Sweden-Number/dlls/mlang/tests/mlang.c

2066 lines
84 KiB
C

/*
* Unit test suite for MLANG APIs.
*
* Copyright 2004 Dmitry Timoshkov
* Copyright 2009 Detlef Riekenberg
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
*/
#define COBJMACROS
#include <stdarg.h>
#include <stdio.h>
#include "windef.h"
#include "winbase.h"
#include "winerror.h"
#include "initguid.h"
#include "mlang.h"
#include "wine/test.h"
#ifndef CP_UNICODE
#define CP_UNICODE 1200
#endif
/* #define DUMP_CP_INFO */
/* #define DUMP_SCRIPT_INFO */
static BOOL (WINAPI *pGetCPInfoExA)(UINT, DWORD, LPCPINFOEXA);
static HRESULT (WINAPI *pConvertINetMultiByteToUnicode)(LPDWORD, DWORD, LPCSTR,
LPINT, LPWSTR, LPINT);
static HRESULT (WINAPI *pConvertINetUnicodeToMultiByte)(LPDWORD, DWORD, LPCWSTR,
LPINT, LPSTR, LPINT);
static HRESULT (WINAPI *pRfc1766ToLcidA)(LCID *, LPCSTR);
static HRESULT (WINAPI *pLcidToRfc1766A)(LCID, LPSTR, INT);
typedef struct lcid_tag_table {
LPCSTR rfc1766;
LCID lcid;
HRESULT hr;
LCID broken_lcid;
LPCSTR broken_rfc;
} lcid_table_entry;
/* en, ar and zh use SUBLANG_NEUTRAL for the rfc1766 name without the country
all others suppress the country with SUBLANG_DEFAULT.
For 3 letter language codes, the rfc1766 is too small for the country */
static const lcid_table_entry lcid_table[] = {
{"e", -1, E_FAIL},
{"", -1, E_FAIL},
{"-", -1, E_FAIL},
{"e-", -1, E_FAIL},
{"ar", 1, S_OK},
{"zh", 4, S_OK},
{"de", 0x0407, S_OK},
{"de-ch", 0x0807, S_OK},
{"de-at", 0x0c07, S_OK},
{"de-lu", 0x1007, S_OK},
{"de-li", 0x1407, S_OK},
{"en", 9, S_OK},
{"en-gb", 0x809, S_OK},
{"en-GB", 0x809, S_OK},
{"EN-GB", 0x809, S_OK},
{"en-US", 0x409, S_OK},
{"en-us", 0x409, S_OK},
{"fr", 0x040c, S_OK},
{"fr-be", 0x080c, S_OK},
{"fr-ca", 0x0c0c, S_OK},
{"fr-ch", 0x100c, S_OK},
{"fr-lu", 0x140c, S_OK},
{"fr-mc", 0x180c, S_OK, 0x040c, "fr"},
{"it", 0x0410, S_OK},
{"it-ch", 0x0810, S_OK},
{"nl", 0x0413, S_OK},
{"nl-be", 0x0813, S_OK},
{"pl", 0x0415, S_OK},
{"ru", 0x0419, S_OK},
{"kok", 0x0457, S_OK, 0x0412, "x-kok"}
};
#define TODO_NAME 1
typedef struct info_table_tag {
LCID lcid;
LANGID lang;
DWORD todo;
LPCSTR rfc1766;
LPCWSTR localename;
LPCWSTR broken_name;
} info_table_entry;
static const WCHAR de_en[] = {'E','n','g','l','i','s','c','h',0};
static const WCHAR de_enca[] = {'E','n','g','l','i','s','c','h',' ',
'(','K','a','n','a','d','a',')',0};
static const WCHAR de_engb[] = {'E','n','g','l','i','s','c','h',' ',
'(','G','r','o',0xDF,'b','r','i','t','a','n','n','i','e','n',')',0};
static const WCHAR de_engb2[] ={'E','n','g','l','i','s','c','h',' ',
'(','V','e','r','e','i','n','i','g','t','e','s',' ',
'K',0xF6,'n','i','g','r','e','i','c',0};
static const WCHAR de_enus[] = {'E','n','g','l','i','s','c','h',' ',
'(','U','S','A',')',0};
static const WCHAR de_de[] = {'D','e','u','t','s','c','h',' ',
'(','D','e','u','t','s','c','h','l','a','n','d',')',0};
static const WCHAR de_deat[] = {'D','e','u','t','s','c','h',' ',
'(',0xD6,'s','t','e','r','r','e','i','c','h',')',0};
static const WCHAR de_dech[] = {'D','e','u','t','s','c','h',' ',
'(','S','c','h','w','e','i','z',')',0};
static const WCHAR en_en[] = {'E','n','g','l','i','s','h',0};
static const WCHAR en_enca[] = {'E','n','g','l','i','s','h',' ',
'(','C','a','n','a','d','a',')',0};
static const WCHAR en_engb[] = {'E','n','g','l','i','s','h',' ',
'(','U','n','i','t','e','d',' ','K','i','n','g','d','o','m',')',0};
static const WCHAR en_enus[] = {'E','n','g','l','i','s','h',' ',
'(','U','n','i','t','e','d',' ','S','t','a','t','e','s',')',0};
static const WCHAR en_de[] = {'G','e','r','m','a','n',' ',
'(','G','e','r','m','a','n','y',')',0};
static const WCHAR en_deat[] = {'G','e','r','m','a','n',' ',
'(','A','u','s','t','r','i','a',')',0};
static const WCHAR en_dech[] = {'G','e','r','m','a','n',' ',
'(','S','w','i','t','z','e','r','l','a','n','d',')',0};
static const WCHAR fr_en[] = {'A','n','g','l','a','i','s',0};
static const WCHAR fr_enca[] = {'A','n','g','l','a','i','s',' ',
'(','C','a','n','a','d','a',')',0};
static const WCHAR fr_engb[] = {'A','n','g','l','a','i','s',' ',
'(','R','o','y','a','u','m','e','-','U','n','i',')',0};
static const WCHAR fr_enus[] = {'A','n','g','l','a','i','s',' ',
'(',0xC9, 't','a','t','s','-','U','n','i','s',')',0};
static const WCHAR fr_enus2[] ={'A','n','g','l','a','i','s',' ',
'(','U','.','S','.',')',0};
static const WCHAR fr_de[] = {'A','l','l','e','m','a','n','d',' ',
'(','A','l','l','e','m','a','g','n','e',')',0};
static const WCHAR fr_de2[] = {'A','l','l','e','m','a','n','d',' ',
'(','S','t','a','n','d','a','r','d',')',0};
static const WCHAR fr_deat[] = {'A','l','l','e','m','a','n','d',' ',
'(','A','u','t','r','i','c','h','e',')',0};
static const WCHAR fr_dech[] = {'A','l','l','e','m','a','n','d',' ',
'(','S','u','i','s','s','e',')',0};
static const info_table_entry info_table[] = {
{MAKELANGID(LANG_ENGLISH, SUBLANG_NEUTRAL), MAKELANGID(LANG_ENGLISH, SUBLANG_NEUTRAL),
0, "en", en_en},
{MAKELANGID(LANG_ENGLISH, SUBLANG_DEFAULT), MAKELANGID(LANG_ENGLISH, SUBLANG_NEUTRAL),
0, "en-us", en_enus},
{MAKELANGID(LANG_ENGLISH, SUBLANG_ENGLISH_UK), MAKELANGID(LANG_ENGLISH, SUBLANG_NEUTRAL),
0, "en-gb", en_engb},
{MAKELANGID(LANG_ENGLISH, SUBLANG_ENGLISH_US), MAKELANGID(LANG_ENGLISH, SUBLANG_NEUTRAL),
0, "en-us", en_enus},
{MAKELANGID(LANG_ENGLISH, SUBLANG_ENGLISH_CAN), MAKELANGID(LANG_ENGLISH, SUBLANG_NEUTRAL),
0, "en-ca", en_enca},
{MAKELANGID(LANG_GERMAN, SUBLANG_DEFAULT), MAKELANGID(LANG_ENGLISH, SUBLANG_NEUTRAL),
0, "de", en_de},
{MAKELANGID(LANG_GERMAN, SUBLANG_GERMAN), MAKELANGID(LANG_ENGLISH, SUBLANG_NEUTRAL),
0, "de", en_de},
{MAKELANGID(LANG_GERMAN, SUBLANG_GERMAN_SWISS), MAKELANGID(LANG_ENGLISH, SUBLANG_NEUTRAL),
0, "de-ch", en_dech},
{MAKELANGID(LANG_GERMAN, SUBLANG_GERMAN_AUSTRIAN), MAKELANGID(LANG_ENGLISH, SUBLANG_NEUTRAL),
0, "de-at", en_deat},
{MAKELANGID(LANG_ENGLISH, SUBLANG_NEUTRAL), MAKELANGID(LANG_GERMAN, SUBLANG_DEFAULT),
TODO_NAME, "en", de_en},
{MAKELANGID(LANG_ENGLISH, SUBLANG_DEFAULT), MAKELANGID(LANG_GERMAN, SUBLANG_DEFAULT),
TODO_NAME, "en-us", de_enus},
{MAKELANGID(LANG_ENGLISH, SUBLANG_ENGLISH_UK), MAKELANGID(LANG_GERMAN, SUBLANG_DEFAULT),
TODO_NAME, "en-gb", de_engb, de_engb2},
{MAKELANGID(LANG_ENGLISH, SUBLANG_ENGLISH_US), MAKELANGID(LANG_GERMAN, SUBLANG_DEFAULT),
TODO_NAME, "en-us", de_enus},
{MAKELANGID(LANG_ENGLISH, SUBLANG_ENGLISH_CAN), MAKELANGID(LANG_GERMAN, SUBLANG_DEFAULT),
TODO_NAME, "en-ca", de_enca},
{MAKELANGID(LANG_GERMAN, SUBLANG_DEFAULT), MAKELANGID(LANG_GERMAN, SUBLANG_DEFAULT),
TODO_NAME, "de", de_de},
{MAKELANGID(LANG_GERMAN, SUBLANG_GERMAN), MAKELANGID(LANG_GERMAN, SUBLANG_DEFAULT),
TODO_NAME, "de",de_de},
{MAKELANGID(LANG_GERMAN, SUBLANG_GERMAN_SWISS), MAKELANGID(LANG_GERMAN, SUBLANG_DEFAULT),
TODO_NAME, "de-ch", de_dech},
{MAKELANGID(LANG_GERMAN, SUBLANG_GERMAN_AUSTRIAN), MAKELANGID(LANG_GERMAN, SUBLANG_DEFAULT),
TODO_NAME, "de-at", de_deat},
{MAKELANGID(LANG_ENGLISH, SUBLANG_NEUTRAL), MAKELANGID(LANG_FRENCH, SUBLANG_DEFAULT),
TODO_NAME, "en", fr_en},
{MAKELANGID(LANG_ENGLISH, SUBLANG_DEFAULT), MAKELANGID(LANG_FRENCH, SUBLANG_DEFAULT),
TODO_NAME, "en-us", fr_enus, fr_enus2},
{MAKELANGID(LANG_ENGLISH, SUBLANG_ENGLISH_UK), MAKELANGID(LANG_FRENCH, SUBLANG_DEFAULT),
TODO_NAME, "en-gb", fr_engb},
{MAKELANGID(LANG_ENGLISH, SUBLANG_ENGLISH_US), MAKELANGID(LANG_FRENCH, SUBLANG_DEFAULT),
TODO_NAME, "en-us", fr_enus, fr_enus2},
{MAKELANGID(LANG_ENGLISH, SUBLANG_ENGLISH_CAN), MAKELANGID(LANG_FRENCH, SUBLANG_DEFAULT),
TODO_NAME, "en-ca", fr_enca},
{MAKELANGID(LANG_GERMAN, SUBLANG_DEFAULT), MAKELANGID(LANG_FRENCH, SUBLANG_DEFAULT),
TODO_NAME, "de", fr_de, fr_de2},
{MAKELANGID(LANG_GERMAN, SUBLANG_GERMAN), MAKELANGID(LANG_FRENCH, SUBLANG_DEFAULT),
TODO_NAME, "de", fr_de, fr_de2},
{MAKELANGID(LANG_GERMAN, SUBLANG_GERMAN_SWISS), MAKELANGID(LANG_FRENCH, SUBLANG_DEFAULT),
TODO_NAME, "de-ch", fr_dech},
{MAKELANGID(LANG_GERMAN, SUBLANG_GERMAN_AUSTRIAN), MAKELANGID(LANG_FRENCH, SUBLANG_DEFAULT),
TODO_NAME, "de-at", fr_deat}
};
static BOOL init_function_ptrs(void)
{
HMODULE hMlang;
hMlang = LoadLibraryA("mlang.dll");
if (!hMlang)
{
skip("mlang not available\n");
return FALSE;
}
pConvertINetMultiByteToUnicode = (void *)GetProcAddress(hMlang, "ConvertINetMultiByteToUnicode");
pConvertINetUnicodeToMultiByte = (void *)GetProcAddress(hMlang, "ConvertINetUnicodeToMultiByte");
pRfc1766ToLcidA = (void *)GetProcAddress(hMlang, "Rfc1766ToLcidA");
pLcidToRfc1766A = (void *)GetProcAddress(hMlang, "LcidToRfc1766A");
pGetCPInfoExA = (void *)GetProcAddress(GetModuleHandleA("kernel32.dll"), "GetCPInfoExA");
return TRUE;
}
#define ok_w2(format, szString1, szString2) \
\
if (lstrcmpW((szString1), (szString2)) != 0) \
{ \
CHAR string1[256], string2[256]; \
WideCharToMultiByte(CP_ACP, 0, (szString1), -1, string1, 256, NULL, NULL); \
WideCharToMultiByte(CP_ACP, 0, (szString2), -1, string2, 256, NULL, NULL); \
ok(0, (format), string1, string2); \
}
static void test_multibyte_to_unicode_translations(IMultiLanguage2 *iML2)
{
/* these APIs are broken regarding constness of the input buffer */
