November 2013 - Wine-devel - list.winehq.org

Re: kernel32: use ANSI code page if current thread locale has no code page (try 3)
by Alexandre Julliard 05 Nov '13

05 Nov '13

Daniel Lehman <dlehman(a)esri.com> writes: > +static void test_threadcp(void) > +{ > + CPINFOEXA cpi; > + int i, ret; > + UINT cp; > + LCID last; > + struct test > + { > + LCID lcid; > + UINT defacp; > + UINT threadcp; > + } lcids[] = > + { > + { 0x0409 /* 1033 - English - United States */, 1252, 1252 /* ANSI Latin 1 */}, > + { 0x0439 /* 1081 - Hindi - India */, 0 /* None */, 1252 /* ANSI Latin 1 */}, > + { 0x0419 /* 1049 - Russian - Russia */, 1251, 1251 /* ANSI Cyrillic */}, > + { 0x0411 /* 1041 - Japanese - Japan */, 932, 932 /* Shift-JIS */} > + }; > + > + last = GetThreadLocale(); > + > + for (i = 0; i < sizeof(lcids)/sizeof(lcids[0]); i++) > + { > + SetThreadLocale(lcids[i].lcid); > + > + GetLocaleInfoA(lcids[i].lcid, LOCALE_IDEFAULTANSICODEPAGE|LOCALE_RETURN_NUMBER, (LPSTR)&cp, sizeof(cp)); > + ok(cp == lcids[i].defacp, "wrong codepage %i for lcid %04x, should be %04x\n", cp, lcids[i].defacp, cp); > + > + memset(&cpi, 0, sizeof(cpi)); > + GetCPInfoExA(CP_ACP, 0, &cpi); > + ok(cpi.CodePage == 1252 /* ANSI Latin 1 */, "wrong codepage %i for lcid %04x, should be 1252\n", > + cpi.CodePage, lcids[i].lcid); > + > + SetLastError(0xdeadbeef); > + ret = WideCharToMultiByte(CP_ACP, 0, foobarW, -1, NULL, 0, NULL, NULL); > + ok(ret == 7, "ret is %d (%04x)\n", ret, lcids[i].lcid); > + ok(GetLastError() == 0xdeadbeef, "GetLastError() is %u (%04x)\n", GetLastError(), lcids[i].lcid); > + > + GetCPInfoExA(CP_THREAD_ACP, 0, &cpi); > + ok(cpi.CodePage == lcids[i].threadcp, "wrong codepage %i for lcid %04x, should be %i\n", > + cpi.CodePage, lcids[i].lcid, lcids[i].threadcp); > + > + SetLastError(0xdeadbeef); > + ret = WideCharToMultiByte(CP_THREAD_ACP, 0, foobarW, -1, NULL, 0, NULL, NULL); > + ok(ret == 7, "ret is %d (%04x)\n", ret, lcids[i].lcid); > + ok(GetLastError() == 0xdeadbeef, "GetLastError() is %u (%04x)\n", GetLastError(), lcids[i].lcid); > + } > + > + SetThreadLocale(last); That's better, but you should use appropriate constants for the locale ids, and test the other functions that use get_codepage_table to make sure the behavior is common to all of them. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: msvcrt/tests: Enable compilation with -D__WINESRC__
by Marvin 05 Nov '13

05 Nov '13

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at https://newtestbot.winehq.org/JobDetails.pl?Key=3213 Your paranoid android. === wxppro (32 bit file) === file.c:1417: Test failed: bad position 0 expecting 8 file.c:1419: Test failed: Couldn't read back the data file.c:1987: Test failed: i 0, got 0 file.c:1990: Test failed: expected to read 'Hello world', got '' file.c:1987: Test failed: i 1, got 0 file.c:1990: Test failed: expected to read 'Hello world', got '' file.c:1987: Test failed: i 2, got 0 file.c:1990: Test failed: expected to read 'Hello world', got '' file.c:2024: Test failed: fread() returned 0: ferror=0 file.c:2027: Test failed: got '' expected 'Hello worldHello worldHello world'

1 0

How to undo authors list generated automatically.
by Tae Wong 04 Nov '13

04 Nov '13

How to undo the commit: 5da31c8817293ff98b4defceb962f7d0fbc25829: AUTHORS: Regenerate the authors list from git using a .mailmap file.? It's generated automatically causing word spaces and diacritics to be ignored.

1 0

Re: kernel32: use ANSI code page if current thread locale has no code page (try 2)
by Alexandre Julliard 04 Nov '13

04 Nov '13

Daniel Lehman <dlehman(a)esri.com> writes: > @@ -394,10 +398,48 @@ static void test_undefined_byte_char(void) > } > } > > +static void test_unicode_only(void) > +{ > + LCID last; > + BOOL ret; > + > + if (!pSetThreadLocale || !pGetThreadLocale) > + { > + win_skip("SetThreadLocale and/or GetThreadLocale not found\n"); > + return; > + } > + > + if (!IsValidCodePage(1081)) /* Hindi_India */ > + { > + skip("Codepage 1081 not available\n"); > + return; > + } > + > + last = pGetThreadLocale(); > + > + pSetThreadLocale(1252); > + SetLastError(0xdeadbeef); > + ret = WideCharToMultiByte(1081, 0, foobarW, -1, NULL, 0, NULL, NULL); > + ok(ret == 7, "ret is %d\n", ret); > + ok(GetLastError() == 0xdeadbeef, "GetLastError() is %u\n", GetLastError()); This doesn't make any sense. 1081 is not a codepage, and 1252 is not a locale. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: Fwd: msvcr: Implemented _chsize_s
by Alexandre Julliard 04 Nov '13

