January 2010 - Wine-devel - list.winehq.org

errno-ntstatus conversion
by Mike Kaplinskiy 24 Jan '10

24 Jan '10

There are many places where we need to do a conversion between the unix errno and the windows ntstatus (and then sometimes to dos errors). Here are some that I found: http://source.winehq.org/source/server/file.c#L582 http://source.winehq.org/source/server/sock.c#L702 http://source.winehq.org/source/dlls/wininet/netconnection.c#L512 http://source.winehq.org/source/dlls/winhttp/net.c#L153 http://source.winehq.org/source/dlls/ntdll/file.c#L312 Would it be a good idea to perhaps create an include file for both the server and the dlls to convert unix error codes to status/winsock error codes? The reason for 2 functions is winsock error codes do not actually map to any status code, they seem to be there to make porting easier (ie STATUS_CONNECTION_RESET maps to ERROR_CONNECTION_RESET). Does this seem reasonable, or should we just keep duplicating the conversion code everywhere? I have not checked whether there are places that have different statuses for the same unix error, but presumably we can pick a reasonable default and just have special cases within the module. This is coming from some conversion within the server. The server should not return wsa-based error codes in its error events list, and just use wsa conversion when sending window messages. Mike.

1 0

Re: [PATCH] Add missing classes for DOMDocument in msxml, and add test-cases back it up.
by Austin English 23 Jan '10

23 Jan '10

On Sat, Jan 23, 2010 at 9:24 AM, Terry <tbos(a)verizon.net> wrote: Forgot the patch. -- -Austin

1 0

cygwin, perl, wine, rebase, and ReBaseImage()
by Dan Kegel 22 Jan '10

22 Jan '10

While building chromium with visual c++ 2005 on wine, step 'webcore_bindings_sources' tries to run WebCore/bindings/scripts/generate-bindings.pl, and sometimes it fails with 13 [main] perl 29 C:\chromium\src\third_party\cygwin\bin\perl.exe: *** fatal error - unable to remap C:\chromium\src\third_party\cygwin\lib\perl5\5.10\i686-cygwin\auto\List\Util\Util.dll to same address as parent(0x1350000) != 0x52760000^M 14 [main] perl 74 fork: child 29 - died waiting for dll loading, errno 11^M http://www.tishler.net/jason/software/rebase/rebase-2.2.README indicates that cygwin's fork() relies on everything mapping to the same address in the child. (Gulp.) This is an occasional problem even on Windows; the usual advice is to run cygwin's rebaseall (carefully). Sure enough, running wine c:\\cygwin\\bin\ash c:\\cygwin\\bin\\rebaseall seems to have solved the problem. Or maybe I just got lucky. Cygwin's rebase command has its own implementation of ReBaseImage(), so it doesn't matter that Wine's is a stub. (Gee, maybe we should ask Ralf sometime if he's willing to relicense it from gpl. We'd have to translate it from C++ to C, too.) - Dan

1 0

comctl32 listview notification assert
by Aric Stewart 22 Jan '10

22 Jan '10

Hello, I am seeing something in listview.c that really make me feel like there is something really wrong going on. So maybe it is something I am having a lack of understanding in. I have an application that is sending 'A' style notification codes to the listview header. Specifically I am seeing an HDN_ITEMCHANGEDA. In the listview code the first thing HDN_ITEMCHANGEDA and HDN_ITEMCHANGEDW does is notify_forward_header. notify_forward_header does not understand HDN_ITEMCHANGEDA and then asserts. It looks straight forward to add the A versions of the notifications to that case statement. But often when something looks that straight forward there is something more underlying it. Especailly since this looks like something we would be seeing a lot of if it was as much of an issue as it appears. does anyone have a deeper understanding of the listview code that would help me figure out what is going on. Or maybe just tell me that I am over thinking it and it really is as easy of a fix as it looks. thanks! -aric

1 0

Re: shell32: Dynamically load an ordinal only export.
by Alexandre Julliard 22 Jan '10