char stringA[] = "Just a test string\0"; /* double 0 for CP_UNICODE tests */
WCHAR stringW[] = {'J','u','s','t',' ','a',' ','t','e','s','t',' ','s','t','r','i','n','g',0};
char bufA[256];
WCHAR bufW[256];
UINT lenA, lenW, expected_len;
HRESULT ret;
/* IMultiLanguage2_ConvertStringToUnicode tests */
memset(bufW, 'x', sizeof(bufW));
lenA = 0;
lenW = sizeof(bufW)/sizeof(bufW[0]);
ret = IMultiLanguage2_ConvertStringToUnicode(iML2, NULL, 1252, stringA, &lenA, bufW, &lenW);
ok(ret == S_OK, "IMultiLanguage2_ConvertStringToUnicode failed: %08x\n", ret);
ok(lenA == 0, "expected lenA 0, got %u\n", lenA);
ok(lenW == 0, "expected lenW 0, got %u\n", lenW);
memset(bufW, 'x', sizeof(bufW));
lenA = -1;
lenW = sizeof(bufW)/sizeof(bufW[0]);
ret = IMultiLanguage2_ConvertStringToUnicode(iML2, NULL, 1252, stringA, &lenA, bufW, &lenW);
ok(ret == S_OK, "IMultiLanguage2_ConvertStringToUnicode failed: %08x\n", ret);
ok(lenA == lstrlenA(stringA), "expected lenA %u, got %u\n", lstrlenA(stringA), lenA);
ok(lenW == lstrlenW(stringW), "expected lenW %u, got %u\n", lstrlenW(stringW), lenW);
if (lenW < sizeof(bufW)/sizeof(bufW[0])) {
/* can only happen if the convert call fails */
ok(bufW[lenW] != 0, "buf should not be 0 terminated\n");
bufW[lenW] = 0; /* -1 doesn't include 0 terminator */
}
ok(!lstrcmpW(bufW, stringW), "bufW/stringW mismatch\n");
memset(bufW, 'x', sizeof(bufW));
lenA = -1;
lenW = 5;
ret = IMultiLanguage2_ConvertStringToUnicode(iML2, NULL, 1252, stringA, &lenA, bufW, &lenW);
ok(ret == E_FAIL, "IMultiLanguage2_ConvertStringToUnicode should fail: %08x\n", ret);
ok(lenW == 0, "expected lenW 0, got %u\n", lenW);
/* still has to do partial conversion */
ok(!memcmp(bufW, stringW, 5 * sizeof(WCHAR)), "bufW/stringW mismatch\n");
memset(bufW, 'x', sizeof(bufW));
lenA = -1;
lenW = sizeof(bufW)/sizeof(bufW[0]);
ret = IMultiLanguage2_ConvertStringToUnicode(iML2, NULL, CP_UNICODE, stringA, &lenA, bufW, &lenW);
ok(ret == S_OK, "IMultiLanguage2_ConvertStringToUnicode failed: %08x\n", ret);
ok(lenA == lstrlenA(stringA), "expected lenA %u, got %u\n", lstrlenA(stringA), lenA);
ok(lenW == lstrlenW(stringW)/(int)sizeof(WCHAR), "wrong lenW %u\n", lenW);
ok(bufW[lenW] != 0, "buf should not be 0 terminated\n");
bufW[lenW] = 0; /* -1 doesn't include 0 terminator */
ok(!lstrcmpA((LPCSTR)bufW, stringA), "bufW/stringA mismatch\n");
memset(bufW, 'x', sizeof(bufW));
lenA = lstrlenA(stringA);
lenW = 0;
ret = IMultiLanguage2_ConvertStringToUnicode(iML2, NULL, 1252, stringA, &lenA, NULL, &lenW);
ok(ret == S_OK, "IMultiLanguage2_ConvertStringToUnicode failed: %08x\n", ret);
ok(lenA == lstrlenA(stringA), "expected lenA %u, got %u\n", lstrlenA(stringA), lenA);
expected_len = MultiByteToWideChar(1252, 0, stringA, lenA, NULL, 0);
ok(lenW == expected_len, "expected lenW %u, got %u\n", expected_len, lenW);
memset(bufW, 'x', sizeof(bufW));
lenA = lstrlenA(stringA);
lenW = sizeof(bufW)/sizeof(bufW[0]);
ret = pConvertINetMultiByteToUnicode(NULL, 1252, stringA, (INT *)&lenA, NULL, (INT *)&lenW);
ok(ret == S_OK, "ConvertINetMultiByteToUnicode failed: %08x\n", ret);
ok(lenA == lstrlenA(stringA), "expected lenA %u, got %u\n", lstrlenA(stringA), lenA);
expected_len = MultiByteToWideChar(1252, 0, stringA, lenA, NULL, 0);
ok(lenW == expected_len, "expected lenW %u, got %u\n", expected_len, lenW);
memset(bufW, 'x', sizeof(bufW));
lenA = lstrlenA(stringA);
lenW = 0;
ret = pConvertINetMultiByteToUnicode(NULL, 1252, stringA, (INT *)&lenA, NULL, (INT *)&lenW);
ok(ret == S_OK, "ConvertINetMultiByteToUnicode failed: %08x\n", ret);
ok(lenA == lstrlenA(stringA), "expected lenA %u, got %u\n", lstrlenA(stringA), lenA);
expected_len = MultiByteToWideChar(1252, 0, stringA, lenA, NULL, 0);
ok(lenW == expected_len, "expected lenW %u, got %u\n", expected_len, lenW);
/* IMultiLanguage2_ConvertStringFromUnicode tests */
memset(bufA, 'x', sizeof(bufA));
lenW = 0;
lenA = sizeof(bufA);
ret = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, stringW, &lenW, bufA, &lenA);
ok(ret == S_OK, "IMultiLanguage2_ConvertStringFromUnicode failed: %08x\n", ret);
ok(lenA == 0, "expected lenA 0, got %u\n", lenA);
ok(lenW == 0, "expected lenW 0, got %u\n", lenW);
memset(bufA, 'x', sizeof(bufA));
lenW = -1;
lenA = sizeof(bufA);
ret = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, stringW, &lenW, bufA, &lenA);
ok(ret == S_OK, "IMultiLanguage2_ConvertStringFromUnicode failed: %08x\n", ret);
ok(lenA == lstrlenA(stringA), "expected lenA %u, got %u\n", lstrlenA(stringA), lenA);
ok(lenW == lstrlenW(stringW), "expected lenW %u, got %u\n", lstrlenW(stringW), lenW);
ok(bufA[lenA] != 0, "buf should not be 0 terminated\n");
bufA[lenA] = 0; /* -1 doesn't include 0 terminator */
ok(!lstrcmpA(bufA, stringA), "bufA/stringA mismatch\n");
memset(bufA, 'x', sizeof(bufA));
lenW = -1;
lenA = 5;
ret = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, stringW, &lenW, bufA, &lenA);
ok(ret == E_FAIL, "IMultiLanguage2_ConvertStringFromUnicode should fail: %08x\n", ret);
ok(lenA == 0, "expected lenA 0, got %u\n", lenA);
/* still has to do partial conversion */
ok(!memcmp(bufA, stringA, 5), "bufW/stringW mismatch\n");
memset(bufA, 'x', sizeof(bufA));
lenW = -1;
lenA = sizeof(bufA);
ret = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, CP_UNICODE, stringW, &lenW, bufA, &lenA);
ok(ret == S_OK, "IMultiLanguage2_ConvertStringFromUnicode failed: %08x\n", ret);
ok(lenA == lstrlenA(stringA) * (int)sizeof(WCHAR), "wrong lenA %u\n", lenA);
ok(lenW == lstrlenW(stringW), "expected lenW %u, got %u\n", lstrlenW(stringW), lenW);
ok(bufA[lenA] != 0 && bufA[lenA+1] != 0, "buf should not be 0 terminated\n");
bufA[lenA] = 0; /* -1 doesn't include 0 terminator */
bufA[lenA+1] = 0; /* sizeof(WCHAR) */
ok(!lstrcmpW((LPCWSTR)bufA, stringW), "bufA/stringW mismatch\n");
memset(bufA, 'x', sizeof(bufA));
lenW = lstrlenW(stringW);
lenA = 0;
ret = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, stringW, &lenW, NULL, &lenA);
ok(ret == S_OK, "IMultiLanguage2_ConvertStringFromUnicode failed: %08x\n", ret);
ok(lenW == lstrlenW(stringW), "expected lenW %u, got %u\n", lstrlenW(stringW), lenW);
expected_len = WideCharToMultiByte(1252, 0, stringW, lenW, NULL, 0, NULL, NULL);
ok(lenA == expected_len, "expected lenA %u, got %u\n", expected_len, lenA);
}
static void cpinfo_cmp(MIMECPINFO *cpinfo1, MIMECPINFO *cpinfo2)
{
ok(cpinfo1->dwFlags == cpinfo2->dwFlags, "dwFlags mismatch: %08x != %08x\n", cpinfo1->dwFlags, cpinfo2->dwFlags);
ok(cpinfo1->uiCodePage == cpinfo2->uiCodePage, "uiCodePage mismatch: %u != %u\n", cpinfo1->uiCodePage, cpinfo2->uiCodePage);
ok(cpinfo1->uiFamilyCodePage == cpinfo2->uiFamilyCodePage, "uiFamilyCodePage mismatch: %u != %u\n", cpinfo1->uiFamilyCodePage, cpinfo2->uiFamilyCodePage);
ok(!lstrcmpW(cpinfo1->wszDescription, cpinfo2->wszDescription), "wszDescription mismatch\n");
ok(!lstrcmpW(cpinfo1->wszWebCharset, cpinfo2->wszWebCharset), "wszWebCharset mismatch\n");
ok(!lstrcmpW(cpinfo1->wszHeaderCharset, cpinfo2->wszHeaderCharset), "wszHeaderCharset mismatch\n");
ok(!lstrcmpW(cpinfo1->wszBodyCharset, cpinfo2->wszBodyCharset), "wszBodyCharset mismatch\n");
ok(!lstrcmpW(cpinfo1->wszFixedWidthFont, cpinfo2->wszFixedWidthFont), "wszFixedWidthFont mismatch\n");
ok(!lstrcmpW(cpinfo1->wszProportionalFont, cpinfo2->wszProportionalFont), "wszProportionalFont mismatch\n");
ok(cpinfo1->bGDICharset == cpinfo2->bGDICharset, "bGDICharset mismatch: %d != %d\n", cpinfo1->bGDICharset, cpinfo2->bGDICharset);
}
#ifdef DUMP_CP_INFO
static const char *dump_mime_flags(DWORD flags)
{
static char buf[1024];
buf[0] = 0;
if (flags & MIMECONTF_MAILNEWS) strcat(buf, " MIMECONTF_MAILNEWS");
if (flags & MIMECONTF_BROWSER) strcat(buf, " MIMECONTF_BROWSER");
if (flags & MIMECONTF_MINIMAL) strcat(buf, " MIMECONTF_MINIMAL");
if (flags & MIMECONTF_IMPORT) strcat(buf, " MIMECONTF_IMPORT");
if (flags & MIMECONTF_SAVABLE_MAILNEWS) strcat(buf, " MIMECONTF_SAVABLE_MAILNEWS");
if (flags & MIMECONTF_SAVABLE_BROWSER) strcat(buf, " MIMECONTF_SAVABLE_BROWSER");
if (flags & MIMECONTF_EXPORT) strcat(buf, " MIMECONTF_EXPORT");
if (flags & MIMECONTF_PRIVCONVERTER) strcat(buf, " MIMECONTF_PRIVCONVERTER");
if (flags & MIMECONTF_VALID) strcat(buf, " MIMECONTF_VALID");
if (flags & MIMECONTF_VALID_NLS) strcat(buf, " MIMECONTF_VALID_NLS");
if (flags & MIMECONTF_MIME_IE4) strcat(buf, " MIMECONTF_MIME_IE4");
if (flags & MIMECONTF_MIME_LATEST) strcat(buf, " MIMECONTF_MIME_LATEST");
if (flags & MIMECONTF_MIME_REGISTRY) strcat(buf, " MIMECONTF_MIME_REGISTRY");
return buf;
}
#endif
static HRESULT check_convertible(IMultiLanguage2 *iML2, UINT from, UINT to)
{
CHAR convert[MAX_PATH];
BYTE dest[MAX_PATH];
HRESULT hr;
UINT srcsz, destsz;
static WCHAR strW[] = {'a','b','c',0};
/* Check to see if the target codepage has these characters or not */
if (from != CP_UTF8)
{
BOOL fDefaultChar;
char ch[10];
int cb;
cb = WideCharToMultiByte( from, WC_NO_BEST_FIT_CHARS | WC_COMPOSITECHECK | WC_DEFAULTCHAR,
strW, 3, ch, sizeof(ch), NULL, &fDefaultChar);
if(cb == 0 || fDefaultChar)
{