04 Nov '13

Andrey Zhezherun <zhezherun(a)yandex.ru> writes: > On 1 November 2013 09:37, Piotr Caban <piotr.caban(a)gmail.com> wrote: >> On 11/01/13 00:40, Andrey Zhezherun wrote: >>> +__int64 __cdecl _chsize_s(int,__int64); >> There should be int (or errno_t) instead of __int64. > Fixed, revised patch is attached. > > Regards, > Andrey > > --- wine-1.7.5.orig/dlls/crtdll/crtdll.spec 2013-10-25 18:45:30.000000000 +0100 > +++ wine-1.7.5/dlls/crtdll/crtdll.spec 2013-10-30 11:09:56.000000000 +0000 > @@ -58,6 +58,7 @@ > @ cdecl _chgsign(double) msvcrt._chgsign > @ cdecl _chmod(str long) msvcrt._chmod > @ cdecl _chsize(long long) msvcrt._chsize > +@ cdecl _chsize_s(long int64) msvcrt._chsize_s > @ cdecl _clearfp() msvcrt._clearfp > @ cdecl _close(long) msvcrt._close > @ cdecl _commit(long) msvcrt._commit It's not exported in old dlls, you shouldn't be adding it. -- Alexandre Julliard julliard(a)winehq.org

1 0

using mono dll components on wine(C#)
by 中川祥 04 Nov '13

04 Nov '13

I do not know how to use COM classes on wine lib,for example , System.dll wine dlls should be written by C,but I found many .Net methods equivalent to Win32 APIs. Isn't it good idea?

2 1

Re: [PATCH 3/3] winmm/tests: Compile with -D__WINESRC__
by Marvin 03 Nov '13

03 Nov '13

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at https://newtestbot.winehq.org/JobDetails.pl?Key=3187 Your paranoid android. === wxppro (32 bit mci) === mci.c:1212: Test failed: mci status "C:\Documents and Settings\winetest\My Documents\tempfile.wav" mode returned MCIERR_INVALID_DEVICE_NAME mci.c:1277: Test failed: mci re-auto-open status mode returned MCIERR_INVALID_DEVICE_NAME mci.c:1293: Test failed: mci auto-open pause returned MCIERR_INVALID_DEVICE_NAME === wvista (32 bit mci) === mci.c:989: Test failed: 395ms is not in the expected 400-600ms range mci.c:996: Test failed: 395ms is not even in the expected 400-1000ms range === w8 (32 bit mci) === mci.c:1123: Test failed: got 0001 instead of MCI_NOTIFY_xyz 0004 from command play (aborted by close) === w7pro64 (64 bit mci) === mci.c:989: Test failed: 389ms is not in the expected 400-600ms range mci.c:996: Test failed: 389ms is not even in the expected 400-1000ms range

1 0

Re: msvcr: Implemented _chsize_s
by Piotr Caban 03 Nov '13

03 Nov '13

On 11/01/13 00:40, Andrey Zhezherun wrote: > --- wine-1.7.5.orig/include/msvcrt/io.h 2013-10-25 18:45:30.000000000 +0100 > +++ wine-1.7.5/include/msvcrt/io.h 2013-10-30 11:34:59.000000000 +0000 > @@ -87,6 +87,7 @@ > int __cdecl _access(const char*,int); > int __cdecl _chmod(const char*,int); > int __cdecl _chsize(int,__msvcrt_ulong); > +__int64 __cdecl _chsize_s(int,__int64); > int __cdecl _close(int); > int __cdecl _commit(int); > int __cdecl _creat(const char*,int); There should be int (or errno_t) instead of __int64.

2 2

Re: msvcp: fixed complex division
by Andrey Zhezherun 03 Nov '13

03 Nov '13

On 1 November 2013 10:40, Piotr Caban <piotr.caban(a)gmail.com> wrote: > How about making it more numerically safe? Makes sense, although I checked what libstdc++ does and it has the vanilla version, similar to what I had before. Not sure what the native msvcp does though. A revised patch is attached. Regards, Andrey

1 0

Re: msvcr: Added _crt_debugger_hook function
by Piotr Caban 03 Nov '13

03 Nov '13

On 11/01/13 00:41, Andrey Zhezherun wrote: > +/********************************************************************* > + * _crt_debugger_hook (MSVCRT.@) > + */ > +void CDECL MSVCRT__crt_debugger_hook(int reserved) > +{ > + TRACE(""); > +} The trace needs to end with '\n' character. Probably it's better to use ERR or FIXME here. Why do you need this function? In all cases I've seen the function was called only when the application was working incorrectly. Cheers, Piotr

2 1