February 2007 - Wine-Devel - list.winehq.org

Re: [PATCH 3/6] oleaut32: Comment out "dual" attribute of IKindaEnumWidget in tmarshal test to cause the interface to be marshaled correctly.
by Alexandre Julliard Feb. 7, 2007

Feb. 7, 2007

Robert Shearman <rob(a)codeweavers.com> writes: > This fixes a bunch of tmarshal tests. It doesn't seem to make any difference for me: ../../../tools/runtest -q -P wine -M oleaut32.dll -T ../../.. -p oleaut32_test.exe.so tmarshal.c && touch tmarshal.ok tmarshal.c:139: Test failed: CoMarshalInterface failed with error 0x80040155 tmarshal.c:765: Test failed: CoUnmarshalInterface failed with error 0x80040155 tmarshal.c:784: Test failed: EXCEPINFO differs from expected: wCode = 0x2, scode = 0x7bc7a2c8 tmarshal.c:803: Test failed: EXCEPINFO differs from expected: wCode = 0x2, scode = 0x7bc7a2c8 tmarshal.c:915: Test failed: EXCEPINFO differs from expected: wCode = 0x2, scode = 0x7bc7a2c8 tmarshal.c:942: Test failed: EXCEPINFO differs from expected: wCode = 0x2, scode = 0x7bc7a2c8 tmarshal.c:977: Test failed: EXCEPINFO differs from expected: wCode = 0x2, scode = 0x80004001 make: *** [tmarshal.ok] Error 7 -- Alexandre Julliard julliard(a)winehq.org

2 2

Re: msi: Register the typelib
by Alexandre Julliard Feb. 7, 2007

Feb. 7, 2007

Mike McCormack <mike(a)codeweavers.com> writes: > +ITypeLib *get_msi_typelib( LPWSTR *path ) > +{ > + static CRITICAL_SECTION cs = { NULL, -1, 0, 0, 0, 0 }; Please use the standard pattern for static critical sections. Setting the debug info to NULL is reserved for global sections. -- Alexandre Julliard julliard(a)winehq.org

1 0

msi: Make MsiInstallProduct conformance test depend on proper UI level processing.
by Misha Koshelev Feb. 7, 2007

Feb. 7, 2007

Hi, just wanted to double-check with msi ppl and other wine developers if it's okay to add my conformance test to the test_MsiInstallProduct function. My reasons for doing this versus a completely separate test function are as follows: 1. The test is already testing a full product install (installs files, services, registry values, etc) and my modification will simply result in either no modification to these actions if UI level processing is good or complete failure if it is bad, which will make the problem very easy to diagnose. 2. Looking through instructions on msdn it seems that there isn't a kosher way to make a very simple MSI file that just, say, writes a registry value (without doing costing, install validation, having a full features table, installing some features, etc.) which would make a separate function easy to implement and it does not really seem necessary to copy and paste the whole test_MsiInstallProduct (or similar) function just to check UI level processing. I could do a separate function, it just seems unnecessary. Misha p.s. For some reason, I can't compile the msi crosstests (it gives an error on linking with -lcabinet, whereas, say, kernel32 crosstests compile file), so if anyone can crosstest it for me that would be great :)

1 1

Re: [urlmon 1/2] Cast-qual warning fix
by Joris Huizer Feb. 7, 2007

Feb. 7, 2007

--- Paul Vriens <paul.vriens.wine(a)gmail.com> wrote: > Hi, > > Fix a warning and remove the now not needed cast. > > Changelog > Cast-qual warning fix > > Cheers, > > Paul. > > > > diff --git a/dlls/urlmon/umon.c b/dlls/urlmon/umon.c > index 49260ab..d593c81 100644 > --- a/dlls/urlmon/umon.c > +++ b/dlls/urlmon/umon.c > @@ -1605,7 +1605,7 @@ HRESULT WINAPI > URLDownloadToCacheFileW(LPUNKNOWN lpUnkCaller, > LPCWSTR szURL, LPW > HRESULT hr; > LPWSTR ext; > > - static const WCHAR header[] = { > + static WCHAR header[] = { > 'H','T','T','P','/','1','.','0',' > ','2','0','0',' ', > > 'O','K','\\','r','\\','n','\\','r','\\','n',0 > }; > @@ -1631,7 +1631,7 @@ HRESULT WINAPI > URLDownloadToCacheFileW(LPUNKNOWN lpUnkCaller, > LPCWSTR szURL, LPW > modified.dwLowDateTime = 0; > > if (!CommitUrlCacheEntryW(szURL, cache_path, > expire, modified, NORMAL_CACHE_ENTRY, > - (LPWSTR)header, > sizeof(header), NULL, NULL)) > + header, > sizeof(header), NULL, NULL)) > return E_FAIL; > > if (lstrlenW(cache_path) > dwBufLength) > -- If CommitUrlCacheEntryW does not take a constant pointer, this might be dangerous - if it changes the string, the next time this code executes the wrong (modified) string gets used If this is indeed possible, the static keyword needs to be dropped too (ensuring the string is reinitialised with the next call) If it is only not possible in this particular case, sorry for complaining; (though it may be possible to factor out that part of the code that is used, thus ensuring constness) If this is not possible in general, the function CommitUrlCacheEntryW should take a constant pointer instead (same for the other patch) regards, Joris ____________________________________________________________________________________ Do you Yahoo!? Everyone is raving about the all-new Yahoo! Mail beta. http://new.mail.yahoo.com