22 Jan '10

Huw Davies <huw(a)codeweavers.com> writes: > --- > dlls/shell32/tests/shlfileop.c | 18 +++++++++--------- > 1 files changed, 9 insertions(+), 9 deletions(-) This shouldn't be needed, import by ordinal works just fine. -- Alexandre Julliard julliard(a)winehq.org

2 3

Re: winmm: Add a bunch of new mmio tests which discover some bugs in mmio handling.
by Alexandre Julliard 22 Jan '10

22 Jan '10

Dmitry Timoshkov <dmitry(a)codeweavers.com> writes: > @@ -45,6 +45,20 @@ static DWORD RIFF_buf[] = > 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 > }; > > +#define expect_buf_offset(_hmmio, _off) \ > +do \ > +{ \ > + MMIOINFO _mmio; \ > + LONG _ret; \ > + memset(&_mmio, 0, sizeof(_mmio)); \ > + _ret = mmioGetInfo((_hmmio), &_mmio, 0); \ > + ok(_ret == MMSYSERR_NOERROR, "mmioGetInfo error %u\n", _ret); \ > + ok(_mmio.lBufOffset == 0, "expected 0, got %d\n", _mmio.lBufOffset); \ > + _ret = mmioSeek((_hmmio), 0, SEEK_CUR); \ > + ok(_ret == (_off), "expected %d, got %d\n", (_off), _ret); \ > +} \ > +while (0) Please avoid such multi-line macros, define a function instead. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH 1/2] ws2_32: use ntstatus in overlapped functions (try 2)
by Alexandre Julliard 22 Jan '10

22 Jan '10

Mike Kaplinskiy <mike.kaplinskiy(a)gmail.com> writes: > Also some style fixes and move case 0 to the top - it's the most likely Nothing says that likely cases have to be at the top, and anyway 0 is probably the least likely of all (and should in fact never happen). -- Alexandre Julliard julliard(a)winehq.org

1 0

"Mockba to Berlin" game - mouse not clicking
by Pavel Troller 22 Jan '10

22 Jan '10

Hi! My son bought a game DVD bound to a youngster's magazine. It contains a 5 years old game by Monte Cristo Multimedia, called "Mockba to Berlin". We were trying this game with wine 1.1.36. The game PERFECTLY installs and then PERFECTLY launches, going through all the initial screens and videos. It then renders a primary menu on the screen, allowing to start a new game, load a save, perform some settings, etc. The mouse cursor is fully functional, the hovered item gets brightened and an explanatory text appears, which tells, what the particular item is for. BUT: The mouse buttons are DEAD. All. I cannot left-click, right-click, mid-click, simply nothing. It' very frustrating, it looks that just ONE CLICK remains to start the game, but you can't do it! There is no console output related to the problem, just some warnings from the GL renederer, which seem harmless. There is no AppDB entry for this game (we didn't found any), so we cannot follow any hints, how to solve the problem. The game was tested on 3 different machines but with the same wine installation. The results were identical. Are there any general tips, what we can try to solve the problem ? With regards, Pavel

4 5

Re: [2/2] xmllite/reader: Basic input object creation on IXmlReader::SetInput()
by Paul Vriens 22 Jan '10

22 Jan '10

On 01/21/2010 07:08 PM, Nikolay Sivov wrote: > This patch depends on previously sent, forget it properly, sorry. > Basic input object creation on IXmlReader::SetInput() > > > > Hi Nikolay, This one doesn't apply cleanly on top of: "xmllite/tests: Test query for supported interface sequence while creating IXmlReaderInput instance" -- Cheers, Paul.

2 1

Re: [PATCH 4/5] xmllite/tests: Add basic test structure for IXmlReader
by Paul Vriens 22 Jan '10

22 Jan '10

On 01/18/2010 10:28 PM, Nikolay Sivov wrote: > Add basic test structure for IXmlReader > Hi Nikolay, This one crashes on Vista without a servicepack at: + IXmlReader_Release(reader); +} All the calls before that one look ok (they produce the expected output). I guess "vista without an SP" can be considered broken(), but how do deal with that? -- Cheers, Paul.

3 10