October 2010 - Wine-Devel - list.winehq.org

Re: crypt32/tests(8/10): Add macros to declare tests

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5873 Your paranoid android. === W98SE (32 bit chain) === chain.c:4117: Test failed: ignoredInvalidDateBasePolicyCheck[0](#0001): expected 800b0101, got 800b0102 chain.c:4108: Test failed: invalidExtensionPolicyCheck[0](#0001): expected 800b0105, got 800b0102 chain.c:4117: Test failed: iTunesPolicyCheckWithoutMatchingName[0](#0004): expected 800b010f, got 800b0109 chain.c:4117: Test failed: googlePolicyCheckWithMatchingName[0](#0004): expected 00000000, got 800b0101 chain.c:4117: Test failed: opensslPolicyCheckWithMatchingName[0](#0004): expected 00000000, got 800b0109 chain.c:4117: Test failed: opensslPolicyCheckWithoutMatchingName[0](#0004): expected 800b010f, got 800b0109 chain.c:4117: Test failed: opensslPolicyCheckWithoutMatchingName[0](#0004): expected 800b010f, got 800b0109 chain.c:4117: Test failed: opensslPolicyCheckWithoutMatchingName[0](#0004): expected 800b010f, got 800b0109 chain.c:4117: Test failed: stanfordPolicyCheckWithMatchingName[0](#0004): expected 00000000, got 800b0109 chain.c:4117: Test failed: stanfordPolicyCheckWithMatchingName[0](#0004): expected 00000000, got 800b0109 chain.c:4117: Test failed: stanfordPolicyCheckWithoutMatchingName[0](#0004): expected 800b010f, got 800b0109 chain.c:2752: Test failed: Chain 31: expected error 00000001, got 00000003. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2674: Test failed: Chain 31, element [0,1]: expected error 00000000, got 00000002. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2674: Test failed: Chain 31, element [0,2]: expected error 00000000, got 00000001. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2752: Test failed: Chain 32: expected error 00000001, got 00010001. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2712: Test failed: Chain 32: expected 3 elements, got 2 === WNT4WSSP6 (32 bit chain) === chain.c:4108: Test failed: invalidExtensionPolicyCheck[0](#0001): expected 800b0105, got 800b0102

Oct. 6, 2010

Re: crypt32/tests(7/10): Add name to test to help identify test failures

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5872 Your paranoid android. === W98SE (32 bit chain) === chain.c:4117: Test failed: ignoredInvalidDateBasePolicyCheck[0](#0001): expected 800b0101, got 800b0102 chain.c:4108: Test failed: invalidExtensionPolicyCheck[0](#0001): expected 800b0105, got 800b0102 chain.c:4117: Test failed: iTunesPolicyCheckWithoutMatchingName[0](#0004): expected 800b010f, got 800b0109 chain.c:4117: Test failed: googlePolicyCheckWithMatchingName[0](#0004): expected 00000000, got 800b0101 chain.c:4117: Test failed: opensslPolicyCheckWithMatchingName[0](#0004): expected 00000000, got 800b0109 chain.c:4117: Test failed: opensslPolicyCheckWithoutMatchingName[0](#0004): expected 800b010f, got 800b0109 chain.c:4117: Test failed: opensslPolicyCheckWithoutMatchingName[0](#0004): expected 800b010f, got 800b0109 chain.c:4117: Test failed: opensslPolicyCheckWithoutMatchingName[0](#0004): expected 800b010f, got 800b0109 chain.c:4117: Test failed: stanfordPolicyCheckWithMatchingName[0](#0004): expected 00000000, got 800b0109 chain.c:4117: Test failed: stanfordPolicyCheckWithMatchingName[0](#0004): expected 00000000, got 800b0109 chain.c:4117: Test failed: stanfordPolicyCheckWithoutMatchingName[0](#0004): expected 800b010f, got 800b0109 chain.c:2752: Test failed: Chain 31: expected error 00000001, got 00000003. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2674: Test failed: Chain 31, element [0,1]: expected error 00000000, got 00000002. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2674: Test failed: Chain 31, element [0,2]: expected error 00000000, got 00000001. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2752: Test failed: Chain 32: expected error 00000001, got 00010001. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2712: Test failed: Chain 32: expected 3 elements, got 2 === WNT4WSSP6 (32 bit chain) === chain.c:4108: Test failed: invalidExtensionPolicyCheck[0](#0001): expected 800b0105, got 800b0102

Oct. 6, 2010

Re: crypt32(6/10): Set CERT_TRUST_HAS_NOT_SUPPORTED_CRITICAL_EXT when appropriate

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5871 Your paranoid android. === W98SE (32 bit chain) === chain.c:4113: Test failed: #0001[0]: expected 800b0101, got 800b0102 chain.c:4105: Test failed: #0001[0]: expected 800b0105, got 800b0102 chain.c:2752: Test failed: Chain 31: expected error 00000001, got 00000003. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2674: Test failed: Chain 31, element [0,1]: expected error 00000000, got 00000002. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2674: Test failed: Chain 31, element [0,2]: expected error 00000000, got 00000001. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2752: Test failed: Chain 32: expected error 00000001, got 00010001. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2712: Test failed: Chain 32: expected 3 elements, got 2 === WNT4WSSP6 (32 bit chain) === chain.c:4105: Test failed: #0001[0]: expected 800b0105, got 800b0102

Oct. 6, 2010

Re: crypt32/tests(5/10): Test the base and SSL policies against a certificate with an invalid critical extension

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5870 Your paranoid android. === W98SE (32 bit chain) === chain.c:4113: Test failed: #0001[0]: expected 800b0101, got 800b0102 chain.c:4105: Test failed: #0001[0]: expected 800b0105, got 800b0102 chain.c:2752: Test failed: Chain 31: expected error 00000001, got 00000003. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2674: Test failed: Chain 31, element [0,1]: expected error 00000000, got 00000002. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2674: Test failed: Chain 31, element [0,2]: expected error 00000000, got 00000001. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2752: Test failed: Chain 32: expected error 00000001, got 00010001. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2712: Test failed: Chain 32: expected 3 elements, got 2 === WNT4WSSP6 (32 bit chain) === chain.c:4105: Test failed: #0001[0]: expected 800b0105, got 800b0102