trace("target codepage %i does not contain 'abc'\n",from);
return E_FAIL;
}
}
srcsz = lstrlenW(strW) + 1;
destsz = MAX_PATH;
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, from, strW,
&srcsz, convert, &destsz);
if (hr != S_OK)
return S_FALSE;
srcsz = -1;
destsz = MAX_PATH;
hr = IMultiLanguage2_ConvertString(iML2, NULL, from, to, (BYTE *)convert,
&srcsz, dest, &destsz);
if (hr != S_OK)
return S_FALSE;
return S_OK;
}
static void test_EnumCodePages(IMultiLanguage2 *iML2, DWORD flags)
{
IEnumCodePage *iEnumCP = NULL;
MIMECPINFO *cpinfo;
MIMECPINFO cpinfo2;
HRESULT ret;
ULONG i, n;
UINT total;
total = 0;
ret = IMultiLanguage2_GetNumberOfCodePageInfo(iML2, &total);
ok(ret == S_OK && total != 0, "IMultiLanguage2_GetNumberOfCodePageInfo: expected S_OK/!0, got %08x/%u\n", ret, total);
trace("total mlang supported codepages %u\n", total);
ret = IMultiLanguage2_EnumCodePages(iML2, flags, LANG_NEUTRAL, &iEnumCP);
ok(ret == S_OK && iEnumCP, "IMultiLanguage2_EnumCodePages: expected S_OK/!NULL, got %08x/%p\n", ret, iEnumCP);
ret = IEnumCodePage_Reset(iEnumCP);
ok(ret == S_OK, "IEnumCodePage_Reset: expected S_OK, got %08x\n", ret);
n = 65536;
ret = IEnumCodePage_Next(iEnumCP, 0, NULL, &n);
if (ret == S_FALSE)
ok(n == 0 && ret == S_FALSE, "IEnumCodePage_Next: expected 0/S_FALSE, got %u/%08x\n", n, ret);
else if (ret == E_FAIL)
ok(n == 65536 && ret == E_FAIL, "IEnumCodePage_Next: expected 65536/E_FAIL, got %u/%08x\n", n, ret);
ret = IEnumCodePage_Next(iEnumCP, 0, NULL, NULL);
if (ret == S_FALSE)
ok(ret == S_FALSE, "IEnumCodePage_Next: expected S_FALSE, got %08x\n", ret);
else if (ret == E_FAIL)
ok(n == 65536 && ret == E_FAIL, "IEnumCodePage_Next: expected 65536/E_FAIL, got %u/%08x\n", n, ret);
cpinfo = HeapAlloc(GetProcessHeap(), 0, sizeof(*cpinfo) * total * 2);
n = total * 2;
ret = IEnumCodePage_Next(iEnumCP, 0, cpinfo, &n);
ok(ret == S_FALSE && n == 0, "IEnumCodePage_Next: expected S_FALSE/0, got %08x/%u\n", ret, n);
n = total * 2;
ret = IEnumCodePage_Next(iEnumCP, n, cpinfo, &n);
ok(ret == S_OK && n != 0, "IEnumCodePage_Next: expected S_OK/!0, got %08x/%u\n", ret, n);
trace("flags %08x, enumerated codepages %u\n", flags, n);
if (!flags)
{
ok(n == total, "IEnumCodePage_Next: expected %u, got %u\n", total, n);
flags = MIMECONTF_MIME_LATEST;
}
total = n;
for (i = 0; i < n; i++)
{
CHARSETINFO csi;
MIMECSETINFO mcsi;
HRESULT convertible = S_OK;
static const WCHAR autoW[] = {'_','a','u','t','o',0};
static const WCHAR feffW[] = {'u','n','i','c','o','d','e','F','E','F','F',0};
#ifdef DUMP_CP_INFO
trace("MIMECPINFO #%u:\n"
"dwFlags %08x %s\n"
"uiCodePage %u\n"
"uiFamilyCodePage %u\n"
"wszDescription %s\n"
"wszWebCharset %s\n"
"wszHeaderCharset %s\n"
"wszBodyCharset %s\n"
"wszFixedWidthFont %s\n"
"wszProportionalFont %s\n"
"bGDICharset %d\n\n",
i,
cpinfo[i].dwFlags, dump_mime_flags(cpinfo[i].dwFlags),
cpinfo[i].uiCodePage,
cpinfo[i].uiFamilyCodePage,
wine_dbgstr_w(cpinfo[i].wszDescription),
wine_dbgstr_w(cpinfo[i].wszWebCharset),
wine_dbgstr_w(cpinfo[i].wszHeaderCharset),
wine_dbgstr_w(cpinfo[i].wszBodyCharset),
wine_dbgstr_w(cpinfo[i].wszFixedWidthFont),
wine_dbgstr_w(cpinfo[i].wszProportionalFont),
cpinfo[i].bGDICharset);
#endif
ok(cpinfo[i].dwFlags & flags, "enumerated flags %08x do not include requested %08x\n", cpinfo[i].dwFlags, flags);
if (TranslateCharsetInfo((DWORD *)(INT_PTR)cpinfo[i].uiFamilyCodePage, &csi, TCI_SRCCODEPAGE))
ok(cpinfo[i].bGDICharset == csi.ciCharset, "%d != %d\n", cpinfo[i].bGDICharset, csi.ciCharset);
else
if (winetest_debug > 1)
trace("TranslateCharsetInfo failed for cp %u\n", cpinfo[i].uiFamilyCodePage);
#ifdef DUMP_CP_INFO
trace("%u: codepage %u family %u\n", i, cpinfo[i].uiCodePage, cpinfo[i].uiFamilyCodePage);
#endif
/* support files for some codepages might be not installed, or
* the codepage is just an alias.
*/
if (IsValidCodePage(cpinfo[i].uiCodePage))
{
ret = IMultiLanguage2_IsConvertible(iML2, cpinfo[i].uiCodePage, CP_UNICODE);
ok(ret == S_OK, "IMultiLanguage2_IsConvertible(%u -> CP_UNICODE) = %08x\n", cpinfo[i].uiCodePage, ret);
ret = IMultiLanguage2_IsConvertible(iML2, CP_UNICODE, cpinfo[i].uiCodePage);
ok(ret == S_OK, "IMultiLanguage2_IsConvertible(CP_UNICODE -> %u) = %08x\n", cpinfo[i].uiCodePage, ret);
convertible = check_convertible(iML2, cpinfo[i].uiCodePage, CP_UTF8);
if (convertible != E_FAIL)
{
ret = IMultiLanguage2_IsConvertible(iML2, cpinfo[i].uiCodePage, CP_UTF8);
ok(ret == convertible, "IMultiLanguage2_IsConvertible(%u -> CP_UTF8) = %08x\n", cpinfo[i].uiCodePage, ret);
ret = IMultiLanguage2_IsConvertible(iML2, CP_UTF8, cpinfo[i].uiCodePage);
ok(ret == convertible, "IMultiLanguage2_IsConvertible(CP_UTF8 -> %u) = %08x\n", cpinfo[i].uiCodePage, ret);
}
}
else
if (winetest_debug > 1)
trace("IsValidCodePage failed for cp %u\n", cpinfo[i].uiCodePage);
if (memcmp(cpinfo[i].wszWebCharset,feffW,sizeof(WCHAR)*11)==0)
skip("Legacy windows bug returning invalid charset of unicodeFEFF\n");
else
{
ret = IMultiLanguage2_GetCharsetInfo(iML2, cpinfo[i].wszWebCharset, &mcsi);
/* _autoxxx charsets are a fake and GetCharsetInfo fails for them */
if (memcmp(cpinfo[i].wszWebCharset, autoW, 5 * sizeof(WCHAR)))
{
ok (ret == S_OK, "IMultiLanguage2_GetCharsetInfo failed: %08x\n", ret);
#ifdef DUMP_CP_INFO
trace("%s: %u %u %s\n", wine_dbgstr_w(cpinfo[i].wszWebCharset), mcsi.uiCodePage, mcsi.uiInternetEncoding, wine_dbgstr_w(mcsi.wszCharset));
#endif
ok(!lstrcmpiW(cpinfo[i].wszWebCharset, mcsi.wszCharset), "%s != %s\n",
wine_dbgstr_w(cpinfo[i].wszWebCharset), wine_dbgstr_w(mcsi.wszCharset));
if (0)
{
/* native mlang returns completely messed up encodings in some cases */
ok(mcsi.uiInternetEncoding == cpinfo[i].uiCodePage || mcsi.uiInternetEncoding == cpinfo[i].uiFamilyCodePage,
"%u != %u || %u\n", mcsi.uiInternetEncoding, cpinfo[i].uiCodePage, cpinfo[i].uiFamilyCodePage);
ok(mcsi.uiCodePage == cpinfo[i].uiCodePage || mcsi.uiCodePage == cpinfo[i].uiFamilyCodePage,
"%u != %u || %u\n", mcsi.uiCodePage, cpinfo[i].uiCodePage, cpinfo[i].uiFamilyCodePage);
}
}
}
if (memcmp(cpinfo[i].wszHeaderCharset,feffW,sizeof(WCHAR)*11)==0)
skip("Legacy windows bug returning invalid charset of unicodeFEFF\n");
else
{
ret = IMultiLanguage2_GetCharsetInfo(iML2, cpinfo[i].wszHeaderCharset, &mcsi);
/* _autoxxx charsets are a fake and GetCharsetInfo fails for them */
if (memcmp(cpinfo[i].wszHeaderCharset, autoW, 5 * sizeof(WCHAR)))
{
ok (ret == S_OK, "IMultiLanguage2_GetCharsetInfo failed: %08x\n", ret);
#ifdef DUMP_CP_INFO
trace("%s: %u %u %s\n", wine_dbgstr_w(cpinfo[i].wszHeaderCharset), mcsi.uiCodePage, mcsi.uiInternetEncoding, wine_dbgstr_w(mcsi.wszCharset));
#endif
ok(!lstrcmpiW(cpinfo[i].wszHeaderCharset, mcsi.wszCharset), "%s != %s\n",
wine_dbgstr_w(cpinfo[i].wszHeaderCharset), wine_dbgstr_w(mcsi.wszCharset));
if (0)
{
/* native mlang returns completely messed up encodings in some cases */
ok(mcsi.uiInternetEncoding == cpinfo[i].uiCodePage || mcsi.uiInternetEncoding == cpinfo[i].uiFamilyCodePage,
"%u != %u || %u\n", mcsi.uiInternetEncoding, cpinfo[i].uiCodePage, cpinfo[i].uiFamilyCodePage);
ok(mcsi.uiCodePage == cpinfo[i].uiCodePage || mcsi.uiCodePage == cpinfo[i].uiFamilyCodePage,
"%u != %u || %u\n", mcsi.uiCodePage, cpinfo[i].uiCodePage, cpinfo[i].uiFamilyCodePage);
}
}
}
if (memcmp(cpinfo[i].wszBodyCharset,feffW,sizeof(WCHAR)*11)==0)
skip("Legacy windows bug returning invalid charset of unicodeFEFF\n");
else
{
ret = IMultiLanguage2_GetCharsetInfo(iML2, cpinfo[i].wszBodyCharset, &mcsi);
/* _autoxxx charsets are a fake and GetCharsetInfo fails for them */
if (memcmp(cpinfo[i].wszBodyCharset, autoW, 5 * sizeof(WCHAR)))
{
ok (ret == S_OK, "IMultiLanguage2_GetCharsetInfo failed: %08x\n", ret);
#ifdef DUMP_CP_INFO
trace("%s: %u %u %s\n", wine_dbgstr_w(cpinfo[i].wszBodyCharset), mcsi.uiCodePage, mcsi.uiInternetEncoding, wine_dbgstr_w(mcsi.wszCharset));
#endif
ok(!lstrcmpiW(cpinfo[i].wszBodyCharset, mcsi.wszCharset), "%s != %s\n",
wine_dbgstr_w(cpinfo[i].wszBodyCharset), wine_dbgstr_w(mcsi.wszCharset));
if (0)
{
/* native mlang returns completely messed up encodings in some cases */
ok(mcsi.uiInternetEncoding == cpinfo[i].uiCodePage || mcsi.uiInternetEncoding == cpinfo[i].uiFamilyCodePage,
"%u != %u || %u\n", mcsi.uiInternetEncoding, cpinfo[i].uiCodePage, cpinfo[i].uiFamilyCodePage);
ok(mcsi.uiCodePage == cpinfo[i].uiCodePage || mcsi.uiCodePage == cpinfo[i].uiFamilyCodePage,
"%u != %u || %u\n", mcsi.uiCodePage, cpinfo[i].uiCodePage, cpinfo[i].uiFamilyCodePage);
}
}
}
}
/* now IEnumCodePage_Next should fail, since pointer is at the end */
n = 1;
ret = IEnumCodePage_Next(iEnumCP, 1, &cpinfo2, &n);
ok(ret == S_FALSE && n == 0, "IEnumCodePage_Next: expected S_FALSE/0, got %08x/%u\n", ret, n);
ret = IEnumCodePage_Reset(iEnumCP);
ok(ret == S_OK, "IEnumCodePage_Reset: expected S_OK, got %08x\n", ret);
n = 0;
ret = IEnumCodePage_Next(iEnumCP, 1, &cpinfo2, &n);
ok(n == 1 && ret == S_OK, "IEnumCodePage_Next: expected 1/S_OK, got %u/%08x\n", n, ret);
cpinfo_cmp(cpinfo, &cpinfo2);
if (0)
{
/* Due to a bug in MS' implementation of IEnumCodePage_Skip
* it's not used here.
