June 2009 - Wine-Devel - list.winehq.org

[RESEND] mountmgr.sys: implement forced setting of fixed/local disk
by Ben Klein June 4, 2009

June 4, 2009

define_unix_drive in mountmgr.c includes all cases used in query_unix_drive except the DRIVE_FIXED == DEVICE_HARDDISK_VOL mapping. This patch fixes 17618 where setting the drive type to "Local disk" in winecfg does not save the change to the registry. I've CC'd wine-devel in the hope of getting some feedback this time. This is the 3rd or 4th time I've submitted this patch.

2 3

Re: (resend, 2nd) kernel32: Implement QueryFullProcessImageNameA (2nd try)
by Alexandre Julliard June 4, 2009

June 4, 2009

Stefan Leichter <Stefan.Leichter(a)camline.com> writes: > /****************************************************************** > + * QueryFullProcessImageNameA (KERNEL32.@) > + */ > +BOOL WINAPI QueryFullProcessImageNameA(HANDLE hProcess, DWORD dwFlags, LPSTR lpExeName, PDWORD pdwSize) > +{ > + BOOL retval; > + DWORD pdwSizeW = *pdwSize; > + LPWSTR lpExeNameW = HeapAlloc( GetProcessHeap(), HEAP_ZERO_MEMORY, *pdwSize * sizeof(WCHAR)); > + > + retval = QueryFullProcessImageNameW(hProcess, dwFlags, lpExeNameW, &pdwSizeW); > + > + if(retval) { > + if(lpExeName) > + *pdwSize = FILE_name_WtoA(lpExeNameW, pdwSizeW, lpExeName, *pdwSize); > + else { > + /* this is a difference between the acsii and the unicode version (vista) > + * the unicode version crashes when the size is big enough to hold the result > + * ascii version throughs an error > + */ > + retval = FALSE; > + SetLastError(ERROR_INVALID_PARAMETER); > + } > + } I'd expect that error to be a side effect of WideCharToMultiByte, which is most likely what you want to use here. Also you have to handle buffer overflows correctly, with test cases. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: jscript: Fix DateConstr_value (with one argument) implementation (3/9)
by Alexandre Julliard June 4, 2009

June 4, 2009

Piotr Caban <piotr.caban(a)gmail.com> writes: > + FILETIME ft_now, ft_epoch; > + SYSTEMTIME st_epoch; > + DOUBLE time_now, time_epoch; > + > + memset(&st_epoch, 0, sizeof(SYSTEMTIME)); > + st_epoch.wYear = 1970; > + st_epoch.wMonth = 1; > + st_epoch.wDay = 1; > + SystemTimeToFileTime(&st_epoch, &ft_epoch); > + > + GetSystemTimeAsFileTime(&ft_now); > > - GetSystemTimeAsFileTime(&time); > + time_now = (DOUBLE)ft_now.dwHighDateTime/10000*UINT_MAX > + + (DOUBLE)ft_now.dwLowDateTime/10000; > + time_epoch = (DOUBLE)ft_epoch.dwHighDateTime/10000*UINT_MAX > + + (DOUBLE)ft_epoch.dwLowDateTime/10000; This is not correct, you need a <<32. Also there's no reason to use doubles for the computation, and you shouldn't have to use SystemTimeToFileTime for the epoch, it can be a constant. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: user32/tests: Add tests for message sequence of window destruction at thread exit.
by Paul Vriens June 4, 2009

June 4, 2009

Hi Alexandre, I assume you've already seen the test failures on some boxes. Looks like the WmExitThreadSeq sequence is not always the same. -- Cheers, Paul.

1 0

Re: (try2) [1/2] kernel32/tests: Add tests for CreateFile errors and opening volumes
by Paul Vriens June 4, 2009

June 4, 2009

Guy Albertelli wrote: > Changelog: > - Add tests to verify error codes returned for certain error cases. > - Also add tests for opening volumes and handle the case where we do > not have the rights to access the volume. > --- > dlls/kernel32/tests/file.c | 326 +++++++++++++++++++++++++++++++++++++++++++- > 1 files changed, 324 insertions(+), 2 deletions(-) > > > > ------------------------------------------------------------------------ > > Hi Guy, The newly added tests introduce some failures on Win9x and WinME. Could you have a look? -- Cheers, Paul.

