "Arabic", 1256, "t\xca\x8c.txt" },
{ "Japanese", 932, "t\xb8\xd5.txt" },
{ "Chinese", 936, "t\x81\x40\xfd\x71.txt" },
- };
It doesn't make sense to have tests_a and test_w tables. It basically duplicates the tests (you're converting the strings anyway). It also introduces some problems - e.g. DeleteFileA will not work with `UTF-8` encoded tests.