*/
ret = IEnumCodePage_Skip(iEnumCP, 1);
ok(ret == S_OK, "IEnumCodePage_Skip: expected S_OK, got %08x\n", ret);
}
for (i = 0; i < total - 1; i++)
{
n = 0;
ret = IEnumCodePage_Next(iEnumCP, 1, &cpinfo2, &n);
ok(n == 1 && ret == S_OK, "IEnumCodePage_Next: expected 1/S_OK, got %u/%08x\n", n, ret);
cpinfo_cmp(&cpinfo[i + 1], &cpinfo2);
}
HeapFree(GetProcessHeap(), 0, cpinfo);
IEnumCodePage_Release(iEnumCP);
}
static void scriptinfo_cmp(SCRIPTINFO *sinfo1, SCRIPTINFO *sinfo2)
{
ok(sinfo1->ScriptId == sinfo2->ScriptId, "ScriptId mismatch: %d != %d\n", sinfo1->ScriptId, sinfo2->ScriptId);
ok(sinfo1->uiCodePage == sinfo2->uiCodePage, "uiCodePage mismatch: %u != %u\n", sinfo1->uiCodePage, sinfo2->uiCodePage);
ok(!lstrcmpW(sinfo1->wszDescription, sinfo2->wszDescription), "wszDescription mismatch\n");
ok(!lstrcmpW(sinfo1->wszFixedWidthFont, sinfo2->wszFixedWidthFont), "wszFixedWidthFont mismatch\n");
ok(!lstrcmpW(sinfo1->wszProportionalFont, sinfo2->wszProportionalFont), "wszProportionalFont mismatch\n");
}
static void test_EnumScripts(IMultiLanguage2 *iML2, DWORD flags)
{
IEnumScript *iEnumScript = NULL;
SCRIPTINFO *sinfo;
SCRIPTINFO sinfo2;
HRESULT ret;
ULONG i, n;
UINT total;
total = 0;
ret = IMultiLanguage2_GetNumberOfScripts(iML2, &total);
ok(ret == S_OK && total != 0, "IMultiLanguage2_GetNumberOfScripts: expected S_OK/!0, got %08x/%u\n", ret, total);
trace("total mlang supported scripts %u\n", total);
ret = IMultiLanguage2_EnumScripts(iML2, flags, LANG_NEUTRAL, &iEnumScript);
ok(ret == S_OK && iEnumScript, "IMultiLanguage2_EnumScripts: expected S_OK/!NULL, got %08x/%p\n", ret, iEnumScript);
ret = IEnumScript_Reset(iEnumScript);
ok(ret == S_OK, "IEnumScript_Reset: expected S_OK, got %08x\n", ret);
n = 65536;
ret = IEnumScript_Next(iEnumScript, 0, NULL, &n);
ok(n == 65536 && ret == E_FAIL, "IEnumScript_Next: expected 65536/E_FAIL, got %u/%08x\n", n, ret);
ret = IEnumScript_Next(iEnumScript, 0, NULL, NULL);
ok(ret == E_FAIL, "IEnumScript_Next: expected E_FAIL, got %08x\n", ret);
sinfo = HeapAlloc(GetProcessHeap(), 0, sizeof(*sinfo) * total * 2);
n = total * 2;
ret = IEnumScript_Next(iEnumScript, 0, sinfo, &n);
ok(ret == S_FALSE && n == 0, "IEnumScript_Next: expected S_FALSE/0, got %08x/%u\n", ret, n);
n = total * 2;
ret = IEnumScript_Next(iEnumScript, n, sinfo, &n);
ok(ret == S_OK && n != 0, "IEnumScript_Next: expected S_OK, got %08x/%u\n", ret, n);
trace("flags %08x, enumerated scripts %u\n", flags, n);
if (!flags)
{
ok(n == total, "IEnumScript_Next: expected %u, got %u\n", total, n);
}
total = n;
for (i = 0; pGetCPInfoExA && i < n; i++)
{
#ifdef DUMP_SCRIPT_INFO
trace("SCRIPTINFO #%u:\n"
"ScriptId %08x\n"
"uiCodePage %u\n"
"wszDescription %s\n"
"wszFixedWidthFont %s\n"
"wszProportionalFont %s\n\n",
i,
sinfo[i].ScriptId,
sinfo[i].uiCodePage,
wine_dbgstr_w(sinfo[i].wszDescription),
wine_dbgstr_w(sinfo[i].wszFixedWidthFont),
wine_dbgstr_w(sinfo[i].wszProportionalFont));
trace("%u codepage %u\n", i, sinfo[i].uiCodePage);
#endif
}
/* now IEnumScript_Next should fail, since pointer is at the end */
n = 1;
ret = IEnumScript_Next(iEnumScript, 1, &sinfo2, &n);
ok(ret == S_FALSE && n == 0, "IEnumScript_Next: expected S_FALSE/0, got %08x/%u\n", ret, n);
ret = IEnumScript_Reset(iEnumScript);
ok(ret == S_OK, "IEnumScript_Reset: expected S_OK, got %08x\n", ret);
n = 0;
ret = IEnumScript_Next(iEnumScript, 1, &sinfo2, &n);
ok(n == 1 && ret == S_OK, "IEnumScript_Next: expected 1/S_OK, got %u/%08x\n", n, ret);
scriptinfo_cmp(sinfo, &sinfo2);
if (0)
{
/* Due to a bug in MS' implementation of IEnumScript_Skip
* it's not used here.
*/
ret = IEnumScript_Skip(iEnumScript, 1);
ok(ret == S_OK, "IEnumScript_Skip: expected S_OK, got %08x\n", ret);
}
for (i = 0; i < total - 1; i++)
{
n = 0;
ret = IEnumScript_Next(iEnumScript, 1, &sinfo2, &n);
ok(n == 1 && ret == S_OK, "IEnumScript_Next: expected 1/S_OK, got %u/%08x\n", n, ret);
scriptinfo_cmp(&sinfo[i + 1], &sinfo2);
}
HeapFree(GetProcessHeap(), 0, sinfo);
IEnumScript_Release(iEnumScript);
}
static void IMLangFontLink_Test(IMLangFontLink* iMLFL)
{
DWORD dwCodePages, dwManyCodePages;
DWORD dwCmpCodePages;
UINT CodePage;
static const WCHAR str[] = { 'd', 0x0436, 0xff90 };
LONG processed;
HRESULT ret;
dwCodePages = ~0u;
ret = IMLangFontLink_CodePageToCodePages(iMLFL, -1, &dwCodePages);
ok(ret == E_FAIL, "IMLangFontLink_CodePageToCodePages should fail: %x\n", ret);
ok(dwCodePages == 0, "expected 0, got %u\n", dwCodePages);
dwCodePages = 0;
ret = IMLangFontLink_CodePageToCodePages(iMLFL, 932, &dwCodePages);
ok(ret == S_OK, "IMLangFontLink_CodePageToCodePages error %x\n", ret);
ok(dwCodePages == FS_JISJAPAN, "expected FS_JISJAPAN, got %08x\n", dwCodePages);
CodePage = 0;
ret = IMLangFontLink_CodePagesToCodePage(iMLFL, dwCodePages, 1035, &CodePage);
ok(ret == S_OK, "IMLangFontLink_CodePagesToCodePage error %x\n", ret);
ok(CodePage == 932, "Incorrect CodePage Returned (%i)\n",CodePage);
dwManyCodePages = 0;
ret = IMLangFontLink_CodePageToCodePages(iMLFL, 1252, &dwManyCodePages);
ok(ret == S_OK, "IMLangFontLink_CodePageToCodePages error %x\n", ret);
ok(dwManyCodePages == FS_LATIN1, "expected FS_LATIN1, got %08x\n", dwManyCodePages);
dwCodePages = 0;
ret = IMLangFontLink_CodePageToCodePages(iMLFL, 1256, &dwCodePages);
ok(ret == S_OK, "IMLangFontLink_CodePageToCodePages error %x\n", ret);
ok(dwCodePages == FS_ARABIC, "expected FS_ARABIC, got %08x\n", dwCodePages);
dwManyCodePages |= dwCodePages;
ret = IMLangFontLink_CodePageToCodePages(iMLFL, 874, &dwCodePages);
ok(ret == S_OK, "IMLangFontLink_CodePageToCodePages error %x\n", ret);
ok(dwCodePages == FS_THAI, "expected FS_THAI, got %08x\n", dwCodePages);
dwManyCodePages |= dwCodePages;
ret = IMLangFontLink_CodePagesToCodePage(iMLFL, dwManyCodePages, 1256, &CodePage);
ok(ret == S_OK, "IMLangFontLink_CodePagesToCodePage error %x\n", ret);
ok(CodePage == 1256, "Incorrect CodePage Returned (%i)\n",CodePage);
ret = IMLangFontLink_CodePagesToCodePage(iMLFL, dwManyCodePages, 936, &CodePage);
ok(ret == S_OK, "IMLangFontLink_CodePagesToCodePage error %x\n", ret);
ok(CodePage == 1252, "Incorrect CodePage Returned (%i)\n",CodePage);
/* Tests for GetCharCodePages */
/* Latin 1 */
dwCmpCodePages = FS_LATIN1 | FS_LATIN2 | FS_CYRILLIC | FS_GREEK | FS_TURKISH
| FS_HEBREW | FS_ARABIC | FS_BALTIC | FS_VIETNAMESE | FS_THAI
| FS_JISJAPAN | FS_CHINESESIMP | FS_WANSUNG | FS_CHINESETRAD;
ret = IMLangFontLink_GetCharCodePages(iMLFL, 'd', &dwCodePages);
ok(ret == S_OK, "IMLangFontLink_GetCharCodePages error %x\n", ret);
ok(dwCodePages == dwCmpCodePages, "expected %x, got %x\n", dwCmpCodePages, dwCodePages);
dwCodePages = 0;
processed = 0;
ret = IMLangFontLink_GetStrCodePages(iMLFL, str, 1, 0, &dwCodePages, &processed);
ok(ret == S_OK, "IMLangFontLink_GetStrCodePages error %x\n", ret);
ok(dwCodePages == dwCmpCodePages, "expected %x, got %x\n", dwCmpCodePages, dwCodePages);
ok(processed == 1, "expected 1, got %d\n", processed);
/* Cyrillic */
dwCmpCodePages = FS_CYRILLIC | FS_JISJAPAN | FS_CHINESESIMP | FS_WANSUNG;
ret = IMLangFontLink_GetCharCodePages(iMLFL, 0x0436, &dwCodePages);
ok(ret == S_OK, "IMLangFontLink_GetCharCodePages error %x\n", ret);
ok(dwCodePages == dwCmpCodePages, "expected %x, got %x\n", dwCmpCodePages, dwCodePages);
dwCodePages = 0;
processed = 0;
ret = IMLangFontLink_GetStrCodePages(iMLFL, &str[1], 1, 0, &dwCodePages, &processed);
ok(ret == S_OK, "IMLangFontLink_GetStrCodePages error %x\n", ret);
ok(dwCodePages == dwCmpCodePages, "expected %x, got %x\n", dwCmpCodePages, dwCodePages);
ok(processed == 1, "expected 1, got %d\n", processed);
/* Japanese */
dwCmpCodePages = FS_JISJAPAN;
ret = IMLangFontLink_GetCharCodePages(iMLFL, 0xff90, &dwCodePages);
ok(ret == S_OK, "IMLangFontLink_GetCharCodePages error %x\n", ret);
ok(dwCodePages == dwCmpCodePages, "expected %x, got %x\n", dwCmpCodePages, dwCodePages);
dwCodePages = 0;
processed = 0;
ret = IMLangFontLink_GetStrCodePages(iMLFL, &str[2], 1, 0, &dwCodePages, &processed);
ok(ret == S_OK, "IMLangFontLink_GetStrCodePages error %x\n", ret);
ok(dwCodePages == dwCmpCodePages, "expected %x, got %x\n", dwCmpCodePages, dwCodePages);
ok(processed == 1, "expected 1, got %d\n", processed);
dwCmpCodePages = FS_CYRILLIC | FS_JISJAPAN | FS_CHINESESIMP | FS_WANSUNG;
dwCodePages = 0;
processed = 0;
ret = IMLangFontLink_GetStrCodePages(iMLFL, str, 2, 0, &dwCodePages, &processed);
ok(ret == S_OK, "IMLangFontLink_GetStrCodePages error %x\n", ret);
ok(dwCodePages == dwCmpCodePages, "expected %x, got %x\n", dwCmpCodePages, dwCodePages);
ok(processed == 2, "expected 2, got %d\n", processed);
dwCmpCodePages = FS_JISJAPAN;
dwCodePages = 0;
processed = 0;
ret = IMLangFontLink_GetStrCodePages(iMLFL, str, 3, 0, &dwCodePages, &processed);
ok(ret == S_OK, "IMLangFontLink_GetStrCodePages error %x\n", ret);
ok(dwCodePages == dwCmpCodePages, "expected %x, got %x\n", dwCmpCodePages, dwCodePages);
ok(processed == 3, "expected 3, got %d\n", processed);
dwCodePages = 0xffff;
processed = -1;
ret = IMLangFontLink_GetStrCodePages(iMLFL, &str[2], 1, 0, &dwCodePages, &processed);
ok(ret == S_OK, "IMLangFontLink_GetStrCodePages error %x\n", ret);
ok(dwCodePages == dwCmpCodePages, "expected %x, got %x\n", dwCmpCodePages, dwCodePages);
ok(processed == 1, "expected 0, got %d\n", processed);
ret = IMLangFontLink_GetStrCodePages(iMLFL, &str[2], 1, 0, NULL, NULL);
ok(ret == S_OK, "IMLangFontLink_GetStrCodePages error %x\n", ret);
dwCodePages = 0xffff;
processed = -1;
ret = IMLangFontLink_GetStrCodePages(iMLFL, str, -1, 0, &dwCodePages, &processed);
