July 2008 - Wine-Devel - list.winehq.org

re: Tests and Valgrinding

by Dan Kegel

Zachary Goldberg wrote: > test.winehq.org ... is fantastic, but it doesn't give a very good feel for progress > that all of these recent patches have been making. I would love to see a feature added to test.winehq.org which showed one tests's results across time (and platform). That would let you see regressions (and progress) at a glance. > Also, Dan Kegel has been extraordinarily diligent in valgrinding of > late (/applause) and I just wanted to ask if we're doing anything to > support him? A few people are starting to run Valgrind themselves. How about this: if you ever post a patch that causes a valgrind warning, you send yourself to valgrind jail for a month, and valgrind every patch you write before you send it in... > perhaps making Valgrind-passing a requirement (along > with not breaking tests) for being committed. I think perhaps > we should take a look at this and seriously consider implementing it. As a first step, James and I could (along with posting the errors to wine-devel) also post patches to revert the guilty changes. (And if we do it quickly enough, perhaps Alexandre can do the revert on the same day...) Then the author could fix and resubmit. How's that sound? - Dan

July 2, 2008

Re: secur32: Allow loading external schannel.dll.

by Huang, Zhangrong

Hi, 2008/6/25 Austin Lund <austin.lund(a)gmail.com>: > I cannot get this to work with googletalk.exe. No luck, googletalk.exe still can't login. > > I get the following messages: > > trace:secur32:SECUR32_initializeProviders > trace:secur32:_tryLoadProvider loaded L"schannel.dll", > InitSecurityInterfaceA is 0x7ec2fae0, InitSecurityInterfaceW is > 0x7ec2fb00 > trace:secur32:_tryLoadProvider L"schannel.dll" has built-in providers, > skip adding This message means secur32 is loading schannel.dll, and find built-in providers. > > Even though I have removed schannel.dll completely from the libraries. > > Then the rest of the messages come up reporting stubs. To get googletalk.exe work, first step is let secur32 load schannel.dll which has LSA/user mode functions tables, next need to implement the stubs in schannel.dll. -- $ apt-get moo (__) (oo) /------\/ / | || * /\---/\ ~~ ~~ ...."Have you mooed today?"...

July 2, 2008

Re: winecfg.cpl: Add new control panel applet for winecfg

by Owen Rudge

> Wine's builtin control does just launch the shell's version of the > control panel, unless it's given an argument like "DESKTOP", > "DATE/TIME", etc. See programs/control.c. Indeed, I'm aware of that. Effectively, there are two versions of the control panel in the shell of course, the namespace version and the built-in simple version. The latter didn't check the registry for control panels - I've just implemented such a check, which will fix this bug. I'll submit a patch shortly. > You might try to make some of those that have an analogue work in > Wine, especially the "Legacy Control Panel Commands" "desktop" and > "color". That's something I can look into, indeed. Cheers, -- Owen Rudge http://www.owenrudge.net/

July 2, 2008

Re: winecfg.cpl: Add new control panel applet for winecfg

by Juan Lang

> I'll look into that, then. I know that the shell namespace control panel > does check the registry for entries (could you try using a program like the > ReactOS Explorer to browse the control panel and see if it shows up there?), > but I'm not sure if the standalone control.exe does. Ideally, the standalone > control.exe would be replaced with a stub that launches the Explorer, but of > course Wine doesn't have a built-in Explorer yet. Wine's builtin control does just launch the shell's version of the control panel, unless it's given an argument like "DESKTOP", "DATE/TIME", etc. See programs/control.c. Those string arguments to control aren't functional, because the control panel applets they reference don't exist in Wine. Launching control panel items via name is described in MSDN: http://msdn.microsoft.com/en-us/library/bb776838.aspx See especially "Executing Control Panel Items", http://msdn.microsoft.com/en-us/library/bb776841(VS.85).aspx You might try to make some of those that have an analogue work in Wine, especially the "Legacy Control Panel Commands" "desktop" and "color". --Juan

July 2, 2008

Re: gdi32: Implement automatic font substitution

by Huang, Zhangrong

Hi, 2008/7/2 Liu Qishuai <lqs.buaa(a)gmail.com>: > No. > > 1. In Windows XP, I deleted the FontLink register key and rebooted, > and the Chinese characters are still displaying correctly. I guess gdi32.dll has hard-coded the default FontLink registry, of course no prove. > > 2. I moved simsun.ttc (The default font in Simplified Chinese) to > another directory and rebooted. Windows XP automatically uses > simhei.ttf (another Chinese font) to display Chinese characters. There > is no register key for this font. This doesn't prove anything, creating a non-existent font will return mostly related font. On Japanese Windows, if user select a Latin font in notepad, CJK chars will display incorrect. And I have a Simplified Chinese Windows XP box, which also displays CJK chars as squares, I have no idea what's wrong. -- $ apt-get moo (__) (oo) /------\/ / | || * /\---/\ ~~ ~~ ...."Have you mooed today?"...