Oct. 6, 2010

Re: crypt32/tests(4/10): Test invalid critical extension in intermediate cert

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5869 Your paranoid android. === W98SE (32 bit chain) === chain.c:4108: Test failed: #0001[0]: expected 800b0101, got 800b0102 chain.c:2752: Test failed: Chain 31: expected error 00000001, got 00000003. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2674: Test failed: Chain 31, element [0,1]: expected error 00000000, got 00000002. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2674: Test failed: Chain 31, element [0,2]: expected error 00000000, got 00000001. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2752: Test failed: Chain 32: expected error 00000001, got 00010001. 00000020 is expected if no valid Verisign root certificate is available. chain.c:2712: Test failed: Chain 32: expected 3 elements, got 2

Oct. 6, 2010

Re: crypt32(2/10): Improve error checking for the base policy

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5868 Your paranoid android. === W98SE (32 bit chain) === chain.c:4015: Test failed: #0001[0]: expected 800b0101, got 800b0102

Oct. 6, 2010

Re: crypt32/tests(1/10): Add more tests of verifying the base policy, including flags to ignore certain errors

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5867 Your paranoid android. === W98SE (32 bit chain) === chain.c:4012: Test failed: #0001[0]: expected 800b0101, got 800b0102

Oct. 6, 2010

Re: (6th try) winhlp32: Added Hebrew translation

by Paul Vriens

Hi Yaron, I'm see this one is still not applied. Maybe because there are still some small issues? See below. On 10/03/2010 11:12 PM, Yaron Shahrabani wrote: > +LANGUAGE LANG_HEBREW, SUBLANG_DEFAULT This line should go, it's redundant (it's already stated at the beginning). This line was present in the English resources because there is a difference between SUBLANG_DEFAULT and SUBLANG_NEUTRAL in that resource file. For example 'colors' vs. 'colours'. > + > +IDD_INDEX DIALOG 0, 0, 200, 190 > +STYLE DS_MODALFRAME | WS_CAPTION | WS_SYSMENU | WS_EX_LAYOUTRTL You should move the WS_EX_LAYOUTRTL to an EXSTYLE like the IDD_SEARCH_DIALOG below. Adding it to the STYLE has no effect. > +FONT 8, "MS Shell Dlg" > +CAPTION "מפתח" > +{ > + LISTBOX IDC_INDEXLIST, 10, 10, 180, 150, LBS_NOINTEGRALHEIGHT | WS_VSCROLL | WS_BORDER > +} > + > +IDD_SEARCH DIALOG 0, 0, 200, 190 > +STYLE DS_MODALFRAME | WS_CAPTION | WS_SYSMENU > +EXSTYLE WS_EX_LAYOUTRTL > +FONT 8, "MS Shell Dlg" -- Cheers, Paul.

Oct. 5, 2010

Re: user32/tests: Test FindWindowEx for empty window title (try 2)

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5829 Your paranoid android. === W98SE (32 bit win) === win.c:2970: Test failed: message 000f available win.c:2982: Test failed: message 000f available win.c:2991: Test failed: message 000f available win.c:3000: Test failed: message 000f available === W7PROX64 (64 bit win) === win.c:2684: Test failed: GetActiveWindow() = 0000000000000000 win.c:2684: Test failed: GetFocus() = 0000000000000000 win.c:2710: Test failed: GetActiveWindow() = 0000000000000000 win.c:2710: Test failed: GetFocus() = 0000000000000000 win.c:2769: Test failed: GetActiveWindow() = 0000000000000000 win.c:2769: Test failed: GetFocus() = 0000000000000000

Oct. 5, 2010

Re: [PATCH] shlwapi: Handle URL_WININET_COMPATIBLITY flag in UrlCanonicalize

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5828 Your paranoid android. === W98SE (32 bit url) === url.c:746: Test failed: UrlCanonicalizeA dwFlags 0x00000000 url 'c:dir\file' Expected "file:///c:dir/file", but got "file:///c:/dir/file", index 46

Oct. 5, 2010

winmm/tests: Don't test the non-default Midi devices if the 'PreferredOnly' Windows setting is set.

by Joerg-Cyril.Hoehle＠t-systems.com

Hi, a few month ago when the discussion about "PreferredOnly" registry setting came up, I queried Google and found that many more users use this setting than one may believe - i.e. it is not exotic. Therefore, I don't think it's that great an idea to reduce the number of devices that the tests stress using that method: we'll get less feedback from a number of real machines. OTOH, I have no solution to the vmware MIDI trouble on Francois' machines. Regards, Jörg Höhle

Oct. 5, 2010

Re: msi: Put the uninstall key for 32-bit packages under Wow6432Node on 64-bit.

by Hans Leidekker

On Tue, 2010-10-05 at 16:59 +0200, testbot(a)testbot.winehq.org wrote: > 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 > http://testbot.winehq.org/JobDetails.pl?Key=5824 > > Your paranoid android. > > > === W7PROX64 (64 bit install) === > install.c:10427: Test failed: key not created This failure existed before my patch.

Oct. 5, 2010

Re: msi: Put the uninstall key for 32-bit packages under Wow6432Node on 64-bit.

