From: Daniel Lehman <dlehman@esri.com>
Subject: [PATCH] make_unicode: Change handling of Turkish i to match Windows
Message-Id: <6a002e96d93944db98f8fa16f194e430@RED-INF-MXMB-P1.esri.com>
Date: Thu, 29 Sep 2016 20:44:46 +0000

From 825c50aa4f9f55ab8162b6c11ab85a07b0bf9cdd Mon Sep 17 00:00:00 2001
From: Daniel Lehman <dlehman@esri.com>
Date: Wed, 28 Sep 2016 14:45:43 -0700
Subject: [PATCH] make_unicode: Change handling of Turkish i to match Windows

Windows does not lowercase the Turkish dotted i (0x130) as 'i' (0x69)
It also does not uppercase the small dotless I (0x131) to 'I' (0x49)

Wine does the conversion for both, but Windows leaves them as-is

the change here is to make Wine mimic Windows for the Turkish i

Signed-off-by: Daniel Lehman <dlehman@esri.com>
---
 dlls/msvcrt/tests/string.c | 70 ++++++++++++++++++++++++++++++++++++++++++++++
 tools/make_unicode         |  4 +++
 2 files changed, 74 insertions(+)

diff --git a/dlls/msvcrt/tests/string.c b/dlls/msvcrt/tests/string.c
index 8e78695..dbe0103 100644
--- a/dlls/msvcrt/tests/string.c
+++ b/dlls/msvcrt/tests/string.c
@@ -88,6 +88,12 @@ static errno_t (__cdecl *p_mbslwr_s)(unsigned char *str, size_t numberOfElements
 static int (__cdecl *p_wctob)(wint_t);
 static size_t (__cdecl *p_wcrtomb)(char*, wchar_t, mbstate_t*);
 static int (__cdecl *p_tolower)(int);
+static wint_t (__cdecl *p_towlower)(wint_t);
+static wint_t (__cdecl *p__towlower_l)(wint_t,_locale_t);
+static wint_t (__cdecl *p_towupper)(wint_t);
+static wint_t (__cdecl *p__towupper_l)(wint_t,_locale_t);
+static _locale_t (__cdecl *p__create_locale)(int,const char*);
+static void (__cdecl *p__free_locale)(_locale_t);
 static size_t (__cdecl *p_mbrlen)(const char*, size_t, mbstate_t*);
 static size_t (__cdecl *p_mbrtowc)(wchar_t*, const char*, size_t, mbstate_t*);
 static int (__cdecl *p__atodbl_l)(_CRT_DOUBLE*,char*,_locale_t);
@@ -3113,6 +3119,63 @@ static void test__mbscmp(void)
     ok(ret == 1, "got %d\n", ret);
 }
 
+static void test_casemap(void)
+{
+    int i;
+    wint_t ret;
+    _locale_t locale;
+    const char *locales[] = {"Turkish_Turkey.1254", "C"};
+
+    /* test case mappings for certain characters
+       where Windows deviates from Unicode */
+
+    for (i = 0; i < sizeof(locales)/sizeof(locales[0]); i++)
+    {
+        if(!setlocale(LC_ALL, locales[i]))
+        {
+            win_skip("locale %s not available.  skipping\n", locales[i]);
+            continue;
+        }
+
+        ret = p_towlower('I');
+        ok(ret == 'i', "ret = %x\n", ret);
+
+        ret = p_towupper('i');
+        ok(ret == 'I', "ret = %x\n", ret);
+
+        /* Turkish capital dotted i - should lowercase to 'i' 0x69 */
+        ret = p_towlower(0x130);
+        ok(ret == 0x130, "ret = %x\n", ret);
+
+        /* Turkish small dotless i - should uppercase to 'I' 0x49 */
+        ret = p_towupper(0x131);
+        ok(ret == 0x131, "ret = %x\n", ret);
+    }
+
+    if (!p__towlower_l || !p__towupper_l || !p__create_locale)
+    {
+        win_skip("_towlower_l/_towupper_l/_create_locale not available\n");
+        return;
+    }
+
+    for (i = 0; i < sizeof(locales)/sizeof(locales[0]); i++)
+    {
+        if (!(locale = p__create_locale(LC_ALL, locales[i])))
+        {
+            win_skip("locale %s not available.  skipping\n", locales[i]);
+            continue;
+        }
+
+        ret = p__towlower_l(0x130, locale);
+        ok(ret == 0x130, "ret = %x\n", ret);
+
+        ret = p__towupper_l(0x131, locale);
+        ok(ret == 0x131, "ret = %x\n", ret);
+
+        p__free_locale(locale);
+    }
+}
+
 START_TEST(string)
 {
     char mem[100];
@@ -3157,6 +3220,12 @@ START_TEST(string)
     p_wctob = (void*)GetProcAddress(hMsvcrt, "wctob");
     p_wcrtomb = (void*)GetProcAddress(hMsvcrt, "wcrtomb");
     p_tolower = (void*)GetProcAddress(hMsvcrt, "tolower");
+    p_towlower = (void*)GetProcAddress(hMsvcrt, "towlower");
+    p__towlower_l = (void*)GetProcAddress(hMsvcrt, "_towlower_l");
+    p_towupper = (void*)GetProcAddress(hMsvcrt, "towupper");
+    p__towupper_l = (void*)GetProcAddress(hMsvcrt, "_towupper_l");
+    p__create_locale = (void*)GetProcAddress(hMsvcrt, "_create_locale");
+    p__free_locale = (void*)GetProcAddress(hMsvcrt, "_free_locale");
     p_mbrlen = (void*)GetProcAddress(hMsvcrt, "mbrlen");
     p_mbrtowc = (void*)GetProcAddress(hMsvcrt, "mbrtowc");
     p_mbsrtowcs = (void*)GetProcAddress(hMsvcrt, "mbsrtowcs");
@@ -3227,4 +3296,5 @@ START_TEST(string)
     test__strnset_s();
     test__wcsset_s();
     test__mbscmp();
+    test_casemap();
 }
diff --git a/tools/make_unicode b/tools/make_unicode
index a0bcf8f..10821b8 100755
--- a/tools/make_unicode
+++ b/tools/make_unicode
@@ -542,6 +542,10 @@ sub READ_DEFAULTS($)
         my $flag = $ctype{$cat};
         foreach my $i (@{$special_categories{$cat}}) { $category_table[$i] |= $flag; }
     }
+
+    # Windows does not convert these characters
+    $tolower_table[0x0130] = 0x0130; # LATIN CAPITAL LETTER I WITH DOT ABOVE (Turkish)
+    $toupper_table[0x0131] = 0x0131; # LATIN SMALL LETTER DOTLESS I (Turkish)
 }
 
 

-- 
1.9.5