September 2009 - Wine-devel

Re: Add dxdiag help message box
by Henri Verbeet 17 Sep '09

17 Sep '09

2009/9/17 Brian Nguyen <mtxcoll(a)gmail.com>: > +/* Thread-safe function for converting to wide char strings at runtime */ > +LPWSTR PrintWide(LPWSTR buf, size_t len, const char *s) > +{ > + size_t i; > + for (i = 0; i < len && *s != '\0'; i++, s++) { > + buf[i] = *s; > + } > + return buf; > +} That doesn't do what you want, you should use MultiByteToWideChar() to convert to WCHAR. However, I don't think you want to hardcode the strings in the first place, you should load them from the resources so they can be translated. Also, functions like that which are only used in the current file should be static.

2 1

Re: [1/5] mapi32: Scan registry for MAPI providers and load them
by Nikolay Sivov 17 Sep '09

17 Sep '09

Owen Rudge wrote: > --- > dlls/mapi32/Makefile.in | 2 +- > dlls/mapi32/mapi32_main.c | 5 +- > dlls/mapi32/util.c | 148 > +++++++++++++++++++++++++++++++++++++++++++++ > dlls/mapi32/util.h | 28 +++++++++ > 4 files changed, 181 insertions(+), 2 deletions(-) > create mode 100644 dlls/mapi32/util.h Hi. Few comments. > + > +static HMODULE mapi_provider = NULL; > +static HMODULE mapi_ex_provider = NULL; > + This is redundant, static will be initialized to zero anyway. > +static void load_mapi_provider(HKEY hkeyMail, LPCWSTR valueName, HMODULE *mapi_provider) > +{ > + static const WCHAR mapi32_dll[] = {'m','a','p','i','3','2','.','d','l','l',0 }; > + > + DWORD dwType, dwLen = 0; > + LPWSTR dllPath; > + > + if (!mapi_provider) > + return; > + I don't think this check is necessary, it's your internal helper. > +cleanUp: ; Semicolumn isn't needed here.

1 0

please test inconsistent WAVEFORMAT on MS-Windows (compiler needed)
by Joerg-Cyril.Hoehle＠t-systems.com 17 Sep '09

17 Sep '09

Hi, One MS-Windows machine passes all the attached tests, proving that MS-Windows ignores nBlockAlign and nAvgBytesPerSec. It would be helpful to get confirmation from more machines. Please build with the attached patch, execute it in interactive mode set WINETEST_INTERACTIVE=1 winetest.exe wave and listen to the sound and especially the pauses. OTOH, Wine fails at least one test: - One waveOutGetPosition test in WAVE_MAPPER fails. - My recently accepted patches (for OSS and CoreAudio) do something forbidden, that the test was written to reveal: they fix broken values in the WAVEFORMAT description, because both winmm and dsound depend on their correctness. But Wine should not touch user-supplied data. That is a todo_wine until dsound learns how to live with broken input. I'll add todo_wine before submitting these tests as a patch. Try this out with <= 1.1.29 (test with WINETEST_INTERACTIVE=1 make test) and you'll hear how Wine's winmm incorrectly computes the duration of the sound when given inconsistent format descriptions, causing silence where none should be, because MS-Windows does not care about BlockAlign and AvgBytesPerSec. With the upcoming 1.1.30, this will be corrected for OSS and CoreAudio, but not yet with the WAVE_MAPPER (ALSA is still broken). Regards, Jörg Höhle

1 0

Re: winetest: Add the ability to retrieve the dll path for com dlls from the registry.
by Paul Vriens 17 Sep '09

17 Sep '09

On 09/14/2009 03:10 PM, Huw Davies wrote: > --- > programs/winetest/main.c | 87 ++++++++++++++++++++++++++++++++++++++------- > 1 files changed, 73 insertions(+), 14 deletions(-) > > > ------------------------------------------------------------------------ > > Hi Huw, Winetest doesn't do a complete run anymore (tested only on W2K3 and Vista) for now. It ends with a 998 error during enumeration of the test (oleaut32 being the last visible one). I'm crossbuilding a winetest as we speak but just thought I'd mention it. -- Cheers, Paul.

2 3

Re: user32: Skip STM_SETIMAGE message if the static control style doesnot match
by Dmitry Timoshkov 17 Sep '09

17 Sep '09

"Ilya Shpigor" <shpigor(a)etersoft.ru> wrote: > This style checking is needed to the STM_SETIMAGE message processing according > to MSDN: > > http://msdn.microsoft.com/en-us/library/bb760782%28VS.85%29.aspx This change requires a test case. -- Dmitry.

1 0

re: Disabling URL attachments.
by Dan Kegel 17 Sep '09

17 Sep '09

Done.

2 1

dstlen < 0 for MultiByteToWideChar / WideChatToMultiByte
by Michael Stefaniuc 17 Sep '09

17 Sep '09

Hello Detlef, although Smatch could have handled the task too of finding places where Wine passes -1 as destlen to MultiByteToWideChar / WideChatToMultiByte this can be done with the C compiler too. Please see the attached patch. This solution doesn't finds more occurrences that you have found and patched already. But this is really only the stuff that the compiler can figure out at compile time. It has no chance at all to catch stuff like void foo(int bar) { int dstlen = -1; if (bar) dstlen = 10; MultiByteToWideChar(a, b, c, d, e, dstlen) } Those can't be found by a grep either and I think that's why Alexandre is reluctant to apply your patches. bye michael

2 1

Disabling URL attachments.
by Vitaliy Margolen 17 Sep '09

17 Sep '09

Can Bugzilla admin please disable URL attachments? Number of users don't get the concept of what attachment is, and keep attaching urls to things like pastebin, or some file share servers. Instead of attaching actual files. Vitaliy.

1 0

Re: oleaut32: Fail the typelib test when LoadTypeLib fails.
by Jacek Caban 16 Sep '09

16 Sep '09

Hi Vincent, Vincent Povirk wrote: > > As far as I know, this failure only occurs when I do something stupid > like set use_midl_tlb to 1 and don't provide a working > midl_tmarshal.tlb, but I need the test to remind me that I'm an idiot > when I do that. Silently skipping tests is unacceptable. > While you are at this, it would be better to test the exact hr value, so the test should be: ok(hr == S_OK, ...). Jacek

1 0

Status of Quartz
by Keith Muir 16 Sep '09

16 Sep '09

what's the status of wine quartz.dll at present? reason I'm asking is because I'm trying to get a Computer based training(CBT) app Testout Navigator to run which uses Windows Media Player 9 to show videos. Using WINDEBUG with warn+dll seems to show most of the trouble is due to quartz.dll interaction with ntdll.dll Regards, Keith

1 0