September 2009 - Wine-Devel

Re: [kernel32/tests] Fix some test failures on WinMe
by Alexandre Julliard Sept. 24, 2009

Sept. 24, 2009

Paul Vriens <paul.vriens.wine(a)gmail.com> writes: > @@ -1894,10 +1895,18 @@ static void test_FindFirstFileExA(void) > ok(FindNextFile(handle, &search_results), "Fetching third file failed\n"); > ok(CHECK_NAME(search_results.cFileName), "Invalid third entry - %s\n", search_results.cFileName); > > - ok(FindNextFile(handle, &search_results), "Fetching fourth file failed\n"); > + SetLastError(0xdeadbeef); > + ret = FindNextFile(handle, &search_results); > + ok( ret || > + broken(!ret && GetLastError() == ERROR_NO_MORE_FILES), /* WinMe */ > + "Fetching fourth file failed\n"); It sounds like it's actually respecting the FindExSearchLimitToDirectories flag, so I'm not sure this qualifies as broken. Maybe testing without that flag too would be a better test. -- Alexandre Julliard julliard(a)winehq.org

2 2

Re: Fix for Bug 19851 (try 4) - Interlocked instruction support for ARM
by Austin English Sept. 24, 2009

Sept. 24, 2009

On Thu, Sep 24, 2009 at 12:19 PM, Colin Fowler <elethiomel(a)gmail.com> wrote: > Removed trailing white-spaces. I think this is good to go now :) Close ;-). > +#elif defined(__arm__) Like AJ said, if it's generic enough to be used on most archs, it should be the fall back. So instead of '+#elif defined(__arm__)', use '#else', and remove the ' #else', '# error You must implement the interlocked* functions for your CPU'. -- -Austin

1 0

Re: Fix for Bug 19851 (try 3) - Interlocked instruction support for ARM
by André Hentschel Sept. 24, 2009

Sept. 24, 2009

You have some trailing whitespaces in your patch. As far as i had a look it is mostly the intendation from an empty line. -- Best Regards, André Hentschel

1 0

newbie question
by Dmitry Konnov Sept. 24, 2009

Sept. 24, 2009

Hello All, can somebody tell weather it is possible to run NT service on Wine, especially written in Delphi VCL(TService) tx. dima

1 0

Re: Fix for Bug 19851 (try 3) - Interlocked instruction support for ARM
by Alexandre Julliard Sept. 24, 2009

Sept. 24, 2009

Colin Fowler <elethiomel(a)gmail.com> writes: > @@ -396,6 +396,88 @@ > "beq $1,L0xchg_add\n\t" > "mb") > > + > +#elif defined(__arm__) > + > +/* > + * Similar to the Sparc, many ARM CPUs lack the necessary > + * atomic instructions. Fall back in a similar way, but instead > + * use pthreads. This solution is generic enough to work on > + * pretty much any CPU, but is still very fast thanks to > + * Linux's NPTLS. > + */ Actually since it's generic there's no reason to put it in an __arm__ ifdef, it could simply be the default fallback. -- Alexandre Julliard julliard(a)winehq.org

2 1

Re: [1/4] ntdll: add initial test for NtQueryDirectoryFile(), improve test for NtQueryInformationFile() (take 3)
by Alexandre Julliard Sept. 24, 2009

Sept. 24, 2009

Dan Kegel <dank(a)kegel.com> writes: > Change since last try: > - let NtQueryInformationFile() return STATUS_NOT_IMPLEMENTED > (seems to on Vista Home Basic), don't fail if other attributes set Please send unrelated changes as separate patches. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: advapi32: implement LsaLookupSids
by Alexandre Julliard Sept. 24, 2009

Sept. 24, 2009

Aric Stewart <aric(a)codeweavers.com> writes: > + { > + name_size = domain_size = 0; > + (*Names)[i].Use = SidTypeUnknown; > + (*Names)[i].DomainIndex = -1; > + (*Names)[i].Name.Length = 0; > + (*Names)[i].Name.MaximumLength = 0; > + (*Names)[i].Name.Buffer = NULL; > + LookupAccountSidW(NULL, Sids[i], NULL, &name_size, NULL, &domain_size, &use); > + if (GetLastError() != ERROR_NONE_MAPPED) > + { You need to check the function return for failure first, though what you probably want is to check that it succeeded. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: (try 3) shlwapi: implement StrChrNW
by Paul Vriens Sept. 24, 2009

Sept. 24, 2009

On 09/24/2009 02:19 PM, Aric Stewart wrote: > > includes test to show that MSDN is wrong and the search IS case-sensitive. > > thanks for all the comments. > --- > dlls/shlwapi/shlwapi.spec | 1 + > dlls/shlwapi/string.c | 22 ++++++++++++++++++++++ > dlls/shlwapi/tests/string.c | 22 ++++++++++++++++++++++ > 3 files changed, 45 insertions(+), 0 deletions(-) > > > > ------------------------------------------------------------------------ > > Hi Aric, You should return from the function after the win_skip(). -- Cheers, Paul.