ok(ret == E_INVALIDARG, "IMLangFontLink_GetStrCodePages should fail: %x\n", ret);
ok(dwCodePages == 0, "expected %x, got %x\n", dwCmpCodePages, dwCodePages);
ok(processed == 0, "expected 0, got %d\n", processed);
dwCodePages = 0xffff;
processed = -1;
ret = IMLangFontLink_GetStrCodePages(iMLFL, NULL, 1, 0, &dwCodePages, &processed);
ok(ret == E_INVALIDARG, "IMLangFontLink_GetStrCodePages should fail: %x\n", ret);
ok(dwCodePages == 0, "expected %x, got %x\n", dwCmpCodePages, dwCodePages);
ok(processed == 0, "expected 0, got %d\n", processed);
dwCodePages = 0xffff;
processed = -1;
ret = IMLangFontLink_GetStrCodePages(iMLFL, str, 0, 0, &dwCodePages, &processed);
ok(ret == E_INVALIDARG, "IMLangFontLink_GetStrCodePages should fail: %x\n", ret);
ok(dwCodePages == 0, "expected %x, got %x\n", dwCmpCodePages, dwCodePages);
ok(processed == 0, "expected 0, got %d\n", processed);
}
/* copied from libs/wine/string.c */
static WCHAR *strstrW(const WCHAR *str, const WCHAR *sub)
{
while (*str)
{
const WCHAR *p1 = str, *p2 = sub;
while (*p1 && *p2 && *p1 == *p2) { p1++; p2++; }
if (!*p2) return (WCHAR *)str;
str++;
}
return NULL;
}
static void test_rfc1766(IMultiLanguage2 *iML2)
{
IEnumRfc1766 *pEnumRfc1766;
RFC1766INFO info;
ULONG n;
HRESULT ret;
BSTR rfcstr;
ret = IMultiLanguage2_EnumRfc1766(iML2, LANG_NEUTRAL, &pEnumRfc1766);
ok(ret == S_OK, "IMultiLanguage2_EnumRfc1766 error %08x\n", ret);
while (1)
{
ret = IEnumRfc1766_Next(pEnumRfc1766, 1, &info, &n);
if (ret != S_OK) break;
#ifdef DUMP_CP_INFO
trace("lcid %04x rfc_name %s locale_name %s\n",
info.lcid, wine_dbgstr_w(info.wszRfc1766), wine_dbgstr_w(info.wszLocaleName));
#endif
ok(n == 1, "couldn't fetch 1 RFC1766INFO structure\n");
/* verify the Rfc1766 value */
ret = IMultiLanguage2_GetRfc1766FromLcid(iML2, info.lcid, &rfcstr);
ok(ret == S_OK, "Expected S_OK, got %08x\n", ret);
/* not an exact 1:1 correspondence between lcid and rfc1766 in the
* mlang database, e.g., nb-no -> 1044 -> no */
ok(strstrW(info.wszRfc1766, rfcstr) != NULL,
"Expected matching locale names\n");
SysFreeString(rfcstr);
}
IEnumRfc1766_Release(pEnumRfc1766);
}
static void test_GetLcidFromRfc1766(IMultiLanguage2 *iML2)
{
WCHAR rfc1766W[MAX_RFC1766_NAME + 1];
LCID lcid;
HRESULT ret;
DWORD i;
static WCHAR en[] = { 'e','n',0 };
static WCHAR en_them[] = { 'e','n','-','t','h','e','m',0 };
static WCHAR english[] = { 'e','n','g','l','i','s','h',0 };
for(i = 0; i < sizeof(lcid_table) / sizeof(lcid_table[0]); i++) {
lcid = -1;
MultiByteToWideChar(CP_ACP, 0, lcid_table[i].rfc1766, -1, rfc1766W, MAX_RFC1766_NAME);
ret = IMultiLanguage2_GetLcidFromRfc1766(iML2, &lcid, rfc1766W);
/* IE <6.0 guess 0x412 (ko) from "kok" */
ok((ret == lcid_table[i].hr) ||
broken(lcid_table[i].broken_lcid && (ret == S_FALSE)),
"#%02d: HRESULT 0x%x (expected 0x%x)\n", i, ret, lcid_table[i].hr);
ok((lcid == lcid_table[i].lcid) ||
broken(lcid == lcid_table[i].broken_lcid), /* IE <6.0 */
"#%02d: got LCID 0x%x (expected 0x%x)\n", i, lcid, lcid_table[i].lcid);
}
ret = IMultiLanguage2_GetLcidFromRfc1766(iML2, NULL, en);
ok(ret == E_INVALIDARG, "GetLcidFromRfc1766 returned: %08x\n", ret);
ret = IMultiLanguage2_GetLcidFromRfc1766(iML2, &lcid, NULL);
ok(ret == E_INVALIDARG, "GetLcidFromRfc1766 returned: %08x\n", ret);
ret = IMultiLanguage2_GetLcidFromRfc1766(iML2, &lcid, en_them);
ok((ret == E_FAIL || ret == S_FALSE), "GetLcidFromRfc1766 returned: %08x\n", ret);
if (ret == S_FALSE)
{
BSTR rfcstr;
static WCHAR en[] = {'e','n',0};
ret = IMultiLanguage2_GetRfc1766FromLcid(iML2, lcid, &rfcstr);
ok(ret == S_OK, "Expected S_OK, got %08x\n", ret);
ok_w2("Expected \"%s\", got \"%s\"n", en, rfcstr);
}
ret = IMultiLanguage2_GetLcidFromRfc1766(iML2, &lcid, english);
ok((ret == E_FAIL || ret == S_FALSE), "GetLcidFromRfc1766 returned: %08x\n", ret);
if (ret == S_FALSE)
{
BSTR rfcstr;
static WCHAR en[] = {'e','n',0};
ret = IMultiLanguage2_GetRfc1766FromLcid(iML2, lcid, &rfcstr);
ok(ret == S_OK, "Expected S_OK, got %08x\n", ret);
ok_w2("Expected \"%s\", got \"%s\"n", en, rfcstr);
}
}
static void test_Rfc1766ToLcid(void)
{
LCID lcid;
HRESULT ret;
DWORD i;
for(i = 0; i < sizeof(lcid_table) / sizeof(lcid_table[0]); i++) {
lcid = -1;
ret = pRfc1766ToLcidA(&lcid, lcid_table[i].rfc1766);
/* IE <6.0 guess 0x412 (ko) from "kok" */
ok( (ret == lcid_table[i].hr) ||
broken(lcid_table[i].broken_lcid && (ret == S_FALSE)),
"#%02d: HRESULT 0x%x (expected 0x%x)\n", i, ret, lcid_table[i].hr);
ok( (lcid == lcid_table[i].lcid) ||
broken(lcid == lcid_table[i].broken_lcid), /* IE <6.0 */
"#%02d: got LCID 0x%x (expected 0x%x)\n", i, lcid, lcid_table[i].lcid);
}
ret = pRfc1766ToLcidA(&lcid, NULL);
ok(ret == E_INVALIDARG, "got 0x%08x (expected E_INVALIDARG)\n", ret);
ret = pRfc1766ToLcidA(NULL, "en");
ok(ret == E_INVALIDARG, "got 0x%08x (expected E_INVALIDARG)\n", ret);
}
static void test_GetNumberOfCodePageInfo(IMultiLanguage2 *iML2)
{
HRESULT hr;
UINT value;
value = 0xdeadbeef;
hr = IMultiLanguage2_GetNumberOfCodePageInfo(iML2, &value);
ok( (hr == S_OK) && value,
"got 0x%x with %d (expected S_OK with '!= 0')\n", hr, value);
hr = IMultiLanguage2_GetNumberOfCodePageInfo(iML2, NULL);
ok(hr == E_INVALIDARG, "got 0x%x (expected E_INVALIDARG)\n", hr);
}
static void test_GetRfc1766FromLcid(IMultiLanguage2 *iML2)
{
CHAR expected[MAX_RFC1766_NAME];
CHAR buffer[MAX_RFC1766_NAME + 1];
DWORD i;
HRESULT hr;
BSTR rfcstr;
for(i = 0; i < sizeof(lcid_table) / sizeof(lcid_table[0]); i++) {
buffer[0] = '\0';
rfcstr = NULL;
hr = IMultiLanguage2_GetRfc1766FromLcid(iML2, lcid_table[i].lcid, &rfcstr);
ok(hr == lcid_table[i].hr,
"#%02d: HRESULT 0x%x (expected 0x%x)\n", i, hr, lcid_table[i].hr);
if (hr != S_OK)
continue; /* no result-string created */
WideCharToMultiByte(CP_ACP, 0, rfcstr, -1, buffer, sizeof(buffer), NULL, NULL);
LCMapStringA(LOCALE_USER_DEFAULT, LCMAP_LOWERCASE, lcid_table[i].rfc1766,
lstrlenA(lcid_table[i].rfc1766) + 1, expected, MAX_RFC1766_NAME);
/* IE <6.0 return "x-kok" for LCID 0x457 ("kok") */
ok( (!lstrcmpA(buffer, expected)) ||
broken(!lstrcmpA(buffer, lcid_table[i].broken_rfc)),
"#%02d: got '%s' (expected '%s')\n", i, buffer, expected);
SysFreeString(rfcstr);
}
hr = IMultiLanguage2_GetRfc1766FromLcid(iML2, MAKELANGID(LANG_ENGLISH, SUBLANG_ENGLISH_US), NULL);
ok(hr == E_INVALIDARG, "got 0x%x (expected E_INVALIDARG)\n", hr);
}
static void test_LcidToRfc1766(void)
{
CHAR expected[MAX_RFC1766_NAME];
CHAR buffer[MAX_RFC1766_NAME * 2];
HRESULT hr;
DWORD i;
for(i = 0; i < sizeof(lcid_table) / sizeof(lcid_table[0]); i++) {
memset(buffer, '#', sizeof(buffer)-1);
buffer[sizeof(buffer)-1] = '\0';
hr = pLcidToRfc1766A(lcid_table[i].lcid, buffer, MAX_RFC1766_NAME);
/* IE <5.0 does not recognize 0x180c (fr-mc) and 0x457 (kok) */
ok( (hr == lcid_table[i].hr) ||
broken(lcid_table[i].broken_lcid && (hr == E_FAIL)),
"#%02d: HRESULT 0x%x (expected 0x%x)\n", i, hr, lcid_table[i].hr);
if (hr != S_OK)
continue;
LCMapStringA(LOCALE_USER_DEFAULT, LCMAP_LOWERCASE, lcid_table[i].rfc1766,
lstrlenA(lcid_table[i].rfc1766) + 1, expected, MAX_RFC1766_NAME);
/* IE <6.0 return "x-kok" for LCID 0x457 ("kok") */
/* IE <5.0 return "fr" for LCID 0x180c ("fr-mc") */
ok( (!lstrcmpA(buffer, expected)) ||
broken(!lstrcmpA(buffer, lcid_table[i].broken_rfc)),
"#%02d: got '%s' (expected '%s')\n", i, buffer, expected);
}
memset(buffer, '#', sizeof(buffer)-1);
buffer[sizeof(buffer)-1] = '\0';
hr = pLcidToRfc1766A(-1, buffer, MAX_RFC1766_NAME);
ok(hr == E_FAIL, "got 0x%08x and '%s' (expected E_FAIL)\n", hr, buffer);
hr = pLcidToRfc1766A(LANG_ENGLISH, NULL, MAX_RFC1766_NAME);
ok(hr == E_INVALIDARG, "got 0x%08x (expected E_INVALIDARG)\n", hr);
memset(buffer, '#', sizeof(buffer)-1);
buffer[sizeof(buffer)-1] = '\0';
hr = pLcidToRfc1766A(LANG_ENGLISH, buffer, -1);
ok(hr == E_INVALIDARG, "got 0x%08x and '%s' (expected E_INVALIDARG)\n", hr, buffer);
memset(buffer, '#', sizeof(buffer)-1);
buffer[sizeof(buffer)-1] = '\0';
hr = pLcidToRfc1766A(LANG_ENGLISH, buffer, 0);
ok(hr == E_INVALIDARG, "got 0x%08x and '%s' (expected E_INVALIDARG)\n", hr, buffer);
}
static void test_GetRfc1766Info(IMultiLanguage2 *iML2)
{
WCHAR short_broken_name[MAX_LOCALE_NAME];
CHAR rfc1766A[MAX_RFC1766_NAME + 1];
BYTE buffer[sizeof(RFC1766INFO) + 4];
PRFC1766INFO prfc = (RFC1766INFO *) buffer;
HRESULT ret;
DWORD i;
for(i = 0; i < sizeof(info_table) / sizeof(info_table[0]); i++) {
memset(buffer, 'x', sizeof(RFC1766INFO) + 2);
buffer[sizeof(buffer) -1] = 0;
buffer[sizeof(buffer) -2] = 0;
ret = IMultiLanguage2_GetRfc1766Info(iML2, info_table[i].lcid, info_table[i].lang, prfc);
WideCharToMultiByte(CP_ACP, 0, prfc->wszRfc1766, -1, rfc1766A, MAX_RFC1766_NAME, NULL, NULL);
ok(ret == S_OK, "#%02d: got 0x%x (expected S_OK)\n", i, ret);
ok(prfc->lcid == info_table[i].lcid,
"#%02d: got 0x%04x (expected 0x%04x)\n", i, prfc->lcid, info_table[i].lcid);
ok(!lstrcmpA(rfc1766A, info_table[i].rfc1766),
"#%02d: got '%s' (expected '%s')\n", i, rfc1766A, info_table[i].rfc1766);
/* Some IE versions truncate an oversized name one character too short */
if (info_table[i].broken_name) {
lstrcpyW(short_broken_name, info_table[i].broken_name);
short_broken_name[MAX_LOCALE_NAME - 2] = 0;
}
if (info_table[i].todo & TODO_NAME) {
todo_wine
ok( (!lstrcmpW(prfc->wszLocaleName, info_table[i].localename)) ||
(info_table[i].broken_name && (
broken(!lstrcmpW(prfc->wszLocaleName, info_table[i].broken_name)) || /* IE < 6.0 */
broken(!lstrcmpW(prfc->wszLocaleName, short_broken_name)))),
"#%02d: got %s (expected %s)\n", i,