July 2, 2008

New valgrind warnings in msi/db

by Dan Kegel

Hi James, could you have a look at yet more fresh valgrind warnings, triggered by http://source.winehq.org/git/wine.git/?a=commit;h=d9ac95be5bb51de2293326920… ? You might want to invest in a copy of Valgrind yourself sometime. I can get you a 10% off discount :-) http://kegel.com/wine/valgrind/logs-2008-07-02/vg-msi_db-diff.txt + Conditional jump or move depends on uninitialised value(s) + at msihandle2msiinfo (handle.c:161) + by MsiViewExecute (msiquery.c:467) + by test_storages_table (db.c:6058) + by func_db (db.c:6154) + by run_test (test.h:449) + by main (test.h:498) + Uninitialised value was created by a stack allocation + at test_storages_table (db.c:6000) ... + Use of uninitialised value of size 4 + at MsiCloseHandle (handle.c:290) + by test_storages_table (db.c:6112) + by func_db (db.c:6154) + by run_test (test.h:449) + by main (test.h:498) + Uninitialised value was created by a stack allocation + at test_storages_table (db.c:6000) ... + 8 bytes in 1 blocks are definitely lost + at notify_alloc (heap.c:191) + by RtlReAllocateHeap (heap.c:1428) + by msi_realloc (msipriv.h:962) + by msi_update_table_columns (table.c:1065) + by msi_table_load_transform (table.c:2531) + by msi_table_apply_transform (table.c:2653) + by MSI_DatabaseApplyTransformW (msiquery.c:727) + by MsiDatabaseApplyTransformW (msiquery.c:756) + by MsiDatabaseApplyTransformA (msiquery.c:773) + by test_try_transform (db.c:2431) + by func_db (db.c:6133) + by run_test (test.h:449) + by main (test.h:498)

July 2, 2008

Re: gdi32: Implement automatic font substitution

by Liu Qishuai

No. 1. In Windows XP, I deleted the FontLink register key and rebooted, and the Chinese characters are still displaying correctly. 2. I moved simsun.ttc (The default font in Simplified Chinese) to another directory and rebooted. Windows XP automatically uses simhei.ttf (another Chinese font) to display Chinese characters. There is no register key for this font. 2008/7/2 Huang, Zhangrong <hzhrong(a)gmail.com>: > Hi, >> * Append all fonts on child_list so that if a character is not available in the font, >> * other fonts will be automatically used. This is how Windows does to get a CJK character >> * when a latin font is specified. > > That's not correct, actually Windows does this by using FontLink > registry, so does wine now. Your patch seems to break FontLink. > "Automatic font substitution" often means wine will choose the CJK > replacement font which user doesn't prefer to. > > On CJK env, if a latin font doesn't not have a FontLink, it will > fallback to use the FontLink of "Microsoft Sans Serif", so user can > set preferred font there. > > 2008/7/2 Liu Qishuai <lqs.buaa(a)gmail.com>: >> 2008/7/2 Dan Kegel <dank(a)kegel.com>: >>> Can you write a conformance test that verifies this behavior? >> >> Sorry, I don't know how to write a test yet. Maybe I will write a test later. >> >>> Also, can you identify a bug in bugzilla this will fix? >> >> Such as bug 4065, 10864, 14151. > > Please set the following registry key, it should fix those bugs: > > [Software\\Microsoft\\Windows > NT\\CurrentVersion\\FontLink\\SystemLink] 1208451304 > "Microsoft Sans > Serif"=str(7):"simsun.ttc,SimSun\0gulim.ttc,Gulim\0msgothic.ttc,MS UI > Gothic\0mingliu.ttc,PMingLiU\0" > "Tahoma"=str(7):"simsun.ttc,SimSun\0gulim.ttc,Gulim\0msgothic.ttc,MS > UI Gothic\0mingliu.ttc,PMingLiU\0" > > Maybe we need a patch to set up the default FontLink of "Microsoft > Sans Serif" and "Tahoma" now, since FontLink works fine on WIne. > > -- > $ apt-get moo > (__) > (oo) > /------\/ > / | || > * /\---/\ > ~~ ~~ > ...."Have you mooed today?"... >

July 2, 2008

Re: Splitting up ole component

by Dan Kegel

On Wed, Jul 2, 2008 at 6:17 AM, Rob Shearman <robertshearman(a)gmail.com> wrote: > How about these: > > ole32: > OLE (dlls/{ole32,oleacc,olecli32,oledlg,olepro32,olesvr32,olethk32}) > oleaut32: > OLE Automation (dlls/{oleaut32,stdole2.tlb,stdole32.tlb}) > rpc: > DCE/RPC (dlls/rpcrt4, tools/widl and widl-generated files) Done. If they look ok now, we can start moving bugs from ole to one of these three new categories; once they're all moved, we can delete the old ole category.

July 2, 2008

Re: rpcrt4: Fix a test that fails in win2k3

by Rob Shearman

2008/7/2 Kai Blin <kai.blin(a)gmail.com>: > On Wednesday 02 July 2008 11:52:11 Rob Shearman wrote: > >> I think the behaviour on win2k3 should be marked as broken, since I >> can't see any reason why it would need to return more bytes than was >> requested. > > Hm, I seem to recall a Samba RPC fix doing the same thing. This seems to be a > byte alignment thing. If you're interested, I can see if I can dig up the > appropriate discussion on samba-technical. Sure, although I suspect it will be related to packets with authentication, rather than without which is the case here. In any case, alignment of the end of the packet should be a task handled by the RPC layer, not the NDR layer. -- Rob Shearman

July 2, 2008

Re: [quartz/tests] Make tests run on everything lower then Vista (Try 2)

by Alexander Dorofeyev

Paul Vriens wrote: > I've checked this on W2K3 and WinXP-SP3 and it only creates that one > registry > key with one value (Default) underneath. There is another key created which > isn't deleted: > > HKEY_CURRENT_USER\Software\Microsoft\ActiveMovie\devenum\{guid} > > This key has 4 values. > > This key is overwritten when a new registration occurs Hmm, during the course of the test there should definitely be keys created in HKCR\CLSID\{guid} by filtermapper and they were during my tests. I guess they are removed by Unregister... calls (can be checked by commenting them). On a second thought, I guess the UnregisterFilter also should remove Pins subkey, probably that's why there's nothing left there. Can't recheck myself right now, unfortunately. Then deletekey vs deletetree only matters when something in unregistration unexpectedly fails, but still. I'll look into other keys, I guess it's some leftovers from filtermapper2 stuff and more cleanup must be added.

July 2, 2008