2 1

Re: (resend) shlwapi: implement StrChrNW
by Jacek Caban Sept. 24, 2009

Sept. 24, 2009

Hi Aric, Aric Stewart wrote: > --- > dlls/shlwapi/shlwapi.spec | 1 + > dlls/shlwapi/string.c | 23 +++++++++++++++++++++++ > 2 files changed, 24 insertions(+), 0 deletions(-) > I didn't test it myself, but StrChrNW name suggests, that the search shouldn't be case insensitive. Jacek

4 4

Re: shlwapi: implement StrChrNW
by Dmitry Timoshkov Sept. 24, 2009

Sept. 24, 2009

"Aric Stewart" <aric(a)codeweavers.com> wrote: > +LPWSTR WINAPI StrChrNW(LPCWSTR lpszStr, WCHAR ch, UINT cchMax) > +{ > + TRACE("(%s(%i),%i)\n", debugstr_wn(lpszStr,cchMax), cchMax, ch); > + > + if (lpszStr) > + { > + ch = toupperW(ch); > + while (*lpszStr && cchMax > 0) > + { > + if (toupperW(*lpszStr) == ch) > + return (LPWSTR)lpszStr; > + lpszStr++; > + cchMax --; > + } > + lpszStr = NULL; > + } > + return (LPWSTR)lpszStr; > +} Are you sure that the search should be case insensitive? -- Dmitry.

2 1

Re: ole32/tests: Test for nested storage reference counting
by James McKenzie Sept. 24, 2009

Sept. 24, 2009

Markus Stockhausen wrote: > Hi, > > as Nikolay had some tipps for the patch (layout of header and coding) I > attach a new version of the testcase for IStorage reference counters. > Please resend in git format. AJ will reject this patch outright. James McKenzie

2 1

Submitted Bug Link accepted but not showing up
by Erich Hoover Sept. 23, 2009

Sept. 23, 2009

I received an email that a "Bug Link" I requested was accepted, but it's not showing up: http://appdb.winehq.org/objectManager.php?sClass=version&iId=11395 Relevant text: The bug link you submitted between Bug 7929 and Command & Conquer 3: Tiberium Wars Kane's Wrath: 1.0 has been accepted. ... Any thoughts? Erich Hoover ehoover(a)mines.edu

2 1

Coverity access
by John Klehm Sept. 23, 2009

Sept. 23, 2009

Is there a local list of coverity defects for people like me with a spare moment here and there? Or if not is it possible to get me signed up? Thanks, --John Klehm

1 0

jscript: Update French translation
by Frédéric Delanoy Sept. 23, 2009

Sept. 23, 2009

1 0

Re: shlwapi: implement StrChrNW
by Nikolay Sivov Sept. 23, 2009

Sept. 23, 2009

Aric Stewart wrote: > --- > dlls/shlwapi/shlwapi.spec | 1 + > dlls/shlwapi/string.c | 23 +++++++++++++++++++++++ > 2 files changed, 24 insertions(+), 0 deletions(-) Maybe it's better to move this --- +@ stdcall StrChrNW(wstr long long) --- closer to context: --- @ stdcall StrChrA (str long) @ stdcall StrChrIA (str long) @ stdcall StrChrIW (wstr long) @ stdcall StrChrW (wstr long) --- to keep it consistent.

1 0

Re: [4/4] user32: Prevent clipboard viewers that change the clipboard from going into an infinite loop.
by Alexandre Julliard Sept. 23, 2009

Sept. 23, 2009

Alexander Scott-Johns <alexander.scott.johns(a)googlemail.com> writes: > @@ -287,6 +287,7 @@ BOOL WINAPI OpenClipboard( HWND hWnd ) > */ > BOOL WINAPI CloseClipboard(void) > { > + static BOOL bRecursion = FALSE; > BOOL bRet = FALSE; You shouldn't use a static variable, this would have to be per-thread. But it's probably better to find a way to do that based on the current clipboard info. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: Fix for Bug 19851 (try 2) - Interlocked instruction support for ARM
by Dmitry Timoshkov Sept. 22, 2009

Sept. 22, 2009

"Colin Fowler" <elethiomel(a)gmail.com> wrote: > This patch fixes a wrongly named variable in the last patch and also a minor > typo in the comment. Please also set tab size in your editor to 8, and fix spacing problems after that. -- Dmitry.