wine_dbgstr_w(prfc->wszLocaleName), wine_dbgstr_w(info_table[i].localename));
}
else
ok( (!lstrcmpW(prfc->wszLocaleName, info_table[i].localename)) ||
(info_table[i].broken_name && (
broken(!lstrcmpW(prfc->wszLocaleName, info_table[i].broken_name)) || /* IE < 6.0 */
broken(!lstrcmpW(prfc->wszLocaleName, short_broken_name)))),
"#%02d: got %s (expected %s)\n", i,
wine_dbgstr_w(prfc->wszLocaleName), wine_dbgstr_w(info_table[i].localename));
}
/* SUBLANG_NEUTRAL only allowed for English, Arabic, Chinese */
ret = IMultiLanguage2_GetRfc1766Info(iML2, MAKELANGID(LANG_GERMAN, SUBLANG_NEUTRAL), LANG_ENGLISH, prfc);
ok(ret == E_FAIL, "got 0x%x (expected E_FAIL)\n", ret);
ret = IMultiLanguage2_GetRfc1766Info(iML2, MAKELANGID(LANG_ITALIAN, SUBLANG_NEUTRAL), LANG_ENGLISH, prfc);
ok(ret == E_FAIL, "got 0x%x (expected E_FAIL)\n", ret);
/* NULL not allowed */
ret = IMultiLanguage2_GetRfc1766Info(iML2, 0, LANG_ENGLISH, prfc);
ok(ret == E_FAIL, "got 0x%x (expected E_FAIL)\n", ret);
ret = IMultiLanguage2_GetRfc1766Info(iML2, LANG_ENGLISH, LANG_ENGLISH, NULL);
ok(ret == E_INVALIDARG, "got 0x%x (expected E_INVALIDARG)\n", ret);
}
static void test_IMultiLanguage2_ConvertStringFromUnicode(IMultiLanguage2 *iML2)
{
CHAR dest[MAX_PATH];
CHAR invariate[MAX_PATH];
UINT srcsz, destsz;
HRESULT hr;
static WCHAR src[] = {'a','b','c',0};
memset(invariate, 'x', sizeof(invariate));
/* pSrcStr NULL */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src) + 1;
destsz = sizeof(dest);
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, NULL,
&srcsz, dest, &destsz);
ok(hr == S_OK || hr == E_FAIL,"expected S_OK or E_FAIL, got %08x\n",hr);
if (hr == S_OK)
{
ok(srcsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, srcsz);
}
else if (hr == E_FAIL)
{
ok(srcsz == 0,
"Expected %u, got %u\n", 0, srcsz);
}
ok(!memcmp(dest, invariate, sizeof(dest)),
"Expected dest to be unchanged, got %s\n", dest);
ok(destsz == 0, "Expected 0, got %u\n", destsz);
/* pcSrcSize NULL */
memset(dest, 'x', sizeof(dest));
destsz = sizeof(dest);
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, src,
NULL, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(!strncmp(dest, "abc", 3),
"Expected first three chars to be \"abc\"\n");
ok(!memcmp(&dest[3], invariate, sizeof(dest) - 3),
"Expected rest of dest to be unchanged, got %s\n", dest);
ok(destsz == lstrlenW(src),
"Expected %u, got %u\n", lstrlenW(src), destsz);
/* both pSrcStr and pcSrcSize NULL */
memset(dest, 'x', sizeof(dest));
destsz = sizeof(dest);
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, NULL,
NULL, dest, &destsz);
ok(hr == S_OK || hr == E_FAIL, "Expected S_OK or E_FAIL, got %08x\n", hr);
ok(!memcmp(dest, invariate, sizeof(dest)),
"Expected dest to be unchanged, got %s\n", dest);
ok(destsz == 0, "Expected 0, got %u\n", destsz);
/* pDstStr NULL */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src) + 1;
destsz = sizeof(dest);
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, src,
&srcsz, NULL, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, srcsz);
ok(destsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, srcsz);
/* pcDstSize NULL */
memset(dest, 'x', sizeof(dest));
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, src,
&srcsz, dest, NULL);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, srcsz);
ok(!memcmp(dest, invariate, sizeof(dest)),
"Expected dest to be unchanged, got %s\n", dest);
/* pcSrcSize is 0 */
memset(dest, 'x', sizeof(dest));
srcsz = 0;
destsz = sizeof(dest);
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, src,
&srcsz, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == 0, "Expected 0, got %u\n", srcsz);
ok(!memcmp(dest, invariate, sizeof(dest)),
"Expected dest to be unchanged, got %s\n", dest);
ok(destsz == 0, "Expected 0, got %u\n", destsz);
/* pcSrcSize does not include NULL terminator */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src);
destsz = sizeof(dest);
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, src,
&srcsz, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src),
"Expected %u, got %u\n", lstrlenW(src), srcsz);
ok(!strncmp(dest, "abc", 3), "Expected first three chars to be \"abc\"\n");
ok(!memcmp(&dest[3], invariate, sizeof(dest) - 3),
"Expected rest of dest to be unchanged, got %s\n", dest);
ok(destsz == lstrlenW(src),
"Expected %u, got %u\n", lstrlenW(src), destsz);
/* pcSrcSize includes NULL terminator */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src) + 1;
destsz = sizeof(dest);
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, src,
&srcsz, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src) + 1, "Expected 3, got %u\n", srcsz);
ok(!lstrcmpA(dest, "abc"), "Expected \"abc\", got \"%s\"\n", dest);
ok(destsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, destsz);
/* pcSrcSize is -1 */
memset(dest, 'x', sizeof(dest));
srcsz = -1;
destsz = sizeof(dest);
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, src,
&srcsz, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src),
"Expected %u, got %u\n", lstrlenW(src), srcsz);
ok(!strncmp(dest, "abc", 3), "Expected first three chars to be \"abc\"\n");
ok(!memcmp(&dest[3], invariate, sizeof(dest) - 3),
"Expected rest of dest to be unchanged, got %s\n", dest);
ok(destsz == lstrlenW(src),
"Expected %u, got %u\n", lstrlenW(src), destsz);
/* pcDstSize is 0 */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src) + 1;
destsz = 0;
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, src,
&srcsz, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, srcsz);
ok(!memcmp(dest, invariate, sizeof(dest)),
"Expected dest to be unchanged, got %s\n", dest);
ok(destsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, destsz);
/* pcDstSize is not large enough */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src) + 1;
destsz = lstrlenW(src);
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, src,
&srcsz, dest, &destsz);
ok(hr == E_FAIL, "Expected E_FAIL, got %08x\n", hr);
ok(srcsz == 0, "Expected 0, got %u\n", srcsz);
ok(!strncmp(dest, "abc", 3), "Expected first three chars to be \"abc\"\n");
ok(!memcmp(&dest[3], invariate, sizeof(dest) - 3),
"Expected rest of dest to be unchanged, got %s\n", dest);
ok(destsz == 0, "Expected 0, got %u\n", srcsz);
/* pcDstSize (bytes) does not leave room for NULL terminator */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src) + 1;
destsz = lstrlenW(src) * sizeof(WCHAR);
hr = IMultiLanguage2_ConvertStringFromUnicode(iML2, NULL, 1252, src,
&srcsz, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, srcsz);
ok(!lstrcmpA(dest, "abc"), "Expected \"abc\", got \"%s\"\n", dest);
ok(destsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, destsz);
}
static void test_ConvertINetUnicodeToMultiByte(void)
{
CHAR dest[MAX_PATH];
CHAR invariate[MAX_PATH];
INT srcsz, destsz;
HRESULT hr;
static WCHAR src[] = {'a','b','c',0};
memset(invariate, 'x', sizeof(invariate));
/* lpSrcStr NULL */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src) + 1;
destsz = sizeof(dest);
hr = pConvertINetUnicodeToMultiByte(NULL, 1252, NULL, &srcsz, dest, &destsz);
ok(hr == S_OK || hr == E_FAIL, "Expected S_OK or E_FAIL, got %08x\n", hr);
if (hr == S_OK)
ok(srcsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, srcsz);
else if (hr == E_FAIL)
ok(srcsz == 0,
"Expected %u, got %u\n", 0, srcsz);
ok(!memcmp(dest, invariate, sizeof(dest)),
"Expected dest to be unchanged, got %s\n", dest);
ok(destsz == 0, "Expected 0, got %u\n", destsz);
/* lpnWideCharCount NULL */
memset(dest, 'x', sizeof(dest));
destsz = sizeof(dest);
hr = pConvertINetUnicodeToMultiByte(NULL, 1252, src, NULL, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(!strncmp(dest, "abc", 3),
"Expected first three chars to be \"abc\"\n");
ok(!memcmp(&dest[3], invariate, sizeof(dest) - 3),
"Expected rest of dest to be unchanged, got %s\n", dest);
ok(destsz == lstrlenW(src),
"Expected %u, got %u\n", lstrlenW(src), destsz);
/* both lpSrcStr and lpnWideCharCount NULL */
memset(dest, 'x', sizeof(dest));
destsz = sizeof(dest);
hr = pConvertINetUnicodeToMultiByte(NULL, 1252, NULL, NULL, dest, &destsz);
ok(hr == S_OK || hr == E_FAIL, "Expected S_OK or E_FAIL, got %08x\n", hr);
ok(!memcmp(dest, invariate, sizeof(dest)),
"Expected dest to be unchanged, got %s\n", dest);
ok(destsz == 0, "Expected 0, got %u\n", destsz);
/* lpDstStr NULL */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src) + 1;
destsz = sizeof(dest);
hr = pConvertINetUnicodeToMultiByte(NULL, 1252, src, &srcsz, NULL, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, srcsz);
ok(destsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, srcsz);
/* lpnMultiCharCount NULL */
memset(dest, 'x', sizeof(dest));
hr = pConvertINetUnicodeToMultiByte(NULL, 1252, src, &srcsz, dest, NULL);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, srcsz);
ok(!memcmp(dest, invariate, sizeof(dest)),
"Expected dest to be unchanged, got %s\n", dest);
/* lpnWideCharCount is 0 */
memset(dest, 'x', sizeof(dest));
srcsz = 0;
destsz = sizeof(dest);
hr = pConvertINetUnicodeToMultiByte(NULL, 1252, src, &srcsz, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == 0, "Expected 0, got %u\n", srcsz);
ok(!memcmp(dest, invariate, sizeof(dest)),
"Expected dest to be unchanged, got %s\n", dest);
ok(destsz == 0, "Expected 0, got %u\n", destsz);
/* lpnWideCharCount does not include NULL terminator */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src);
destsz = sizeof(dest);