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5824 Your paranoid android. === W7PROX64 (64 bit install) === install.c:10427: Test failed: key not created

Oct. 5, 2010

Re: [2/3] kernel32/process: Implement GetLogicalProcessorInformation using NtQuerySystemInformation

by Vitaliy Margolen

On 10/05/2010 04:46 AM, Rudolf Mayerhofer wrote: > --- > dlls/kernel32/process.c | 21 ++++++++++++++++++--- > 1 files changed, 18 insertions(+), 3 deletions(-) > Please exchange patches 2 and 3. And everywhere in your patches, please stick to _THE SAME_ formatting as the rest of the file you editing. Examples: > + if (status != STATUS_SUCCESS) > + { > + if( status == STATUS_INFO_LENGTH_MISMATCH ) Here you adding 2 differently formatted if () statements. > +static ULONG_PTR sysfs_cpu_topology_getmap(int cpu, const char *type) > +{ > + char buf[256]; > + FILE *fd; > + ULONG_PTR result; > + sprintf(buf, "/sys/devices/system/cpu/cpu%d/topology/%s",cpu,type); Please add empty line between variable declarations and the rest of the code. All commas should always be followed by a space. > + fd = fopen(buf,"r"); > + if(fscanf(fd,"%lx",&result) != 1) You are not checking if fopen() succeeded. > +static int sysfs_cpucache_getint(int cpu, int cache, const char *type) > +{ > + }else{ > + val = -1; > + } > + fclose(fd); > + return val; > +} > ... > + BYTE level = sysfs_cpucache_getint(cpucount,cachecount,"level"); Don't mix signed and unsigned integers. Especially that you return -1 in case of error. In that entire file curly braces go on a separate line. Vitaliy.

Oct. 5, 2010

Re: [1/2] user32: fix FindWindowEx behavior if window name is an empty charachter

by Dmitry Timoshkov

Louis Lenders <xerox_xerox2000(a)yahoo.co.uk> wrote: > >It would be better to avoid duplicate GetWindowTextW() calls. > > I thought a long time how to fix this, but i couldn't find another way than > this one, because we need to find out whether or not the window title is > empty, and afaik we can only do that by calling GetWindowText (again). If > you have another way that i overlooked, please let me know I sent 2 patches that make your FindWindow() tests pass under Wine. Did you run the tests after your fix? They don't pass without patching explorer. Also your test has a typo in the comment, and probably the tests need to go into their own test function, and not hijack the message window tests. -- Dmitry.

Oct. 5, 2010

Re: winex11: Add support for animated cursors in X11 driver (try 5).

by Alexandre Julliard

Erich Hoover <ehoover(a)mines.edu> writes: > + imgs[nFrames++] = image; > + if (!(imgstmp = HeapReAlloc( GetProcessHeap(), 0, imgs, nFrames*sizeof(XcursorImage*) ))) goto cleanup; > + imgs = imgstmp; This is still broken. -- Alexandre Julliard julliard(a)winehq.org

Oct. 5, 2010

Re: ImmReleaseContext() function implementation

by Aric Stewart

I do not believe this is correct. You would be destroying the context every time it is released and that is clearly not what we want to be doing ass ImmGetContext and ImmReleaseContext as a pair are called over and over again in a typical application. -aric On 10/4/10 1:25 PM, Mikhail Gorbunov wrote: > This patch implements (with a basic sanity check) a behaviour of > ImmReleaseContext() > function as described in msdn and in the way ImmGetContext() has already been > impleneted. > > From 56f78bcd9584350b456ca0b0514a357b7ee8cdfd Mon Sep 17 00:00:00 2001 > From: Mikhail Gorbunov<mikhail.gorbunov(a)gmail.com> > Date: Mon, 4 Oct 2010 21:47:17 +0400 > Subject: [PATCH] ImmReleaseContext() function implementation > > --- > dlls/imm32/imm.c | 24 ++++++++++++++++++------ > 1 files changed, 18 insertions(+), 6 deletions(-) > > diff --git a/dlls/imm32/imm.c b/dlls/imm32/imm.c > index 28f482c..87b4c14 100644 > --- a/dlls/imm32/imm.c > +++ b/dlls/imm32/imm.c > @@ -2009,13 +2009,25 @@ BOOL WINAPI ImmRegisterWordW( > */ > BOOL WINAPI ImmReleaseContext(HWND hWnd, HIMC hIMC) > { > - static int shown = 0; > + TRACE("(%p, %p)\n", hWnd, hIMC); > > - if (!shown) { > - FIXME("(%p, %p): stub\n", hWnd, hIMC); > - shown = 1; > - } > - return TRUE; > + if (hIMC) > + { > + InputContextData *data = hIMC; > + > + if (data->IMC.hWnd != hWnd) > + { > + FIXME("(%p, %p): data->IMC.hWnd = %p :: err\n", > + hWnd, hIMC, data->IMC.hWnd); > + return TRUE; > + } > + else > + { > + return ImmDestroyContext(hIMC); > + } > + } > + else > + return TRUE; > } > > /***********************************************************************

Oct. 5, 2010

Re: [1/2] user32: fix FindWindowEx behavior if window name is an empty charachter

by Ricardo Filipe

2010/10/5 Louis Lenders <xerox_xerox2000(a)yahoo.co.uk>: >>It would be better to avoid duplicate GetWindowTextW() calls. > > I thought a long time how to fix this, but i couldn't find another way than this one, because we need to find out whether or not the window title is empty, and afaik we can only do that by calling GetWindowText (again). If you have another way that i overlooked, please let me know > > Regards Louis > > > > > > doesn't it work if you store the GetWindowText result in a variable and check that variable in the condition? or you need the two calls to have the right behaviour (seems unlikely...)?

Oct. 5, 2010

Re: [PATCH 1/6] urlmon: Report error on synchronous bindings that don't have a cache file and are not read in callback.

by Nikolay Sivov

On 10/5/2010 14:09, Alexandre Julliard wrote: > Jacek Caban<jacek(a)codeweavers.com> writes: > >> --- >> dlls/urlmon/binding.c | 17 ++++++++++++----- >> 1 files changed, 12 insertions(+), 5 deletions(-) > This breaks the gameux tests: > > ../../../tools/runtest -q -P wine -M gameux.dll -T ../../.. -p gameux_test.exe.so gameexplorer.c&& touch gameexplorer.ok > gameexplorer.c:571: Test failed: IGameExplorer::AddGame failed (error 0x80004005) > gameexplorer.c:589: Test failed: IGameExplorer::AddGame failed (error 0x80004005) > gameexplorer.c:640: Test failed: IGameExplorer2::InstallGame failed (error 0x80004005) > gameexplorer.c:643: Test failed: cannot find game with GDF path L"Z:\\home\\julliard\\wine\\wine\\dlls\\gameux\\tests\\gameux_test.exe" in scope 2 > make: *** [gameexplorer.ok] Error 4 ..by breaking msxml document loading most likely.