1 0

Re: [TRY 2] user32/tests: Add tests for STM_SETIMAGE and STM_SETICONmessage processing
by Dmitry Timoshkov Sept. 22, 2009

Sept. 22, 2009

"Ilya Shpigor" <shpigor(a)etersoft.ru> wrote: > Subject: [TRY 2] user32/tests: Add tests for STM_SETIMAGE and STM_SETICONmessage processing Do these tests pass on Wine? If not you need to use todo_wine in appropriate places, or post a fix as [1/2] and the tests as [2/2]. Also, please add a check that once first SendMessage which sets an image, next SendMessage returns that image, i.e. that the first SendMessage succeeds. -- Dmitry.

1 0

A student hopes to learn more about the Wine Community and Free/Open Source Software
by Tiebing Shi Sept. 22, 2009

Sept. 22, 2009

Dear friends, I have been reading your postings to this mailing list and some of other documents at the web site of the Wine Community. I have really enjoyed reading about your collaborative creative activities and your perspectives on the free/open source software. My name is Tiebing Shi and I am a Ph.D. student at Queen's University in Canada. I am very interested in learning more about the Free/Open Source Software Community. The study that I am doing for my Ph.D. thesis is called "User Creativity: An Investigation of the Free/Open Source Software Community." The purpose of my thesis is to study how developers who are often users of their creative products collectively create free/open source software. The finding of this study will help other user communities or organizations that hope to facilitate collective creative activities among their own members. I sincerely invite you to participate in my study. I would like to interview you in your role as a developer and user in the Wine Community. Your experiences of developing and using Wine programs and your perspectives on the free/open source software will greatly help me to understand the creative activities of the Wine Community. If your time permits, I would like to interview you via phone. The phone interview will last about 1 hour. If you like, we can conduct our interview via email. Please be kindly informed that there is no obligation to participate in this study. But your participation will be highly appreciated. Your confidentiality will be protected. For example, your real name will NOT be used in my thesis and future publications. Further information about this study and our interview will be provided to you upon request. Please respond to this email to let me know your interest in participating in a phone or an email interview with me. If you select the phone interview format, please select a time for an interview that is most convenient for you. Please also let me know your phone number and time zone (my time zone is UTC-4hours) so that I can call you at your local time you select. Should you have any questions, please do not hesitate to contact me. I will respond to you as soon as possible. Thank you for your support and I look forward to talking with you. Sincerely, Tiebing Shi Queen's School of Business, Queen’s University Kingston, ON, Canada K7L 3N6 Tel: 613-533-2377 (Office) Email: tshi(a)business.queensu.ca

1 0

Re: Wine in Tango
by Ralf Jung Sept. 22, 2009

Sept. 22, 2009

Hi everyone, > I don't know if we can say this. Only when our Wine-supplied icons are > appearing near application-supplied icons do we gain some consistency by > mimicing Windows style, but that consistency is confined to that > particular app. Most icons the user sees are instead going to be > compared with the rest of the desktop and its applications, and adopting > a Tango style rather than a Windows style is the only way to get that > overall consistency. > If we're not careful, Wine apps may continue to stick out rather than be > just another part of the desktop. If wine should integrate well with the surrounding Linux desktop, why don't you use the desktop icon set where possible? By using the Tango icons, wine applications will still stick out on almost all non-Gnome system (most notably KDE) as well as those Gnome desktops where the icon set was changed - that's not what I would call integration. I agree that the Tango icons are prettier than the one currently used, and as a result an improvement, but unfortunately they won't fix the "sticking out". Kind regards, Ralf Jung -- GRATIS für alle GMX-Mitglieder: Die maxdome Movie-FLAT! Jetzt freischalten unter http://portal.gmx.net/de/go/maxdome01

