June 2008 - Wine-devel - list.winehq.org

Wine in the news
by Dan Kegel 06 Jun '08

06 Jun '08

http://www.australianit.news.com.au/story/0,24897,23797416-24169,00.html The new president of Red Hat, when asked about desktop linux, said: "My simple answer is until Apple's iTunes runs on consumer Linux desktops, it's not going to take off. I know you can run Wine, which runs Windows applications on Linux, so techies will beat me up on that, but fundamentally there are some major device driver issues with products that people like to use and this limits adoption. It's going to be slow until these issues are solved." http://news.zdnet.co.uk/software/0,1000000121,39429910,00.htm http://economyaustria.at/Text/?id=3363378 Vienna migrates some Linux workstations to Vista because a single app, Schlaumäuse, crashes under Wine. Looks like they send the software out upon request to teachers; see http://www.schlaumaeuse.de/ (Can somebody send me a copy?)

4 5

Bad free error involving shdocvw and COM?
by Dan Kegel 06 Jun '08

06 Jun '08

In shdocvw/tests/webbrowser.c, at the end of test_WebBrowser, the call IUnknown_Release(unk) results in the following Valgrind error: Invalid free() / delete / delete[] at notify_free (heap.c:199) by RtlFreeHeap (heap.c:1284) by SysFreeString (oleaut.c:196) by DocHost_Release (dochost.c:550) by WebBrowser_Release (webbrowser.c:153) by test_WebBrowser (webbrowser.c:2028) by func_webbrowser (webbrowser.c:2070) by run_test (test.h:421) by main (test.h:470) Address 0x7f01dc14 is 8 bytes after a block of size 28 alloc'd at notify_alloc (heap.c:191) by RtlAllocateHeap (heap.c:1231) by COMPOBJ_DllList_Add (compobj.c:910) by apartment_getclassobject (compobj.c:794) by get_inproc_class_object (compobj.c:2138) by CoGetClassObject (compobj.c:2238) by get_protocol_cf (session.c:111) by get_protocol_info (session.c:219) by parse_canonicalize_url (internet.c:63) by CoInternetParseUrl (internet.c:184) by URLMonikerImpl_Construct (umon.c:1039) by CreateURLMonikerEx (umon.c:1097) by CreateURLMoniker (umon.c:1123) by create_moniker (navigate.c:564) by bind_to_object (navigate.c:600) by navigate_bsc (navigate.c:657) by navigate_bsc_proc (navigate.c:679) by process_dochost_task (dochost.c:43) by shell_embedding_proc (oleobject.c:65) by ??? (library.h:163) Seems like an important bug to fix, since it could conceivably lead to heap corruption. This is all our code, no gecko involved, so we ought to be able to sort it, but it's beyond my ken.

1 0

[website] Fix typo in Spanish translation of home.template
by Ángel Guzmán Maeso 06 Jun '08

06 Jun '08

Fix several typos -- Shakaran

1 1

Double free error involving dsound and winealsa.drv
by Dan Kegel 06 Jun '08

06 Jun '08

Hmm, freeing a buffer that's already been freed is bad, is it showstopper bad? http://bugs.winehq.org/show_bug.cgi?id=13723

2 1

Re: user32: EnumDisplaySettingsExA: fill in all of lpDevmode->dmDeviceName
by Dmitry Timoshkov 06 Jun '08

06 Jun '08

"Dan Kegel" <dank(a)kegel.com> wrote: > 'Cause if we don't, code that expects it to be defined > is mildly surprised to find that bytes after the string's > terminating nul are undefined. > Fixes 16 Valgrind warnings of the form: > > Use of uninitialised value of size 4 > at mbstowcs_sbcs (mbtowc.c:78) > by wine_cp_mbstowcs (mbtowc.c:266) > by MultiByteToWideChar (locale.c:1846) > by GdiConvertToDevmodeW (driver.c:434) > by ChangeDisplaySettingsExA (sysparams.c:3011) > by ChangeDisplaySettingsA (sysparams.c:2980) > by test_ChangeDisplaySettingsEx (monitor.c:161) > Uninitialised value was created by a stack allocation > at test_ChangeDisplaySettingsEx (monitor.c:141) GdiConvertToDevmodeW should be fixed instead. -- Dmitry.

