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

Re: WineD3D: request alpha in backbuffer mode, to work correctly with multiple opengl pixel formats
by Stefan Dösinger Feb. 22, 2008

Feb. 22, 2008

Am Freitag, 22. Februar 2008 23:31:43 schrieb Roderick Colenbrander: > Hi, > > In the default backbuffer based renderer in WineD3D the front and > backbuffer share the same pixel format while D3D itself decouples them. A > lot of applications use alpha on the backbuffer, so make sure we have > alpha. This change is needed because the single pixel format which we had > before always had alpha. > + /* In case of ORM_BACKBUFFER, make sure to request an alpha component for X4R4G4B4/X8R8G8B8 as we might need it for the backbuffer. */ It would be for offscreen buffers, not the back buffer. I'm sorry for the confusion in the earlier mail

1 0

Re: 1/2 WGL: remove the pixel format limitation (!)
by Stefan Dösinger Feb. 22, 2008

Feb. 22, 2008

Am Freitag, 22. Februar 2008 21:16:23 schrieb Roderick Colenbrander: > WineD3D has some small issues due to this change but those should be > investigated by Stefan, Chris, me and others. The code should get a lot of > testing as soon as possible. We discussed this on IRC. The problem is that we're still defaulting to back buffer / aux buffer offscreen rendering, and thus offscreen targets are bound to use the same format as the back buffer. In the D3D9 test this means that the backbuffer is X8R8G8B8, and the onscreen target is A8R8G8B8. The alpha is missing with this patch(used to be always there), so the test fails. The proper fix is to switch to FBOs for onscreen rendering whereever possible(FBO support still has problems that prevent that though), and on cards without FBOs try to use A8R8G8B8 instead of X8R8G8B8 for the backbuffer, and similarly for X1R5G5B5 and X4R4G4B4.

1 1

Re: [Wine] Windows Kernel & Executive implementation
by Volodymyr Shcherbyna Feb. 22, 2008

Feb. 22, 2008

Unfortunately, that code just do stubs, and it is not actually executed in kernel mode. -- with best regards, Volodymyr 2008/2/22, Dan Kegel <dank(a)kegel.com>: > > On Fri, Feb 22, 2008 at 3:02 AM, Volodymyr Shcherbyna > <v_scherbina(a)mvps.org> wrote: > > > I discovered that WineHQ project does not have support for drivers (am > I > > wrong?). In other words, quite a big amount of applications will fail > to > > function properly, because many of them are using helper device drivers > to > > get some extended functionality. I can just mention RegMon, FileMon, > TDIMom, > > and others. > > > I know very little about this, but my impression is that > Wine supports a limited set of NT os kernel APIs already. > The initial support was added by this patch: > http://www.winehq.org/pipermail/wine-cvs/2007-May/032546.html > The current set of supported calls can be seen here: > http://source.winehq.org/source/dlls/ntoskrnl.exe/ntoskrnl.c > > Feel free to implement more, or file enhancement requests > describing what nt os apis a popular app needs. > > - Dan >

1 0

Re: msi: MsiMergeDatabaseA/W stub
by James Hawkins Feb. 22, 2008

Feb. 22, 2008

On Thu, Feb 21, 2008 at 2:26 PM, Johan Dahlin <jdahlin(a)async.com.br> wrote: > Add stubs for MsiMergeDatabaseA/W, used by WiX's msi validation. > > diff --git a/dlls/msi/database.c b/dlls/msi/database.c > index 425486e..e74db8f 100644 > --- a/dlls/msi/database.c > +++ b/dlls/msi/database.c > @@ -1119,3 +1119,22 @@ HRESULT create_msi_remote_database( IUnknown *pOuter, LPVOID *ppObj ) > > return S_OK; > } > + > +UINT WINAPI MsiDatabaseMergeW( MSIHANDLE hDatabase, > + MSIHANDLE hDatabaseMerge, > + LPCWSTR szTableName ) > +{ > + FIXME("(%ld %ld %s): stub!\n", hDatabase, hDatabaseMerge, > + debugstr_w(szTableName)); > + return ERROR_SUCCESS; > +} > + > +UINT WINAPI MsiDatabaseMergeA( MSIHANDLE hDatabase, > + MSIHANDLE hDatabaseMerge, > + LPCSTR szTableName ) > +{ > + FIXME("(%ld %ld %s): stub!\n", hDatabase, hDatabaseMerge, > + szTableName); > + > + return ERROR_SUCCESS; > +} > Please use the same style as the rest of the file (4 space indentation). The A version should be before the W version. -- James Hawkins

2 2

Re: [1/2] user32:scroll: Added some tests for scrollbar [try2]
by Anatoly Lyutin Feb. 22, 2008

Feb. 22, 2008

Alexandre Julliard wrote: > The tests fail: > > ../../../tools/runtest -q -P wine -M user32.dll -T ../../.. -p > user32_test.exe.so scroll.c && touch scroll.ok > scroll.c:56: Test failed: The scrollbar window should be disabled. > scroll.c:122: Test failed: The ShowScrollBar() should failed. > Yes! These tests must be failed in WINE, but it passes clear in Win XP. I just want to show you some incorrect behavior of scrollbar in WINE. Should I mark these tests as "todo_wine"? -- Best regards Anatoly Lyutin.