7 13

ole32:clipboard test failures
by Greg Geldorp Sept. 22, 2009

Sept. 22, 2009

I've been looking into the ole32:clipboard test failures (http://test.winehq.org/data/tests/ole32:clipboard.html) I'm now convinced that Windows simply doesn't handle FORMATETC.ptd correctly when retrieving an IDataObject via OleGetClipboard(). When looking at the test.winehq.org status page mentioned above, it appears that the tests succeed on at least a number of boxes. However, when doing a bit more testing on the retrieved ptd data (see attached diff) it's obvious that the retrieved ptd data contains garbage on all Windows versions >= Win2k (meaning with extra tests in attached diff applied I see failures on all my >= Win2k machines). It appears that at some point Windows is using uninitialized memory. I verified that by hooking HeapAlloc() and filling the allocated memory with a 0x00000008 0x00000008 ... pattern (when the HEAP_ZERO_MEMORY flag was not set). This pattern then showed up in the retrieved ptd data (tdSize was 0x00000008, tdDriverNameOffset was 0x0008, tdDeviceNameOffset was 0x0000). Similarly, when I initialized the allocated memory with a 0x0000000c pattern that pattern showed up in the retrieved ptd data. If Windows is really referencing some uninitialized block of memory then that means we can't just use broken() when comparing expected and actual ptd fields. The status page above indicates that the ole32:clipboard test crashes on a lot of machines. The crashes that I investigated on my machines all happen somewhere inside ole32.dll. This can be expected if uninitialized memory is used. I don't see a way to use broken() to work around this. My proposed solution is to not set the ptd field in DataObjectImpl_CreateComplex() when the created IDataObject is going to be used for OleSetClipboard()/OleGetClipboard(). I don't really like that, so if anyone has ideas for a better solution I'd love to hear them. Ge van Geldorp.

1 0

Re: user32/tests: Add tests for STM_SETIMAGE and STM_SETICON messageprocessing
by Dmitry Timoshkov Sept. 22, 2009

Sept. 22, 2009

"Ilya Shpigor" <shpigor(a)etersoft.ru> wrote: > + SendMessageW(hStatic, STM_SETIMAGE, (WPARAM)IMAGE_BITMAP, (LPARAM)hbm); > + res = SendMessageW(hStatic, STM_SETIMAGE, (WPARAM)IMAGE_BITMAP, (LPARAM)hbm); > + ok(!res, "expected NULL got 0x%lx\n", res); > + > + SendMessageW(hStatic, STM_SETIMAGE, (WPARAM)IMAGE_ICON, (LPARAM)hicon); > + res = SendMessageW(hStatic, STM_SETIMAGE, (WPARAM)IMAGE_ICON, (LPARAM)hicon); > + ok(!res, "expected NULL got 0x%lx\n", res); > + > + SendMessageW(hStatic, STM_SETICON, (WPARAM)hicon, 0); > + res = SendMessageW(hStatic, STM_SETICON, (WPARAM)hicon, 0); > + ok(!res, "expected NULL got 0x%lx\n", res); > + > + SendMessageW(hStatic, STM_SETIMAGE, (WPARAM)IMAGE_ENHMETAFILE, (LPARAM)hemf); > + res = SendMessageW(hStatic, STM_SETIMAGE, (WPARAM)IMAGE_ENHMETAFILE, (LPARAM)hemf); > + ok(!res, "expected NULL got 0x%lx\n", res); Is there a reason you are calling SendMessageW twice (here and in the following tests)? -- Dmitry.

2 1

Re: tools/winegcc: fix linker for Haiku
by Detlef Riekenberg Sept. 22, 2009

Sept. 22, 2009

On Mo, 2009-09-21 at 01:52 -0500, Austin English wrote: > +#ifdef __HAIKU__ > + strarray_add(link_args, "-lroot"); > +#else > + strarray_add(link_args, "-lm"); > + strarray_add(link_args, "-lc"); > +#endif > + } "autoconf.ac" was designed to avoid OS dependant "#ifdef". Test for a function and use the resulting defines from "config.h", when needed. The linker failure message give you a function-name for testing. -- By by ... Detlef

2 1

Request to re-open C&C3 networking bug
by Erich Hoover Sept. 21, 2009

Sept. 21, 2009

I made this request in the bug, but I figured it would be good to post here as well. I've posted a patch to Bug 7929 (C&C 3 network does not work) that I believe addresses the concerns people have addressed over previous attempts to fix the issue. So, I would like to request that the bug be re-opened and for people to take a look and comment on the patch. http://bugs.winehq.org/show_bug.cgi?id=7929#c79 http://bugs2.winehq.org/attachment.cgi?id=23679 Erich Hoover ehoover(a)mines.edu

1 0

Re: Fix for Bug 19851 - Interlocked instruction support for ARM
by André Hentschel Sept. 21, 2009

Sept. 21, 2009

> > This patch fixes bug 19851 http://bugs.winehq.org/show_bug.cgi?id=19851 > One more step towards winelib on ARM :) > > Apologies if this comes through as a double post. I inadvertently sent the mail > before my subscription to the list went through properly. > > Colin > Hi, there is a typo in your patch: > static pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER; > . > . > . > pthread_mutex_lock( &interlocked_mutex ); its mutex and interlocked_mutex. also it might be good to mention a sent patch in the bug, so interested people notice it. -- Best Regards, André Hentschel

2 1