Oct. 5, 2010

Re: [PATCH 1/6] urlmon: Report error on synchronous bindings that don't have a cache file and are not read in callback.

by Alexandre Julliard

Jacek Caban <jacek(a)codeweavers.com> writes: > --- > dlls/urlmon/binding.c | 17 ++++++++++++----- > 1 files changed, 12 insertions(+), 5 deletions(-) This breaks the gameux tests: ../../../tools/runtest -q -P wine -M gameux.dll -T ../../.. -p gameux_test.exe.so gameexplorer.c && touch gameexplorer.ok gameexplorer.c:571: Test failed: IGameExplorer::AddGame failed (error 0x80004005) gameexplorer.c:589: Test failed: IGameExplorer::AddGame failed (error 0x80004005) gameexplorer.c:640: Test failed: IGameExplorer2::InstallGame failed (error 0x80004005) gameexplorer.c:643: Test failed: cannot find game with GDF path L"Z:\\home\\julliard\\wine\\wine\\dlls\\gameux\\tests\\gameux_test.exe" in scope 2 make: *** [gameexplorer.ok] Error 4 -- Alexandre Julliard julliard(a)winehq.org

Oct. 5, 2010

[1/2] user32: fix FindWindowEx behavior if window name is an empty charachter

by Louis Lenders

>It would be better to avoid duplicate GetWindowTextW() calls. I thought a long time how to fix this, but i couldn't find another way than this one, because we need to find out whether or not the window title is empty, and afaik we can only do that by calling GetWindowText (again). If you have another way that i overlooked, please let me know Regards Louis

Oct. 5, 2010

Re: kernel32/process: Implement GetLogicalProcessorInformation (rewrite - try 3)

by Rudolf Mayerhofer

> Have you checked that it doesn't just return a > SYSTEM_LOGICAL_PROCESSOR_INFORMATION ? It seems to be the right > "format" for the name anyway. I'd play with the size argument or try > using NtQuerySystemInformationEx (some new thing in win7). You could > just find a case that passes and implement it for wine (i.e. play with > parameters in tests until it passes on windows), since there doesn't > seem to be any public docs for the api anyway. We can worry about > "correct" implementation once someone figures it out (or uses it at > least). > > Mike. Based on your suggestion i've rewritten the test example MSDN provides for GetLogicalProcessorInformation to use NtQuerySystemInformation/SystemLogicalProcessorInformation instead and that returns exactly the same result as the original one. Based on that we can assume that the buffers returned by both calls are identical and just arrays of SYSTEM_LOGICAL_PROCESSOR_INFORMATION.

Oct. 5, 2010

Re: [1/2] user32: fix FindWindowEx behavior if window name is an empty charachter

by Dmitry Timoshkov

Louis Lenders <xerox_xerox2000(a)yahoo.co.uk> wrote: > @@ -1756,6 +1756,10 @@ HWND WINAPI FindWindowExW( HWND parent, HWND child, LPCWSTR className, LPCWSTR t > { > while (list[i]) > { > + /* Handle special case where empty title bar or no title bar exists, so GetWindowsText would return NULL > + but if title is also an empty char we still have a match */ > + if (!GetWindowTextW( list[i], buffer, len + 1 ) && !title[0]) break; > + > if (GetWindowTextW( list[i], buffer, len + 1 ) && !strcmpiW( buffer, title )) break; > i++; > } It would be better to avoid duplicate GetWindowTextW() calls. -- Dmitry.

Oct. 5, 2010

Re: kernel32/process: Implement GetLogicalProcessorInformation (rewrite - try 3)

by Mike Kaplinskiy

On Mon, Oct 4, 2010 at 2:18 PM, Rudolf Mayerhofer <rm(a)eightyfive.net> wrote: >>I don't know much about kernel32/ntdll, but I'm pretty sure that you >>could get this to work through ntdll. NtQuerySystemInformation with >>SystemProcessorInformation/SystemLogicalProcessorInformation (see >>http://undocumented.ntinternals.net/UserMode/Undocumented%20Functions/Syste…) >>should help. http://native-nt-toolkit.googlecode.com/svn/trunk/ndk/extypes.h >>(read, don't copy) seems to have a pretty exhaustive set of >>definitions, but even that's not complete. Some tests for the NT >>version would be useful. >> >>Mike. > > SystemLogicalProcessorInformation seems to be the right SystemClass for > NtQuerySystemInformation. But here's the Problem. I don't have any idea how i > could any information about the struct returned by calling > NtQuerySystemInformation with SystemLogicalProcessorInformation (= 73). > > The only windows machine i have access to a the moment keeps throwing status > 0xC0000005 (ACCESS_VIOLATION) at me which doesn't help at all and i'm pretty > much stuck right now. If i could get my hands on the struct returned by > NtQuerySystemInformation it should be easy to change the patch to do the > parsing in ntdll and just call that from kernel32. > > > Have you checked that it doesn't just return a SYSTEM_LOGICAL_PROCESSOR_INFORMATION ? It seems to be the right "format" for the name anyway. I'd play with the size argument or try using NtQuerySystemInformationEx (some new thing in win7). You could just find a case that passes and implement it for wine (i.e. play with parameters in tests until it passes on windows), since there doesn't seem to be any public docs for the api anyway. We can worry about "correct" implementation once someone figures it out (or uses it at least). Mike.