2 2

Re: [2/3] qmgr: Implement IUnkown interface for IBackgroundCopyJob.
by Alexandre Julliard Feb. 22, 2008

Feb. 22, 2008

Dan Hipschman <dsh(a)linux.ucla.edu> writes: > +static ULONG WINAPI BITS_IBackgroundCopyJob_AddRef(IBackgroundCopyJob* iface) > +{ > + ICOM_THIS_MULTI(BackgroundCopyJobImpl, lpVtbl, iface); > + return InterlockedIncrement(&This->ref); > +} There's no reason to use something like ICOM_THIS_MULTI with a single interface. And if you really do have multiple interfaces, please use an inline function instead. -- Alexandre Julliard julliard(a)winehq.org

1 0

Printing images
by Alexander Demakin Feb. 22, 2008

Feb. 22, 2008

Hi all, I have some troubles with printing. BitBlt returns error when calling it for printer context due to implementation of PSDRV_SelectBitmap(). Does anybody know how to implement this function? Where can i find any documentation about it? Implementation of SelectBitmap() in EMFDRV and MFDRV just returns 0. Is it correct? Please, share your ideas. Thanks in advance, Demakin Alexander.

1 0

gdiplus: negative stride from dipCreateBitmapFromFile
by bill lam Feb. 22, 2008

Feb. 22, 2008

WHen I read a jpeg file of size about 600x400 pixels using gdiplus like this, GdipCreateBitmapFromFile fn BMP GdipGetImageWidth BMP w GdipGetImageHeight BMP h rt=. 0 0, w, h GdipBitmapLockBits BMP rt ImageLockModeRead PixelFormat32bppARGB DATA 'w h s f p r'=. DATA the member stride inside bitmapdata is negative, and its data buffer cannot be read successfully. when using native gdiplus that stride is positive, and it work ok. wine said itself cannot handle negitive stride yet so why itself create a bitmap wih negative stride in GdipCreateBitmapFromFile?

1 0

Re: New dll. Wbemprox.dll
by Alistair Leslie-Hughes Feb. 22, 2008

Feb. 22, 2008

Ivan Sinitsin wrote: > This patch add new library to the project wine. Name of the library > is - "wbemprox.dll".This library is a part of WMI technology. Now there four > interfaces are described - "IWbemLocator, IWbemServices, > IEnumWbemClassObject, IWbemClassObject". Almost all methods are stubs. > This library is used with some programs like "1C", to get some information > about system. > > static struct regsvr_interface const interface_list[] = { >+ { &IID_IWbemLocator, >+ "IWbemLocator", >+ NULL, >+ 1, /*ÞÉÓÌÏ ÍÅÔÏÄÏ×*/ Comment should be in English. >#define _WIN32_WINNT 0x0600 > >#ifndef __wbemprox_PRIVATE_H__ Defining _WIN32_WINNT doesn't seem right. This patches is large, try breaking it up. Best Regards Alistair Leslie-Hughes

1 0

Fun Project: Add dialogs to builtin IE
by Steven Edwards Feb. 22, 2008

Feb. 22, 2008

Hi, I submitted a proof of concept patch for this a week or so back that added a simple file->exit menu to the builtin Internet Explorer. I've spoken with Jacek a bit about it and it seems shdocvw.dll is the correct place to implement the UI layout if we are targeting IE versions less the 7. There seems to be a plugin interface the GUI uses for browser extensions but if a first pass I don't think we really need to support it. Unfortuntly due to time constraints I am not going to be able to work on this project so if there is anyone out there that wants to contribute with a simple fun project, please feel free to add menus, dialogs and the address bar. I'll try to help as best I can but I expect to be tied up for some time to come. Thanks Steven -- Steven Edwards "There is one thing stronger than all the armies in the world, and that is an idea whose time has come." - Victor Hugo

1 0

Re: msvcrt: Implemented wcscpy_s
by Dmitry Timoshkov Feb. 21, 2008

Feb. 21, 2008

"Alistair Leslie-Hughes" <leslie_alistair(a)hotmail.com> wrote: > --- a/dlls/msvcrt/msvcrt.spec > +++ b/dlls/msvcrt/msvcrt.spec > @@ -761,6 +761,7 @@ > @ cdecl wcscmp(wstr wstr) ntdll.wcscmp > @ cdecl wcscoll(wstr wstr) MSVCRT_wcscoll > @ cdecl wcscpy(ptr wstr) ntdll.wcscpy > +@ cdecl wcscpy_s(wstr long wstr) MSVCRT_wcscpy_s Type of the first parameter should be ptr, not wstr. -- Dmitry.