hr = pConvertINetUnicodeToMultiByte(NULL, 1252, src, &srcsz, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src),
"Expected %u, got %u\n", lstrlenW(src), srcsz);
ok(!strncmp(dest, "abc", 3), "Expected first three chars to be \"abc\"\n");
ok(!memcmp(&dest[3], invariate, sizeof(dest) - 3),
"Expected rest of dest to be unchanged, got %s\n", dest);
ok(destsz == lstrlenW(src),
"Expected %u, got %u\n", lstrlenW(src), destsz);
/* lpnWideCharCount includes NULL terminator */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src) + 1;
destsz = sizeof(dest);
hr = pConvertINetUnicodeToMultiByte(NULL, 1252, src, &srcsz, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src) + 1, "Expected 3, got %u\n", srcsz);
ok(!lstrcmpA(dest, "abc"), "Expected \"abc\", got \"%s\"\n", dest);
ok(destsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, destsz);
/* lpnWideCharCount is -1 */
memset(dest, 'x', sizeof(dest));
srcsz = -1;
destsz = sizeof(dest);
hr = pConvertINetUnicodeToMultiByte(NULL, 1252, src, &srcsz, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src),
"Expected %u, got %u\n", lstrlenW(src), srcsz);
ok(!strncmp(dest, "abc", 3), "Expected first three chars to be \"abc\"\n");
ok(!memcmp(&dest[3], invariate, sizeof(dest) - 3),
"Expected rest of dest to be unchanged, got %s\n", dest);
ok(destsz == lstrlenW(src),
"Expected %u, got %u\n", lstrlenW(src), destsz);
/* lpnMultiCharCount is 0 */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src) + 1;
destsz = 0;
hr = pConvertINetUnicodeToMultiByte(NULL, 1252, src, &srcsz, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, srcsz);
ok(!memcmp(dest, invariate, sizeof(dest)),
"Expected dest to be unchanged, got %s\n", dest);
ok(destsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, destsz);
/* lpnMultiCharCount is not large enough */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src) + 1;
destsz = lstrlenW(src);
hr = pConvertINetUnicodeToMultiByte(NULL, 1252, src, &srcsz, dest, &destsz);
ok(hr == E_FAIL, "Expected E_FAIL, got %08x\n", hr);
ok(srcsz == 0, "Expected 0, got %u\n", srcsz);
ok(!strncmp(dest, "abc", 3), "Expected first three chars to be \"abc\"\n");
ok(!memcmp(&dest[3], invariate, sizeof(dest) - 3),
"Expected rest of dest to be unchanged, got %s\n", dest);
ok(destsz == 0, "Expected 0, got %u\n", srcsz);
/* lpnMultiCharCount (bytes) does not leave room for NULL terminator */
memset(dest, 'x', sizeof(dest));
srcsz = lstrlenW(src) + 1;
destsz = lstrlenW(src) * sizeof(WCHAR);
hr = pConvertINetUnicodeToMultiByte(NULL, 1252, src, &srcsz, dest, &destsz);
ok(hr == S_OK, "Expected S_OK, got %08x\n", hr);
ok(srcsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, srcsz);
ok(!lstrcmpA(dest, "abc"), "Expected \"abc\", got \"%s\"\n", dest);
ok(destsz == lstrlenW(src) + 1,
"Expected %u, got %u\n", lstrlenW(src) + 1, destsz);
}
static void test_JapaneseConversion(void)
{
/* Data */
static WCHAR unc_jp[9][12] = {
{9,0x31,0x20,0x3042,0x3044,0x3046,0x3048,0x304a,0x000d,0x000a},
{9,0x31,0x20,0x30a2,0x30a4,0x30a6,0x30a8,0x30aa,0x000d,0x000a},
{9,0x31,0x20,0xff71,0xff72,0xff73,0xff74,0xff75,0x000d,0x000a},
{9,0x31,0x20,0x3041,0x3043,0x3045,0x3047,0x3049,0x000d,0x000a},
{9,0x31,0x20,0x30a1,0x30a3,0x30a5,0x30a7,0x30a9,0x000d,0x000a},
{9,0x31,0x20,0xff67,0xff68,0xff69,0xff6a,0xff6b,0x000d,0x000a},
{9,0x31,0x20,0x300c,0x65e5,0x672c,0x8a9e,0x300d,0x000d,0x000a},
{7,0x31,0x20,0x25c7,0x25c7,0x3012,0x000d,0x000a},
{11,0x31,0x20,0x203b,0x3010,0x0074,0x0065,0x0073,0x0074,0x3011,0x000d,0x000a}
};
static CHAR jis_jp[9][27] = {
{20,0x31,0x20,0x1b,0x24,0x42,0x24,0x22,0x24,0x24,0x24,0x26,0x24,0x28,
0x24,0x2a,0x1b,0x28,0x42,0x0d,0x0a},
{20,0x31,0x20,0x1b,0x24,0x42,0x25,0x22,0x25,0x24,0x25,0x26,0x25,0x28,
0x25,0x2a,0x1b,0x28,0x42,0x0d,0x0a},
{20,0x31,0x20,0x1b,0x24,0x42,0x25,0x22,0x25,0x24,0x25,0x26,0x25,0x28,
0x25,0x2a,0x1b,0x28,0x42,0x0d,0x0a},
{20,0x31,0x20,0x1b,0x24,0x42,0x24,0x21,0x24,0x23,0x24,0x25,0x24,0x27,
0x24,0x29,0x1b,0x28,0x42,0x0d,0x0a},
{20,0x31,0x20,0x1b,0x24,0x42,0x25,0x21,0x25,0x23,0x25,0x25,0x25,0x27,
0x25,0x29,0x1b,0x28,0x42,0x0d,0x0a},
{20,0x31,0x20,0x1b,0x24,0x42,0x25,0x21,0x25,0x23,0x25,0x25,0x25,0x27,
0x25,0x29,0x1b,0x28,0x42,0x0d,0x0a},
{20,0x31,0x20,0x1b,0x24,0x42,0x21,0x56,0x46,0x7c,0x4b,0x5c,0x38,0x6c,
0x21,0x57,0x1b,0x28,0x42,0x0d,0x0a},
{16,0x31,0x20,0x1b,0x24,0x42,0x21,0x7e,0x21,0x7e,0x22,0x29,0x1b,0x28,
0x42,0x0d,0x0a},
{26,0x31,0x20,0x1b,0x24,0x42,0x22,0x28,0x21,0x5a,0x1b,0x28,0x42,0x74,
0x65,0x73,0x74,0x1b,0x24,0x42,0x21,0x5b,0x1b,0x28,0x42,0x0d,0x0a}
};
static CHAR sjis_jp[9][15] = {
{14,0x31,0x20,0x82,0xa0,0x82,0xa2,0x82,0xa4,0x82,0xa6,0x82,0xa8,0x0d,0x0a},
{14,0x31,0x20,0x83,0x41,0x83,0x43,0x83,0x45,0x83,0x47,0x83,0x49,0x0d,0x0a},
{9,0x31,0x20,0xb1,0xb2,0xb3,0xb4,0xb5,0x0d,0x0a},
{14,0x31,0x20,0x82,0x9f,0x82,0xa1,0x82,0xa3,0x82,0xa5,0x82,0xa7,0x0d,0x0a},
{14,0x31,0x20,0x83,0x40,0x83,0x42,0x83,0x44,0x83,0x46,0x83,0x48,0x0d,0x0a},
{9,0x31,0x20,0xa7,0xa8,0xa9,0xaa,0xab,0x0d,0x0a},
{14,0x31,0x20,0x81,0x75,0x93,0xfa,0x96,0x7b,0x8c,0xea,0x81,0x76,0x0d,0x0a},
{10,0x31,0x20,0x81,0x9e,0x81,0x9e,0x81,0xa7,0x0d,0x0a},
{14,0x31,0x20,0x81,0xa6,0x81,0x79,0x74,0x65,0x73,0x74,0x81,0x7a,0x0d,0x0a}
};
static CHAR euc_jp[9][15] = {
{14,0x31,0x20,0xa4,0xa2,0xa4,0xa4,0xa4,0xa6,0xa4,0xa8,0xa4,0xaa,0x0d,0x0a},
{14,0x31,0x20,0xa5,0xa2,0xa5,0xa4,0xa5,0xa6,0xa5,0xa8,0xa5,0xaa,0x0d,0x0a},
{14,0x31,0x20,0x8e,0xb1,0x8e,0xb2,0x8e,0xb3,0x8e,0xb4,0x8e,0xb5,0x0d,0x0a},
{14,0x31,0x20,0xa4,0xa1,0xa4,0xa3,0xa4,0xa5,0xa4,0xa7,0xa4,0xa9,0x0d,0x0a},
{14,0x31,0x20,0xa5,0xa1,0xa5,0xa3,0xa5,0xa5,0xa5,0xa7,0xa5,0xa9,0x0d,0x0a},
{14,0x31,0x20,0x8e,0xa7,0x8e,0xa8,0x8e,0xa9,0x8e,0xaa,0x8e,0xab,0x0d,0x0a},
{14,0x31,0x20,0xa1,0xd6,0xc6,0xfc,0xcb,0xdc,0xb8,0xec,0xa1,0xd7,0x0d,0x0a},
{10,0x31,0x20,0xa1,0xfe,0xa1,0xfe,0xa2,0xa9,0x0d,0x0a},
{14,0x31,0x20,0xa2,0xa8,0xa1,0xda,0x74,0x65,0x73,0x74,0xa1,0xdb,0x0d,0x0a}
};
INT srcsz, destsz;
INT i;
HRESULT hr;
CHAR output[30];
WCHAR outputW[30];
int outlen;
/* test unc->jis */
for (i = 0; i < 9; i++)
{
int j;
destsz = 30;
outlen = jis_jp[i][0];
srcsz = unc_jp[i][0];
hr = pConvertINetUnicodeToMultiByte(NULL, 50220, &unc_jp[i][1], &srcsz, output, &destsz);
if (hr == S_FALSE)
{
skip("Code page identifier 50220 is not supported\n");
break;
}
ok(hr == S_OK,"(%i) Expected S_OK, got %08x\n", i, hr);
ok(destsz == outlen, "(%i) Expected %i, got %i\n",i,outlen,destsz);
ok(srcsz == unc_jp[i][0],"(%i) Expected %i, got %i\n",i,unc_jp[i][0],srcsz);
ok(memcmp(output,&jis_jp[i][1],destsz)==0,"(%i) Strings do not match\n",i);
/* and back */
srcsz = outlen;
destsz = 30;
hr = pConvertINetMultiByteToUnicode(NULL, 50220, output, &srcsz, outputW,&destsz);
/*
* JIS does not have hankata so it get automatically converted to
* zenkata. this means that strings 1 and 2 are identical as well as
* strings 4 and 5.
*/
j = i;
if (i == 2) j = 1;
if (i == 5) j = 4;
ok(hr == S_OK,"(%i) Expected S_OK, got %08x\n",i, hr);
ok(destsz == unc_jp[j][0],"(%i) Expected %i, got %i\n",i,unc_jp[j][0],destsz);
ok(srcsz == outlen,"(%i) Expected %i, got %i\n",i,outlen,srcsz);
ok(memcmp(outputW,&unc_jp[j][1],destsz)==0,"(%i) Strings do not match\n",i);
}
/* test unc->sjis */
for (i = 0; i < 9; i++)
{
destsz = 30;
outlen = sjis_jp[i][0];
srcsz = unc_jp[i][0];
hr = pConvertINetUnicodeToMultiByte(NULL, 932, &unc_jp[i][1], &srcsz, output, &destsz);
if (hr == S_FALSE)
{
skip("Code page identifier 932 is not supported\n");
break;
}
ok(hr == S_OK,"(%i) Expected S_OK, got %08x\n",i,hr);
ok(destsz == outlen,"(%i) Expected %i, got %i\n",i,outlen,destsz);
ok(srcsz == unc_jp[i][0],"(%i) Expected %i, got %i\n",i,unc_jp[i][0],srcsz);
ok(memcmp(output,&sjis_jp[i][1],outlen)==0,"(%i) Strings do not match\n",i);
srcsz = outlen;
destsz = 30;
hr = pConvertINetMultiByteToUnicode(NULL, 932, output, &srcsz, outputW,&destsz);
ok(hr == S_OK,"(%i) Expected S_OK, got %08x\n", i, hr);
ok(destsz == unc_jp[i][0],"(%i) Expected %i, got %i\n",i,unc_jp[i][0],destsz);
ok(srcsz == outlen,"(%i) Expected %i, got %i\n",i,outlen,srcsz);
ok(memcmp(outputW,&unc_jp[i][1],destsz)==0,"(%i) Strings do not match\n",i);
}
/* test unc->euc */
for (i = 0; i < 9; i++)
{
destsz = 30;
outlen = euc_jp[i][0];
srcsz = unc_jp[i][0];
hr = pConvertINetUnicodeToMultiByte(NULL, 51932, &unc_jp[i][1], &srcsz, output, &destsz);
if (hr == S_FALSE)
{
skip("Code page identifier 51932 is not supported\n");
break;
}
ok(hr == S_OK, "(%i) Expected S_OK, got %08x\n",i,hr);
ok(destsz == outlen, "(%i) Expected %i, got %i\n",i,outlen,destsz);
ok(srcsz == unc_jp[i][0],"(%i) Expected %i, got %i\n",i,unc_jp[i][0],destsz);
ok(memcmp(output,&euc_jp[i][1],outlen)==0,"(%i) Strings do not match\n",i);
srcsz = outlen;
destsz = 30;
hr = pConvertINetMultiByteToUnicode(NULL, 51932, output, &srcsz, outputW,&destsz);
ok(hr == S_OK,"(%i) Expected S_OK, got %08x\n",i,hr);
ok(destsz == unc_jp[i][0],"(%i) Expected %i, got %i\n",i,unc_jp[i][0],destsz);
ok(srcsz == outlen,"(%i) Expected %i, got %i\n",i,outlen,srcsz);
ok(memcmp(outputW,&unc_jp[i][1],destsz)==0,"(%i) Strings do not match\n",i);
}
/* Japanese autodetect */
i = 0;
destsz = 30;
srcsz = jis_jp[i][0];
hr = pConvertINetMultiByteToUnicode(NULL, 50932, &jis_jp[i][1], &srcsz, outputW, &destsz);
if (hr == S_FALSE)
{
skip("Code page identifier 50932 is not supported\n");
return;
}
ok(hr == S_OK,"(%i) Expected S_OK, got %08x\n",i,hr);
ok(destsz == unc_jp[i][0],"(%i) Expected %i, got %i\n",i,unc_jp[i][0],destsz);
ok(srcsz == jis_jp[i][0],"(%i) Expected %i, got %i\n",i,jis_jp[i][0],srcsz);
ok(memcmp(outputW,&unc_jp[i][1],destsz)==0,"(%i) Strings do not match\n",i);