Oct. 5, 2010

Re: psapi/tests: macro cleanup

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5813 Your paranoid android. === WVISTAADM (32 bit psapi_main) === psapi_main.c:111: Test failed: cbNeeded=100 === W2K8SE (32 bit psapi_main) === psapi_main.c:111: Test failed: cbNeeded=100 === W7PRO (32 bit psapi_main) === psapi_main.c:111: Test failed: cbNeeded=96 === W7PROX64 (32 bit psapi_main) === psapi_main.c:111: Test failed: cbNeeded=108

Oct. 5, 2010

Re: [PATCH] comctl32/tests: Add tests for negative subitems independent of local settings

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5811 Your paranoid android. === W7PROX64 (64 bit listview) === listview.c:4449: Test failed: NM_HOVER allow test: the msg sequence is not complete: expected 0087 - actual 0000 listview.c:4454: Test failed: NM_HOVER block test: the msg sequence is not complete: expected 0087 - actual 0000

Oct. 5, 2010

Re: [2/2] user32/tests: Test FindWindowEx for empty window title

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5803 Your paranoid android. === W7PROX64 (32 bit win) === win.c:2684: Test failed: GetActiveWindow() = 00000000 win.c:2684: Test failed: GetFocus() = 00000000 win.c:2710: Test failed: GetActiveWindow() = 00000000 win.c:2710: Test failed: GetFocus() = 00000000 win.c:2769: Test failed: GetActiveWindow() = 00000000 win.c:2769: Test failed: GetFocus() = 00000000 === W7PROX64 (64 bit win) === win.c:2684: Test failed: GetActiveWindow() = 0000000000000000 win.c:2684: Test failed: GetFocus() = 0000000000000000 win.c:2710: Test failed: GetActiveWindow() = 0000000000000000 win.c:2710: Test failed: GetFocus() = 0000000000000000 win.c:2769: Test failed: GetActiveWindow() = 0000000000000000 win.c:2769: Test failed: GetFocus() = 0000000000000000

Oct. 5, 2010

Re: [PATCH 2/3] urlmon: Added Abort tests.

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5782 Your paranoid android. === W2K3R2SESP2 (32 bit protocol) === protocol.c:824: Test failed: hrResult = 800c0005, expected: 00000000 Timeout === WVISTAADM (32 bit protocol) === protocol.c:509: Test failed: unexpected call Switch

Oct. 5, 2010

Re: [PATCH 3/3] urlmon: Added Abort support to BindProtocol.

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5783 Your paranoid android. === WINEBUILD (build) === Patch failed

Oct. 5, 2010

Re: cryptui: Fix a crash with certification path

by Juan Lang

Hi André, I see your patch didn't get committed. In case it's waiting for a comment from me, it looks fine: I assume you've imported a certificate that isn't otherwise trusted, and viewing it results in a crash? I don't see the crash here. Thanks, --Juan

Oct. 4, 2010

Re: kernel32/process: Implement GetLogicalProcessorInformation (rewrite - try 3)

by Rudolf Mayerhofer

>I don't know much about kernel32/ntdll, but I'm pretty sure that you >could get this to work through ntdll. NtQuerySystemInformation with >SystemProcessorInformation/SystemLogicalProcessorInformation (see >http://undocumented.ntinternals.net/UserMode/Undocumented%20Functions/Syste…) >should help. http://native-nt-toolkit.googlecode.com/svn/trunk/ndk/extypes.h >(read, don't copy) seems to have a pretty exhaustive set of >definitions, but even that's not complete. Some tests for the NT >version would be useful. > >Mike. SystemLogicalProcessorInformation seems to be the right SystemClass for NtQuerySystemInformation. But here's the Problem. I don't have any idea how i could any information about the struct returned by calling NtQuerySystemInformation with SystemLogicalProcessorInformation (= 73). The only windows machine i have access to a the moment keeps throwing status 0xC0000005 (ACCESS_VIOLATION) at me which doesn't help at all and i'm pretty much stuck right now. If i could get my hands on the struct returned by NtQuerySystemInformation it should be easy to change the patch to do the parsing in ntdll and just call that from kernel32.

Oct. 4, 2010

Re: include: Added d3d10_1.idl.

by Rico Schüller

Am 04.10.2010 18:35, schrieb Jacek Caban: > On 10/3/10 5:44 PM, Rico Schüller wrote: >> Am 02.10.2010 15:29, schrieb Jacek Caban: >> >> +typedef enum D3D10_SRV_DIMENSION1 >> +{ >> + D3D10_1_SRV_DIMENSION_UNKNOWN = 0, >> ... >> + D3D10_1_SRV_DIMENSION_TEXTURECUBE = 9, >> + D3D10_1_SRV_DIMENSION_TEXTURECUBEARRAY = 10 >> +} D3D10_SRV_DIMENSION1; >> >> You probably might add those to d3dcommon.h and only add a typedef. >> It was changed in one of the sdks. > > This can't work for enums. Note that there are D3D10_1_* value names, > not only types. Yeah sure, you need to add it as enums. See e.g. http://msdn.microsoft.com/en-us/library/ff728736%28VS.85%29.aspx . I'm not sure if this could be a problem, but if you include only d3d11.h which includes d3dcommon.h and you use e.g. D3D10_1_SRV_DIMENSION_TEXTURECUBE a build against our headers might fail because D3D10_1_SRV_DIMENSION_TEXTURECUBE is not defined (what a bad example). Well written apps would of course use D3D11_SRV_DIMENSION_TEXTURECUBE or D3D_SRV_DIMENSION_TEXTURECUBE in that case. Note: This is mostly the same for the current d3d10.h, but the stuff was moved to d3dcommon after our d3d10.h was in! Cheers Rico

Oct. 4, 2010

Re: include: Added d3d10_1.idl.

by Jacek Caban

On 10/3/10 5:44 PM, Rico Schüller wrote: > Am 02.10.2010 15:29, schrieb Jacek Caban: > > +typedef enum D3D10_SRV_DIMENSION1 > +{ > + D3D10_1_SRV_DIMENSION_UNKNOWN = 0, > ... > + D3D10_1_SRV_DIMENSION_TEXTURECUBE = 9, > + D3D10_1_SRV_DIMENSION_TEXTURECUBEARRAY = 10 > +} D3D10_SRV_DIMENSION1; > > You probably might add those to d3dcommon.h and only add a typedef. It > was changed in one of the sdks. This can't work for enums. Note that there are D3D10_1_* value names, not only types. Jacek