1 0

Re: appdb: Increase readability by using slightly softer colours on Warning and Notes
by Francois Gouget Feb. 21, 2008

Feb. 21, 2008

On Wed, 20 Feb 2008, Jeff Zaroyko wrote: > The default colours for warnings and notes on the appdb pages feel > rather strong, this patch changes the red and blue to be a bit softer. > > I think this makes it more readable. A comparison can be seen here: > http://jeffz.name/tmp/ In new.png the green looks darker than the others. Maybe it should be lightened up a bit too. -- Francois Gouget <fgouget(a)free.fr> http://fgouget.free.fr/ In a world without fences who needs Gates?

4 7

Re: dmband: Indirection levels fix
by Robert Shearman Feb. 21, 2008

Feb. 21, 2008

Andrew Talbot wrote: > diff --git a/dlls/dmband/band.c b/dlls/dmband/band.c > index 891fb5a..8b89573 100644 > --- a/dlls/dmband/band.c > +++ b/dlls/dmband/band.c > @@ -173,7 +173,7 @@ static HRESULT WINAPI IDirectMusicBandImpl_IDirectMusicObject_SetDescriptor (LPD > This->pDesc->ftDate = pDesc->ftDate; > if (pDesc->dwValidData & DMUS_OBJ_MEMORY) { > This->pDesc->llMemLength = pDesc->llMemLength; > - memcpy (This->pDesc->pbMemData, pDesc->pbMemData, sizeof (pDesc->pbMemData)); > + memcpy (&This->pDesc->pbMemData, &pDesc->pbMemData, sizeof (pDesc->pbMemData)); > } > if (pDesc->dwValidData & DMUS_OBJ_STREAM) { > /* according to MSDN, we copy the stream */ This isn't correct. Judging by the surrounding code, this should be allocating a block of memory of This->pDesc->pbMemData and then passing pDesc->llMemLength into memcpy, possibly validating that pDesc->llMemLength isn't greater than UINT_MAX to avoid an overflow. -- Rob Shearman

2 1

Re: [1/2] ddraw: release interfaces when exiting with error
by Stefan Dösinger Feb. 21, 2008

Feb. 21, 2008

Am Mittwoch, 20. Februar 2008 22:03:51 schrieben Sie: > --- > dlls/ddraw/texture.c | 2 ++ > 1 files changed, 2 insertions(+), 0 deletions(-) Just a general comment(not specific to comments): For such cleanup work it works quite nice to use goto(!), as shown below. Your patch is OK as it is, but I think its a good idea to use this scheme in all places because it helps to prevent new problems from coming in when the code is changed. void *a = NULL, *b = NULL, *c = NULL; a = malloc(10); if(!a) goto error; b = malloc(10); if(!b) goto error; c = malloc(10); if(!c) goto error; if(dosomethingelse() == FAIL) goto error; return SUCCESS; error: f(c) free(c); f(b) free(b); f(a) free(a); return FAIL;

1 0

Re: dplayx: Fix incorrect check in DirectPlayLobbyAImpl_EnumLocalApplications and IDirectPlayLobbyAImpl_EnumAddressTypes.
by Marcus Meissner Feb. 21, 2008

Feb. 21, 2008

On Thu, Feb 21, 2008 at 04:44:34PM +0000, Robert Shearman wrote: > > Dereferencing a function pointer and checking it against NULL doesn't make > any sense, so remove the checks. > - if( !lpEnumAddressTypeCallback || !*lpEnumAddressTypeCallback ) > + if( !lpEnumAddressTypeCallback ) It looks as if they were meant as &&, but that does not make much sense either. Ciao, Marcus

1 0

Re: James Hawkins : itircl: Add a stub implementation of itircl.dll.
by Detlef Riekenberg Feb. 21, 2008

Feb. 21, 2008

On Sa, 2008-02-16 at 09:57 -0600, Alexandre Julliard wrote: > Module: wine > Branch: master > Commit: 58b0686e951364686c3b5b55c18566df0960fd96 > URL: > http://source.winehq.org/git/wine.git/?a=commit;h=58b0686e951364686c3b5b55c… > > Author: James Hawkins <jhawkins(a)codeweavers.com> > Date: Fri Feb 15 17:39:46 2008 -0600 > > itircl: Add a stub implementation of itircl.dll. Hi James. While adding your Patch to http://wiki.winehq.org/Developers-Hints , I found this Page: http://www.helpware.net/aphelp/info.htm itircl.dll: HTML Help 1 (found on w2k and xp) itircl52.dll: MS Help 2 apircl.dll: Assistance Platform 1.0 (Vista) -- By by ... Detlef