i = 1;
destsz = 30;
srcsz = sjis_jp[i][0];
hr = pConvertINetMultiByteToUnicode(NULL, 50932, &sjis_jp[i][1], &srcsz, outputW, &destsz);
ok(hr == S_OK,"(%i) Expected S_OK, got %08x\n",i,hr);
ok(destsz == unc_jp[i][0],"(%i) Expected %i, got %i\n",i,unc_jp[i][0],destsz);
ok(srcsz == sjis_jp[i][0],"(%i) Expected %i, got %i\n",i,sjis_jp[i][0],srcsz);
ok(memcmp(outputW,&unc_jp[i][1],destsz)==0,"(%i) Strings do not match\n",i);
}
static void test_GetScriptFontInfo(IMLangFontLink2 *font_link)
{
HRESULT hr;
UINT nfonts;
SCRIPTFONTINFO sfi[1];
nfonts = 0;
hr = IMLangFontLink2_GetScriptFontInfo(font_link, sidAsciiLatin, 0, &nfonts, NULL);
ok(hr == S_OK, "GetScriptFontInfo failed %u\n", GetLastError());
ok(nfonts, "unexpected result\n");
nfonts = 0;
hr = IMLangFontLink2_GetScriptFontInfo(font_link, sidAsciiLatin, SCRIPTCONTF_FIXED_FONT, &nfonts, NULL);
ok(hr == S_OK, "GetScriptFontInfo failed %u\n", GetLastError());
ok(nfonts, "unexpected result\n");
nfonts = 0;
hr = IMLangFontLink2_GetScriptFontInfo(font_link, sidAsciiLatin, SCRIPTCONTF_PROPORTIONAL_FONT, &nfonts, NULL);
ok(hr == S_OK, "GetScriptFontInfo failed %u\n", GetLastError());
ok(nfonts, "unexpected result\n");
nfonts = 1;
memset(sfi, 0, sizeof(sfi));
hr = IMLangFontLink2_GetScriptFontInfo(font_link, sidAsciiLatin, SCRIPTCONTF_FIXED_FONT, &nfonts, sfi);
ok(hr == S_OK, "GetScriptFontInfo failed %u\n", GetLastError());
ok(nfonts == 1, "got %u, expected 1\n", nfonts);
ok(sfi[0].scripts != 0, "unexpected result\n");
ok(sfi[0].wszFont[0], "unexpected result\n");
nfonts = 1;
memset(sfi, 0, sizeof(sfi));
hr = IMLangFontLink2_GetScriptFontInfo(font_link, sidAsciiLatin, SCRIPTCONTF_PROPORTIONAL_FONT, &nfonts, sfi);
ok(hr == S_OK, "GetScriptFontInfo failed %u\n", GetLastError());
ok(nfonts == 1, "got %u, expected 1\n", nfonts);
ok(sfi[0].scripts != 0, "unexpected result\n");
ok(sfi[0].wszFont[0], "unexpected result\n");
}
static void test_CodePageToScriptID(IMLangFontLink2 *font_link)
{
HRESULT hr;
UINT i;
SCRIPT_ID sid;
static const struct
{
UINT cp;
SCRIPT_ID sid;
HRESULT hr;
}
cp_sid[] =
{
{874, sidThai},
{932, sidKana},
{936, sidHan},
{949, sidHangul},
{950, sidBopomofo},
{1250, sidAsciiLatin},
{1251, sidCyrillic},
{1252, sidAsciiLatin},
{1253, sidGreek},
{1254, sidAsciiLatin},
{1255, sidHebrew},
{1256, sidArabic},
{1257, sidAsciiLatin},
{1258, sidAsciiLatin},
{CP_UNICODE, 0, E_FAIL}
};
for (i = 0; i < sizeof(cp_sid)/sizeof(cp_sid[0]); i++)
{
hr = IMLangFontLink2_CodePageToScriptID(font_link, cp_sid[i].cp, &sid);
ok(hr == cp_sid[i].hr, "%u CodePageToScriptID failed 0x%08x %u\n", i, hr, GetLastError());
if (SUCCEEDED(hr))
{
ok(sid == cp_sid[i].sid,
"%u got sid %u for codepage %u, expected %u\n", i, sid, cp_sid[i].cp, cp_sid[i].sid);
}
}
}
static void test_GetFontUnicodeRanges(IMLangFontLink2 *font_link)
{
HRESULT hr;
UINT count;
HFONT hfont, old_hfont;
LOGFONTA lf;
HDC hdc;
UNICODERANGE *ur;
hdc = CreateCompatibleDC(0);
memset(&lf, 0, sizeof(lf));
lstrcpyA(lf.lfFaceName, "Arial");
hfont = CreateFontIndirectA(&lf);
old_hfont = SelectObject(hdc, hfont);
count = 0;
hr = IMLangFontLink2_GetFontUnicodeRanges(font_link, NULL, &count, NULL);
ok(hr == E_FAIL, "expected E_FAIL, got 0x%08x\n", hr);
hr = IMLangFontLink2_GetFontUnicodeRanges(font_link, hdc, NULL, NULL);
ok(hr == E_INVALIDARG, "expected E_FAIL, got 0x%08x\n", hr);
count = 0;
hr = IMLangFontLink2_GetFontUnicodeRanges(font_link, hdc, &count, NULL);
ok(hr == S_OK, "expected S_OK, got 0x%08x\n", hr);
ok(count, "expected count > 0\n");
ur = HeapAlloc(GetProcessHeap(), 0, sizeof(*ur) * count);
hr = IMLangFontLink2_GetFontUnicodeRanges(font_link, hdc, &count, ur);
ok(hr == S_OK, "expected S_OK, got 0x%08x\n", hr);
count--;
hr = IMLangFontLink2_GetFontUnicodeRanges(font_link, hdc, &count, ur);
ok(hr == S_OK, "expected S_OK, got 0x%08x\n", hr);
HeapFree(GetProcessHeap(), 0, ur);
SelectObject(hdc, old_hfont);
DeleteObject(hfont);
DeleteDC(hdc);
}
static void test_IsCodePageInstallable(IMultiLanguage2 *ml2)
{
UINT i;
HRESULT hr;
for (i = 0; i < 0xffff; i++)
{
hr = IMultiLanguage2_IsCodePageInstallable(ml2, i);
/* it would be better to use IMultiLanguage2_ValidateCodePageEx here but that brings
* up an installation dialog on some platforms, even when specifying CPIOD_PEEK.
*/
if (IsValidCodePage(i))
ok(hr == S_OK ||
broken(hr == S_FALSE) || /* win2k */
broken(hr == E_INVALIDARG), /* win2k */
"code page %u is valid but not installable 0x%08x\n", i, hr);
}
}
static void test_GetGlobalFontLinkObject(void)
{
HRESULT ret;
void *unknown;
ret = GetGlobalFontLinkObject(NULL);
ok(ret == E_INVALIDARG, "expected E_INVALIDARG got %#x\n", ret);
unknown = (void *)0xdeadbeef;
ret = GetGlobalFontLinkObject(&unknown);
todo_wine {
ok(ret == S_OK, "expected S_OK got %#x\n", ret);
ok(unknown != NULL && unknown != (void *)0xdeadbeef,
"GetGlobalFontLinkObject() returned %p\n", unknown);
}
}
static void test_IMLangConvertCharset(IMultiLanguage *ml)
{
IMLangConvertCharset *convert;
WCHAR strW[] = {'a','b','c','d',0};
UINT cp, src_size, dst_size;
char strA[] = "abcd";
WCHAR buffW[20];
HRESULT hr;
hr = IMultiLanguage_CreateConvertCharset(ml, CP_ACP, CP_UTF8, 0, &convert);
todo_wine
ok(hr == S_FALSE, "expected S_FALSE got 0x%08x\n", hr);
hr = IMLangConvertCharset_GetSourceCodePage(convert, NULL);
ok(hr == E_INVALIDARG, "got 0x%08x\n", hr);
cp = CP_UTF8;
hr = IMLangConvertCharset_GetSourceCodePage(convert, &cp);
ok(hr == S_OK, "expected S_OK got 0x%08x\n", hr);
ok(cp == CP_ACP, "got %d\n", cp);
hr = IMLangConvertCharset_GetDestinationCodePage(convert, NULL);
ok(hr == E_INVALIDARG, "got 0x%08x\n", hr);
cp = CP_ACP;
hr = IMLangConvertCharset_GetDestinationCodePage(convert, &cp);
ok(hr == S_OK, "expected S_OK got 0x%08x\n", hr);
ok(cp == CP_UTF8, "got %d\n", cp);
/* DoConversionToUnicode */
hr = IMLangConvertCharset_Initialize(convert, CP_UTF8, CP_UNICODE, 0);
ok(hr == S_OK, "expected S_OK got 0x%08x\n", hr);
hr = IMLangConvertCharset_DoConversionToUnicode(convert, NULL, NULL, NULL, NULL);
ok(hr == E_FAIL || broken(hr == S_OK) /* win2k */, "got 0x%08x\n", hr);
src_size = -1;
dst_size = 20;
buffW[0] = 0;
buffW[4] = 4;
hr = IMLangConvertCharset_DoConversionToUnicode(convert, strA, &src_size, buffW, &dst_size);
ok(hr == S_OK, "got 0x%08x\n", hr);
ok(!memcmp(buffW, strW, 4*sizeof(WCHAR)), "got converted string %s\n", wine_dbgstr_wn(buffW, dst_size));
ok(dst_size == 4, "got %d\n", dst_size);
ok(buffW[4] == 4, "got %d\n", buffW[4]);
ok(src_size == 4, "got %d\n", src_size);
src_size = -1;
dst_size = 0;
buffW[0] = 1;
hr = IMLangConvertCharset_DoConversionToUnicode(convert, strA, &src_size, buffW, &dst_size);
ok(hr == S_OK, "got 0x%08x\n", hr);
ok(buffW[0] == 1, "got %d\n", buffW[0]);
ok(dst_size == 4, "got %d\n", dst_size);
ok(src_size == 4, "got %d\n", src_size);
hr = IMLangConvertCharset_Initialize(convert, CP_UNICODE, CP_UNICODE, 0);
ok(hr == S_OK, "expected S_OK got 0x%08x\n", hr);
IMLangConvertCharset_Release(convert);
}
START_TEST(mlang)
{
IMultiLanguage *iML = NULL;
IMultiLanguage2 *iML2 = NULL;
IMLangFontLink *iMLFL = NULL;
IMLangFontLink2 *iMLFL2 = NULL;
HRESULT ret;
if (!init_function_ptrs())
return;
CoInitialize(NULL);
test_Rfc1766ToLcid();
test_LcidToRfc1766();
test_ConvertINetUnicodeToMultiByte();
test_JapaneseConversion();
test_GetGlobalFontLinkObject();
trace("IMultiLanguage\n");
ret = CoCreateInstance(&CLSID_CMultiLanguage, NULL, CLSCTX_INPROC_SERVER,
&IID_IMultiLanguage, (void **)&iML);
if (ret != S_OK || !iML) return;
test_GetNumberOfCodePageInfo((IMultiLanguage2 *)iML);
test_IMLangConvertCharset(iML);
IMultiLanguage_Release(iML);
/* IMultiLanguage2 (IE5.0 and above) */
trace("IMultiLanguage2\n");
ret = CoCreateInstance(&CLSID_CMultiLanguage, NULL, CLSCTX_INPROC_SERVER,
&IID_IMultiLanguage2, (void **)&iML2);
if (ret != S_OK || !iML2) return;
test_rfc1766(iML2);
test_GetLcidFromRfc1766(iML2);
test_GetRfc1766FromLcid(iML2);
test_GetRfc1766Info(iML2);
test_GetNumberOfCodePageInfo(iML2);
test_EnumCodePages(iML2, 0);
test_EnumCodePages(iML2, MIMECONTF_MIME_LATEST);
test_EnumCodePages(iML2, MIMECONTF_BROWSER);
test_EnumCodePages(iML2, MIMECONTF_MINIMAL);
test_EnumCodePages(iML2, MIMECONTF_VALID);
/* FIXME: why MIMECONTF_MIME_REGISTRY returns 0 of supported codepages? */
/*test_EnumCodePages(iML2, MIMECONTF_MIME_REGISTRY);*/
test_EnumScripts(iML2, 0);
test_EnumScripts(iML2, SCRIPTCONTF_SCRIPT_USER);
test_EnumScripts(iML2, SCRIPTCONTF_SCRIPT_USER | SCRIPTCONTF_SCRIPT_HIDE | SCRIPTCONTF_SCRIPT_SYSTEM);
ret = IMultiLanguage2_IsConvertible(iML2, CP_UTF8, CP_UNICODE);
ok(ret == S_OK, "IMultiLanguage2_IsConvertible(CP_UTF8 -> CP_UNICODE) = %08x\n", ret);
ret = IMultiLanguage2_IsConvertible(iML2, CP_UNICODE, CP_UTF8);
ok(ret == S_OK, "IMultiLanguage2_IsConvertible(CP_UNICODE -> CP_UTF8) = %08x\n", ret);
test_multibyte_to_unicode_translations(iML2);
test_IMultiLanguage2_ConvertStringFromUnicode(iML2);
test_IsCodePageInstallable(iML2);
IMultiLanguage2_Release(iML2);
/* IMLangFontLink */
ret = CoCreateInstance(&CLSID_CMultiLanguage, NULL, CLSCTX_INPROC_SERVER,
&IID_IMLangFontLink, (void **)&iMLFL);
if (ret != S_OK || !iMLFL) return;
IMLangFontLink_Test(iMLFL);
IMLangFontLink_Release(iMLFL);
/* IMLangFontLink2 */
ret = CoCreateInstance(&CLSID_CMultiLanguage, NULL, CLSCTX_INPROC_SERVER,
&IID_IMLangFontLink2, (void **)&iMLFL2);
if (ret != S_OK || !iMLFL2) return;
test_GetScriptFontInfo(iMLFL2);
test_GetFontUnicodeRanges(iMLFL2);
test_CodePageToScriptID(iMLFL2);
IMLangFontLink2_Release(iMLFL2);
CoUninitialize();
}