Oct. 4, 2010

Re: winspool.drv: return maximum possible size of the buffer when no buffer specified

by Vitaly Perov

On Monday 04 October 2010 19:03:21 Alexandre Julliard wrote: > Vitaly Perov <vitperov(a)etersoft.ru> writes: > > @@ -1398,7 +1398,7 @@ static void convert_printerinfo_W_to_A(LPBYTE out, > > LPBYTE pPrintersW, dmA = DEVMODEdupWtoA(piW->pDevMode); > > if (dmA) { > > /* align DEVMODEA to a DWORD boundary */ > > - len = (4 - ( (DWORD_PTR) ptr & 3)) & 3; > > + len = ptr ? (4 - ( (DWORD_PTR) ptr & 3)) & 3 : > > 3; > > That doesn't make any sense. May be I misunderstand something. The bug is following: CODE: GetJobW(hPrinter, job->job_id, 2, NULL, 0, &needed); FIXME("needed = %d\n",needed); ji2 = HeapAlloc(GetProcessHeap(), 0, needed); result = GetJobW(hPrinter, job->job_id, 2, (LPBYTE)ji2, needed, &needed); FIXME("needed = %d\n",needed); CONSOLE OUTPUT: fixme:winspool:ScheduleJob needed = 366 fixme:winspool:ScheduleJob needed = 368 and result=false; The patch fixes this bug. -- Best regards, Vitaly Perov

Oct. 4, 2010

Re: psapi/tests: remove an ugly macro

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5776 Your paranoid android. === WVISTAADM (32 bit psapi_main) === psapi_main.c:116: Test failed: cbNeeded=100 === W2K8SE (32 bit psapi_main) === psapi_main.c:116: Test failed: cbNeeded=100 === W7PRO (32 bit psapi_main) === psapi_main.c:116: Test failed: cbNeeded=96 === W7PROX64 (32 bit psapi_main) === psapi_main.c:116: Test failed: cbNeeded=108

Oct. 4, 2010

Re: winspool.drv: return maximum possible size of the buffer when no buffer specified

by Alexandre Julliard

Vitaly Perov <vitperov(a)etersoft.ru> writes: > @@ -1398,7 +1398,7 @@ static void convert_printerinfo_W_to_A(LPBYTE out, LPBYTE pPrintersW, > dmA = DEVMODEdupWtoA(piW->pDevMode); > if (dmA) { > /* align DEVMODEA to a DWORD boundary */ > - len = (4 - ( (DWORD_PTR) ptr & 3)) & 3; > + len = ptr ? (4 - ( (DWORD_PTR) ptr & 3)) & 3 : 3; That doesn't make any sense. -- Alexandre Julliard julliard(a)winehq.org

Oct. 4, 2010

Re: cmd: Add a space at the end of the first echo'ed batch line [try2]

by Alexandre Julliard

Alexandre Goujon <ale.goujon(a)gmail.com> writes: > -(a)pwd@>echo Tests for cmd's builtin commands(a)todo_space@ > +(a)pwd@>echo Tests for cmd's builtin commands You'll need to define a @space@ symbol, normal trailing spaces will get mangled. -- Alexandre Julliard julliard(a)winehq.org

Oct. 4, 2010

Re: New version of exe-thumbnailer (0.7)

by Damjan Jovanovic

On Mon, Oct 4, 2010 at 4:07 AM, Nicholas van Oudtshoorn <vanoudt(a)gmail.com> wrote: > On 10/04/2010 08:39 AM, Octavian Voicu wrote: >> >> On Fri, Oct 1, 2010 at 11:18 PM, Scott Ritchie <scott(a)open-vote..org >> <mailto:scott(a)open-vote.org>> wrote: >> >> Includes much prettier icons, support for Vista icons (with icoutils >> 0.29.1) and more. >> >> Screenshots and a description are at: >> http://wiki.winehq.org/exe-thumbnailer >> >> >> Nice! Any plans for Kubuntu / KDE support? :) >> >> Octavian >> >> >> >> > > Nice! > > Just to note, KDE/Dolphin already has built-in support for thumnbailing exe > files. (Although, in Fedora at least, it's not enabled by default; just tell > dolphin to show previews for "Microsoft Windows Executables". > > I've also got code on the KDE reviewboard that'll take care of LNK files - > if it gets picked up, that'll be natively supported by KDE too... > > Nicholas How does thumbnailing LNK files work? You get a Windows icon path out, which you have to convert to a *nix path - you'd have to use ~/.wine/dosdevices and do case-insensitive filename matching, and even that would only work for LNK files created by Wine, those on a Windows partition have Windows's idea of drive mappings which is probably different from Wine. Damjan Jovanovic

Oct. 4, 2010

Re: winex11: choose header exclusiv

by Alexandre Julliard

André Hentschel <nerv(a)dawncrow.de> writes: > @@ -29,8 +29,7 @@ > > #ifdef HAVE_X11_EXTENSIONS_XF86VMODE_H > #include <X11/extensions/xf86vmode.h> > -#endif > -#ifdef HAVE_X11_EXTENSIONS_XF86VMPROTO_H > +#elif defined (HAVE_X11_EXTENSIONS_XF86VMPROTO_H) > #include <X11/extensions/xf86vmproto.h> > #endif This doesn't make sense, all it will most likely achieve is to disable the gamma support. -- Alexandre Julliard julliard(a)winehq.org

Oct. 4, 2010

Re: New version of exe-thumbnailer (0.7)

by Octavian Voicu

On Mon, Oct 4, 2010 at 5:07 AM, Nicholas van Oudtshoorn <vanoudt(a)gmail.com>wrote: > Just to note, KDE/Dolphin already has built-in support for thumnbailing exe > files. (Although, in Fedora at least, it's not enabled by default; just tell > dolphin to show previews for "Microsoft Windows Executables". > > I've also got code on the KDE reviewboard that'll take care of LNK files - > if it gets picked up, that'll be natively supported by KDE too... > I got it to work, but it's a bit strange since I must activate Preview for the folder to actually get the icons. I agree with Scott, the gnome version looks better and is more useful :)