1 0

Re: comctl32: tab: update the item mask after a TB_SETITEM (fixes bug #7345)
by Felix Nawothnig Feb. 6, 2007

Feb. 6, 2007

Mikołaj Zalewski wrote: > Tab uses a mask to mark which fields are set but it new fields were set > by SetItem the mask wasn't updated. Actually all members are always valid (for the internal structure at least). I already sent a patch which removes the mask. Felix

2 1

qemu-0.9.0 and kqemu-1.3.0pre10 (now GPL) released
by Detlef Riekenberg Feb. 6, 2007

Feb. 6, 2007

http://www.qemu.org/ QEMU 0.9.0 was released. The QEMU Accelerator Module kqemu-1.3.0pre10 was released and is now available under the GPL license. Thanks Fabrice, the qemu-team and every contributor. -- By by ... Detlef

1 0

Re: msi: InstallPackage check for UI level must not disregard flags (updated)
by James Hawkins Feb. 6, 2007

Feb. 6, 2007

On 2/5/07, Misha Koshelev <mk144210(a)bcm.tmc.edu> wrote: > This is an update, a little less hacky thanks to advice from Dmitry > Timoshkov. This fixes bug #6992. The installer for Vector NTI 10 was > getting to the > point where it was trying to launch the .msi files that do the actual > install, but was quitting with the message "This module is not designed > for direct execution." Install worked fine with native MSI. After a > while of playing with it and disassembling the MSI file with dark.exe > tool, I noticed that the .msi files ran in execute mode but not UI mode. > Finally, I checked the UI level that wine msi was reading, and it was > 102 which is INSTALLUILEVEL_NONE | INSTALLUILEVEL_SOURCESONLY. However, > MSI_InstallPackage was only checking if the uilevel >= > INSTALLUILEVEL_REDUCED, which would have been true in this case as > INSTALLUILEVEL_SOURCESONLY > INSTALLUILEVEL_REDUCED but the actual ui > level is INSTALLUILEVEL_NONE which is < INSTALLUILEVEL_REDUCED. This > patch fixes this check, and now the installer finished successfully > (although it seems upon a quick check that there are more install bugs > for this program in wine, as not all files that should be there seem to > be there after this install). > > Changelog: > > * msi: InstallPackage check for UI level must not disregard > INSTALLUILEVEL flags. > > + if ( (msi_get_property_int(package, szUILevel, 0) & 0xf) >= INSTALLUILEVEL_REDUCED ) > Magic numbers are bad. What is 0xF? -- James Hawkins

4 7

Window focus testing
by Duane Clark Feb. 6, 2007

Feb. 6, 2007

Is it possible to do window focus testing in Wine conformance tests? Or does the windows not being mapped mean that this testing cannot be done? Or maybe I am doing something completely wrong; which is likely ;) I am specifically trying to test window focus in this bug: http://bugs.winehq.org/show_bug.cgi?id=7280 As mentioned in the bug, I turned the treeview conformance test into a standalone test with a mapped window, and added testing that does show focus behavior. But so far I have not been able to get it to give good results in a Wine conformance test.

3 2

Re: [PATCH] make CarbonPoker (previously Poker.com) client work.
by Juan Lang Feb. 6, 2007

Feb. 6, 2007

Hi Trent, thanks for the patches. Please send the split up patches to wine-patches instead of wine-devel. The crypt32 patches won't make it in as-is, but I think you can get a better quality hack accepted. As MSDN states, CryptGetMessageCertificates is just a wrapper around CertOpenStore. So you could write something like the following: HCERTSTORE WINAPI CryptGetMessageCertificates( DWORD dwMsgAndCertEncodingType, HCRYPTPROV hCryptProv, DWORD dwFlags, const BYTE* pbSignedBlob, DWORD cbSignedBlob) { CRYPT_DATA_BLOB blob = { cbSignedBlob, (LPBYTE)pbSignedBlob }; TRACE("%d, %p, %d, %p, %d\n", dwMsgAndCertEncodingType, hCryptProv, dwFlags, pbSignedBlob, cbSignedBlob); return CertOpenStore(CERT_STORE_PROV_PKCS7, dwMsgAndCertEncodingType, hCryptProv, dwFlags, &blob); } Then in store.c, add a stub function for opening and closing PKCS7 stores - look at CRYPT_PhysOpenStore for an example. Cheers, --Juan ____________________________________________________________________________________ Get your own web address. Have a HUGE year through Yahoo! Small Business. http://smallbusiness.yahoo.com/domains/?p=BESTDEAL

1 0

Linuxtag 2007 in Berlin
by Stefan Munz Feb. 6, 2007

Feb. 6, 2007

Hi, there is a "CALL FOR PROJECTS" for Linuxtag 2007 in Germany/Berlin from May 30 to June 2. Is anybody interested in participating in a Wine booth? More information (in german) is here: http://www.linuxtag.org/2007/de/community/projects/cfpro/call-for-projects.… Projects should be registered until February 23. cu, Stefan

3 2