2 1

Re: Alexandre Julliard : clock: Build resource files separately instead of including them all from a global file .
by Michael Stefaniuc June 3, 2009

June 3, 2009

Hello Alexandre, Alexandre Julliard wrote: > Module: wine > Branch: master > Commit: aa41526c73a5fbcc2c08a9342107ee76173a2b97 > URL: http://source.winehq.org/git/wine.git/?a=commit;h=aa41526c73a5fbcc2c08a9342… > > Author: Alexandre Julliard <julliard(a)winehq.org> > Date: Sun May 31 20:11:00 2009 +0200 > > clock: Build resource files separately instead of including them all from a global file. there is a minor problem with this approach. wrc --verify-translation becomes useless as each file is separate and a comparison to the English default resource is no longer possible. That's what Mikołaj Zalewski's Wine translation statistics webpage http://www.mikolaj.zalewski.pl/wine-transl/ is using. I also have a makefile patch/hack in my i18n branch to do a wrc --verify-translation pass for each .rc file. Very useful to find problems when doing translation work. bye michael

1 0

The present state of Direct3D in WINE
by MD.IMAM HOSSAIN June 3, 2009

June 3, 2009

Dear, WINE Developers, Thanks to all of you for creating an environment where people can enjoy games. I very much pleased by the present state of WINE's Direct3D. In my short range of views on earlier versions of wine Grend Theft Auto San Andreas runs very slowly. But now it is very playable. On WINE 1.1.22 all the improvement made to Direct3D is good but there are some incompatibilities too. >From my experiment with recent version of WINE 1.1.22 : 3DMark 2001 has vertex shader problem (Under Game Lobby and Vertex Shader test). Hitman 2 Silent Assassin (No Human Body could be seen). Hitman 3 Contracts (No Human Body could be seen). Prince of Persia Sand of Time (Nothing could be seen). Prince of Persia Warrior Within (Nothing could be seen). I only able to test these games. I hope coming WINE release will fix these problem. Your sincerely, MD.IMAM HOSSAIN

3 4

AppDB backlog of test reports for apps with maintainers
by Rosanne DiMesio June 3, 2009

June 3, 2009

There is a growing backlog of test reports for apps with maintainers still awaiting approval. My understanding of how things are supposed to work is that if a maintainer does not process a report within a week, they are sent a couple of reminder notices, and eventually removed as maintainer if they don't respond. This appears to not be working--the backlog goes back to May 19. I can process the reports myself, of course, but can someone check on what's going on with notifying/removing maintainers? There was a similar backlog about a month ago, but someone (I assume another admin) eventually processed all the reports, and I think the underlying cause was never looked into. -- Rosanne DiMesio <dimesio(a)earthlink.net>

1 0

Problem with async. comm from Wine to a USB->Serial-converter
by Henrik Jacobsson June 3, 2009

June 3, 2009

Hi. I just got myself a USB-GPS (GlobalSat BU-353) and wanted to run SeaClearII, a navigation-program with it. Turns out that Wine got confused when the USB->Serial-driver wasn't acting like a real serial-device though. The app does what it's supposed to do, but Wine consumes 100% of the available CPU. After some googling I found a patch that resolved this issue perfectly, and I'm curious why it never made it into trunk. The patch: http://www.nabble.com/serial-04-04%3A-fix%3A-WaitCommEvent()-does-not-work-… I only managed to apply this patch to 1.1.17. Later releases had changed too much for my non-programmer-brain to cope with.. I would really appreciate if someone in the know could check the patch out and if possible apply it to wine-proper.. Thanks in advance.. /Henrik

3 3

Testing with Comctl32 6.0 (and further versions)
by Nikolay Sivov June 3, 2009

June 3, 2009

Hi. I'm thinking about adding a possibility to test both previous comctl32 and version 6 (or greater maybe in next versions). For example if I compiled in a manifest resource for 6.0 I'll get 6.0 loaded on >= XP, but the best variant for me is to load not manifest aware version first, then unload it (if can't be done automatically) and switch to manifest specified version. Is it possible to do it for example with creating activation context manually (i'm not familiar with it)? Maybe it's easier to create another test target (e.g. comct32_v6_test.exe and its cross built counterpart) - the reason here is to run existing tests on 6.0 too, probably building with some #defines specific for V6 version to tweak test results if there are any difference. Any thoughts? Nikolay S.

1 0