Oct. 4, 2010

Re: New version of exe-thumbnailer (0.7)

by Scott Ritchie

On 10/03/2010 07:07 PM, Nicholas van Oudtshoorn wrote: > On 10/04/2010 08:39 AM, Octavian Voicu wrote: >> On Fri, Oct 1, 2010 at 11:18 PM, Scott Ritchie <scott(a)open-vote..org >> <mailto:scott(a)open-vote.org>> wrote: >> >> Includes much prettier icons, support for Vista icons (with icoutils >> 0.29.1) and more. >> >> Screenshots and a description are at: >> http://wiki.winehq.org/exe-thumbnailer >> >> >> Nice! Any plans for Kubuntu / KDE support? :) >> >> Octavian >> >> >> >> > > Nice! > > Just to note, KDE/Dolphin already has built-in support for thumnbailing > exe files. (Although, in Fedora at least, it's not enabled by default; > just tell dolphin to show previews for "Microsoft Windows Executables". > > I've also got code on the KDE reviewboard that'll take care of LNK files > - if it gets picked up, that'll be natively supported by KDE too... > > Nicholas > > > Patches welcome :) (I think the exe-thumbnailer I got produces better images than KDE's built in stuff, fwiw) How are you parsing the LNK files for icons? Thanks Scott Ritchie

Oct. 4, 2010

Re: mono packaging, for wine packagers or anyone else who wants to follow along

by Scott Ritchie

On 10/01/2010 01:19 PM, Vincent Povirk wrote: > Download and extract the mono 2.6.7 source tarball from > http://ftp.novell.com/pub/mono/sources/mono/mono-2.6.7.tar.bz2 > > Use the instructions at http://wiki.winehq.org/Mono to cross-compile > Mono 2.6.7 using mingw32. Tweak the script if you have to (or just > want to), but consider contributing any fixes you have to make back to > winezeug. > I will note that packagers can do the "lazy way" and just follow the instructions to build a binary and then package that. I'm doing this for Ubuntu (for now), as Ubuntu doesn't provide all the Mingw libraries needed to get the package to build. So in the meantime Ubuntu users will get a binary package, and once I'm done packing up the Mingw libraries they'll get a source package that actually builds the binary properly. And thank you Vincent, this is great stuff :) -Scott Ritchie

Oct. 4, 2010

Re: New version of exe-thumbnailer (0.7)

by Nicholas van Oudtshoorn

On 10/04/2010 08:39 AM, Octavian Voicu wrote: > On Fri, Oct 1, 2010 at 11:18 PM, Scott Ritchie <scott(a)open-vote..org > <mailto:scott(a)open-vote.org>> wrote: > > Includes much prettier icons, support for Vista icons (with icoutils > 0.29.1) and more. > > Screenshots and a description are at: > http://wiki.winehq.org/exe-thumbnailer > > > Nice! Any plans for Kubuntu / KDE support? :) > > Octavian > > > > Nice! Just to note, KDE/Dolphin already has built-in support for thumnbailing exe files. (Although, in Fedora at least, it's not enabled by default; just tell dolphin to show previews for "Microsoft Windows Executables". I've also got code on the KDE reviewboard that'll take care of LNK files - if it gets picked up, that'll be natively supported by KDE too... Nicholas

Oct. 4, 2010

Re: New version of exe-thumbnailer (0.7)

by Octavian Voicu

On Fri, Oct 1, 2010 at 11:18 PM, Scott Ritchie <scott(a)open-vote.org> wrote: > Includes much prettier icons, support for Vista icons (with icoutils > 0.29.1) and more. > > Screenshots and a description are at: > http://wiki.winehq.org/exe-thumbnailer Nice! Any plans for Kubuntu / KDE support? :) Octavian

Oct. 4, 2010

Re: Question on use of TextMetricsW

by James McKenzie

On 10/3/10 11:55 AM, Paul Vriens wrote: > On 10/03/2010 08:44 PM, James McKenzie wrote: >> All: >> >> I am running code like the following in a richedit control to set the >> font to Courier 12 point font: >> >> >> + static const WCHAR courier[] = {'C','o','u','r','i','e','r',0}; >> + testFont1 = CreateFontW (12,0,0,0,FW_LIGHT, 0, 0, 0, ANSI_CHARSET, >> + OUT_DEFAULT_PRECIS, CLIP_DEFAULT_PRECIS, DEFAULT_QUALITY, >> DEFAULT_PITCH | >> + FF_DONTCARE, courier); >> + >> + /*Set test font to be Courier font */ >> + SendMessageW(hwndRichEdit, WM_SETFONT, (WPARAM)testFont1, >> MAKELPARAM(TRUE, 0)); >> + /*Verify font is set */ >> + SendMessageW(hwndRichEdit, EM_GETCHARFORMAT, SCF_DEFAULT, (LPARAM) >> &CharFont1Unicode); >> + GetObjectW(testFont1, sizeof(LOGFONTW), &sentLogFont); >> + ok (!lstrcmpW(sentLogFont.lfFaceName, CharFont1Unicode.szFaceName), >> "WM_SETFONT set wrong font. " \ >> + "Font was set to %s or %s, was supposed to be \"Courier\".\n", >> wine_dbgstr_w(CharFont1Unicode.szFaceName), >> wine_dbgstr_w(sentLogFont.lfFaceName)); >> + >> + ret = GetTextMetricsW (hdc, &tmw); >> + ok (ret, "GetTextMetricsW failed\n"); >> + >> + ok (7 == tmw.tmAveCharWidth, "Average Character Width for Courier 12 >> is %d\n", tmw.tmAveCharWidth); >> >> I also set the font to MS Sans Serif, and two different font sizes for >> Arial. However, the tmAveCharWidth value remains at 7 and I am of the >> opinion that this should change for each font change and also for the >> font size change. Am I doing something improper? Should I clear the tmw >> structure between calls to GetTextMetricsW? >> >> When I run this on native WindowsXP (which I have access to) the average >> character width changes to 9, but remains the same for all four font >> types. >> TTIA. >> >> James McKenzie >> >> >> > Hi James, > > Doesn't the WM_SETFONT only set it for the control? You are retrieving > the font characteristics from the DC though. > You probably are correct as one of the other 'metrics' is the Character Set which returned 0 for both Courier and MS San Serif. I'll look at a different way to get back the information I was looking for. Thank you for the correction on the use of GetTextMetricsW. James McKenzie