1 0

Re: kernel32: tests/path.c: Improve test for GetLongPathNameA
by James Hawkins 05 Jun '08

05 Jun '08

On Wed, Jun 4, 2008 at 8:38 PM, Zac Brown <zac(a)zacbrown.org> wrote: > Add test to check for GetLongPathNameA on GetModuleFileNameA's returned > string. > > Found via errors existing in shell32/tests/shelllink.c. Checked against > Windows XP and Windows 2000. > + pGetModuleFileNameA = + (void*)GetProcAddress( GetModuleHandleA("kernel32.dll"), "GetModuleFileNameA" ); GetModuleFileNameA is available on all platforms, and besides, you use it without checking if it's NULL. +/* Try getting the long path of your module name */ + if(pGetLongPathNameA) { + ok(pGetModuleFileNameA(NULL,tmpstr,sizeof(tmpstr)),"GetModuleFileNameA failed %s\n", + tmpstr); + todo_wine ok(pGetLongPathNameA(tmpstr,tmpstr1,MAX_PATH),"GetLongPathNameA failed %s\n", + tmpstr); + } What is this testing? Just that GetLongPathName doesn't fail in this case (I'm not even sure what 'this case' is specifically)? You should be checking the return value and the value of the out variable. -- James Hawkins

1 0

Re: gdi32/tests: Take into account several Win9x bugs in GetObject(). (take 3)
by Paul Vriens 05 Jun '08

05 Jun '08

Francois Gouget wrote: > --- > > This time without the bug that caused tests to fail in Wine and on > Windows XP. > > > + SetLastError(0xdeadbeef); > ret = GetBitmapBits(hbm, 0, NULL); > - ok(ret == bm.bmWidthBytes * bm.bmHeight, "%d != %d\n", ret, bm.bmWidthBytes * bm.bmHeight); > + gle=GetLastError(); > + ok(ret == bm.bmWidthBytes * bm.bmHeight || (ret == 0 && gle == ERROR_INVALID_PARAMETER /* Win9x */), "%d != %d\n", ret, bm.bmWidthBytes * bm.bmHeight); Any reason this one isn't put into a broken() for win9x (as you did with all the others). -- Cheers, Paul.

2 1

re: [PATCH] oleaut32: Fix valgrind warnings in vatype.c test
by Dan Kegel 05 Jun '08

05 Jun '08

Hi Jon, I guess http://www.winehq.org/pipermail/wine-patches/2008-May/055352.html must have been too scary to commit. I independently sent in a fix for the easy part, and it was committed as http://winehq.org/pipermail/wine-cvs/2008-June/043999.html The last part of your patch, having to do with SysAllocStringByteLen(), seems correct to me, offhand. Maybe if you clean up the comment a bit more, sprinkle sugar on it, and resend, it'll make it in this time. It's so gross, though, that it really should serve as a reminder to us to switch to IMalloc... - Dan

2 1

Re: [AppDB] Allow admins to remove their admin rights
by Alexander Nicolaysen Sørnes 04 Jun '08

04 Jun '08

På Onsdag 04 juni 2008 , 01:24:07 skrev Chris Morgan: > What new policy on dissidents? ;-) > > How does this patch enable admins to remove themselves? I'm afraid I > can't see how this change lets an admin remove their own admin status. > > Chris > > Yea, that particular piece of code isn't very clear on what it's doing. I found out that $aClean['iUserId'] would only be set when modifying another user, and it previously compared it to $oUser->iUserId before allowing the admin removal. Now we only check for admin rights, and the oUser->iUserId == aClean['iUserId'] check is performed later to see if we should redirect back to the user admin page. > > On Tue, Jun 3, 2008 at 6:34 AM, Alexander Nicolaysen Sørnes > > <alex(a)thehandofagony.com> wrote: > > This is in accordance with the AppDB's new policy on dissidents. :) > > > > > > Alexander N. Sørnes

2 3

Fixing crashes in Tests (OS version check)
by Alistair Leslie-Hughes 04 Jun '08

04 Jun '08

Hi, I've being investigating crashes that occur on Win98 when the tests run. I have fixed a couple already which where trivial. When I attempted to fix the snmpapi tests, I have found that i will end up removing alot of tests (~ 17). If I add a check for the windows version, I can skip the tests that will crash, also there is already tests with if(0) around them. What is the policy about checking the version of the OS before running certain tests? Best Regards Alistair Leslie-Hughes

8 11