Oct. 4, 2010

Re: kernel32/process: Implement GetLogicalProcessorInformation (rewrite - try 3)

by Mike Kaplinskiy

On Sun, Oct 3, 2010 at 10:25 AM, Rudolf Mayerhofer <rm(a)eightyfive.net> wrote: >>A rewrite of my previously sent patches in order to avoid changing >>SYSTEM_CPU_INFORMATION. > >>This patch supersedes my previously sent patches: >>[PATCH1/3] include/winternl: Add Logical Processor Information to >>SYSTEM_CPU_INFORMATION (try 3) >>[PATCH2/3] kernel32/process: Implement GetLogicalProcessorInformation() (try > 2) >>[PATCH3/3] ntdll: Retrieve Logical Processor Information from SysFS (try 4) > > Try 2 removes a FIXME which is not needed. > Try 3 changes a FIXME to a WARN when an unknown cache type is found > > > --- > dlls/kernel32/process.c | 272 > ++++++++++++++++++++++++++++++++++++++++++++++- > 1 files changed, 270 insertions(+), 2 deletions(-) > > > > > I don't know much about kernel32/ntdll, but I'm pretty sure that you could get this to work through ntdll. NtQuerySystemInformation with SystemProcessorInformation/SystemLogicalProcessorInformation (see http://undocumented.ntinternals.net/UserMode/Undocumented%20Functions/Syste…) should help. http://native-nt-toolkit.googlecode.com/svn/trunk/ndk/extypes.h (read, don't copy) seems to have a pretty exhaustive set of definitions, but even that's not complete. Some tests for the NT version would be useful. Mike.

Oct. 3, 2010

Re: [3/5] mscoree: Add test for creating CLRMetaHost.

by Greg Geldorp

> 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 > http://testbot.winehq.org/JobDetails.pl?Key=5753 > > Your paranoid android. > > === WINEBUILD (build) === > Make in include dir failed Caused by a problem in TestBot, should be fixed now. Greg.

Oct. 3, 2010

Translation patches for wine-1.2.1

by Michael Stefaniuc

Hello Alexandre, attached is the list with the translation patches to cherry-pick for wine-1.2.1. The list was generated with the attached crud patch (modified from the version I used for 1.0.1) and then I have manually removed all my janitorial patches that didn't produce conflicts. Though the removal of the unneeded "Remove the ignored common resource attributes" patches makes "wrc --verify-translation" produce DIFF lines for resources that were added at a later time as those don't have the ignored common resource attributes. bye michael

Oct. 3, 2010

Re: Question on use of TextMetricsW

by James McKenzie

On 10/3/10 11:55 AM, Paul Vriens wrote: > On 10/03/2010 08:44 PM, James McKenzie wrote: >> All: >> >> I am running code like the following in a richedit control to set the >> font to Courier 12 point font: >> >> >> + static const WCHAR courier[] = {'C','o','u','r','i','e','r',0}; >> + testFont1 = CreateFontW (12,0,0,0,FW_LIGHT, 0, 0, 0, ANSI_CHARSET, >> + OUT_DEFAULT_PRECIS, CLIP_DEFAULT_PRECIS, DEFAULT_QUALITY, >> DEFAULT_PITCH | >> + FF_DONTCARE, courier); >> + >> + /*Set test font to be Courier font */ >> + SendMessageW(hwndRichEdit, WM_SETFONT, (WPARAM)testFont1, >> MAKELPARAM(TRUE, 0)); >> + /*Verify font is set */ >> + SendMessageW(hwndRichEdit, EM_GETCHARFORMAT, SCF_DEFAULT, (LPARAM) >> &CharFont1Unicode); >> + GetObjectW(testFont1, sizeof(LOGFONTW), &sentLogFont); >> + ok (!lstrcmpW(sentLogFont.lfFaceName, CharFont1Unicode.szFaceName), >> "WM_SETFONT set wrong font. " \ >> + "Font was set to %s or %s, was supposed to be \"Courier\".\n", >> wine_dbgstr_w(CharFont1Unicode.szFaceName), >> wine_dbgstr_w(sentLogFont.lfFaceName)); >> + >> + ret = GetTextMetricsW (hdc, &tmw); >> + ok (ret, "GetTextMetricsW failed\n"); >> + >> + ok (7 == tmw.tmAveCharWidth, "Average Character Width for Courier 12 >> is %d\n", tmw.tmAveCharWidth); >> >> I also set the font to MS Sans Serif, and two different font sizes for >> Arial. However, the tmAveCharWidth value remains at 7 and I am of the >> opinion that this should change for each font change and also for the >> font size change. Am I doing something improper? Should I clear the tmw >> structure between calls to GetTextMetricsW? >> >> When I run this on native WindowsXP (which I have access to) the average >> character width changes to 9, but remains the same for all four font >> types. >> TTIA. >> >> James McKenzie >> >> >> > Hi James, > > Doesn't the WM_SETFONT only set it for the control? You are retrieving > the font characteristics from the DC though. > Even if I am retrieving the DC for the control through: hdc = GetDC(hwndRichEdit) and hwndRichEdit is the handle to the control? Just trying to get a grasp on where GetTextMetricsW is getting its information from. James McKenzie

Oct. 3, 2010

Re: [5/5] mscoree: Add a test that enumerates .NET runtimes.

by testbot＠testbot.winehq.org

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 http://testbot.winehq.org/JobDetails.pl?Key=5754 Your paranoid android. === WINEBUILD (build) === Make in include dir failed

Oct. 3, 2010