1 0

Re: slc: Added stub for SLGetWindowsInformationDWORD
by Paul Vriens Feb. 21, 2008

Feb. 21, 2008

On Thu, Feb 21, 2008 at 10:20 AM, Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> wrote: > Hi, > > Changelog: > slc: Added stub for SLGetWindowsInformationDWORD > > Best Regards > Alistair Leslie-Hughes > Hi Alistair, I know that function descriptions should always be your own words but I guess this is also true for header files. The comments in slerror.h are exactly like the comments on MSDN: http://msdn2.microsoft.com/en-us/library/aa965835(VS.85).aspx Cheers, Paul.

3 2

Re: Adding devices in setupapi
by Erich Hoover Feb. 20, 2008

Feb. 20, 2008

> > I think it probably does something similar to: > http://www.codeproject.com/KB/system/RemoveDriveByLetter.aspx > > Except instead of mapping a drive to usb device, it does it the other > way around. > > After looking more into it it really looks like this is the thing it > does. (I need a bunch of hacks for it to work, but alas.) However it > doesn't get me much further since our IOCTL_STORAGE_GET_DEVICE_NUMBER > ioctl is a stub. I would need something that can at least guarantee > uniqueness per volume for it, perhaps a conversion of drive letter to > integer. Now I just need to figure that one out. > > Cheers, > Maarten. > > I do not believe the device number matching is the problem, I have been exploring this problem a little with RemoveDriveByLetter and I believe that it is not working because the SymbolicLink entry is not added as a "real" thing in Wine. Performing an experiment with accessing the same drive letter in three different ways (results below, last parameter is "hDrive") I see that only the call with the registry-returned value for the drive fails (even if it is added to "HKLM/System/MountedDevices"). I believe that a call to IoCreateSymbolicLink must be made in order for the name returned by the registry to be recognized as an alias for the drive. I think iTunes is probably going through the device tree looking for its vendor and product code. It should get this back from CM_Get_Device_ID, but I do not have Windows so I cannot fake iTunes out to confirm this. SetupDiGetDeviceInterfaceDetail '\\?\USBSTOR#Disk&Ven_Apple&Prod_iPod&Rev_1.62#000A27001420C124&0#{53f56307-b6bf-11d0-94f2-00a0c91efb8b}': -1 SetupDiGetDeviceInterfaceDetail '\\?\Volume{00000000-0000-0000-0000-000000000048}': 60 SetupDiGetDeviceInterfaceDetail '\\.\H:': 64 Device number comparison: 1 1

1 0

No new Wine Weekly News?
by Austin English Feb. 20, 2008

Feb. 20, 2008

Hasn't been a new Wine Weekly News in about a month...quite a bit has been going on, with CS2 working out of the box, test suite fixes, etc. Anyone know what's going on? -Austin

2 1

Re: appdb: database issues
by Chris Morgan Feb. 20, 2008

Feb. 20, 2008

In include/query.php in query_appdb() we only call mysql_connect() if the global $hAppdbLink isn't a valid resource. So, when a user connects to display a page the first query should open the connection and upon completion of the script $hAppdbLink will be freed and the connection should be closed. So I'm not sure that this is the cause of the out-of-connections issue although I don't see a reason why we shouldn't just let php take care of this internally and apply a variation of your change where we always attempt to connect and select the appropriate database and remove the global handle and the resource check etc. Chris On Wed, Feb 20, 2008 at 6:01 PM, Jeff Zaroyko <jeffzaroyko(a)gmail.com> wrote: > The mysql database for the appdb/bugzilla appears to have gone down again, > I decided to look at the connection code in the appdb, as the appdb is thought > to be the culprit. > > I came across the call to mysql_connect which is using the new_link parameter > set to true, which is supposedly to force a new connection to be established. > > So, I searched through the git commit history using gitk and saw that Paul van > Schayck made a change introducing this, apparently because mysql was having > stability issues, back in 2005. > > http://www.winehq.org/pipermail/wine-patches/2005-January/014985.html > > my understanding is that php will automagically reuse connections if this is not > set, would there by any objection to trialling a removal of this > option to see if > stability increases? > > > >

1 0

Kudos for Photoshop work
by Dan Kegel Feb. 20, 2008

Feb. 20, 2008

In http://linux.slashdot.org/comments.pl?sid=460616&cid=22493322, an anonymous coward said: --- snip --- Preamble: I'm a photographer needing to process tens of thousands of photos relatively swiftly. The functionality I need isn't all that advanced (curves, levels, an occasional straighten horizon (measure + arbitrary rotate), crop, unsharp mask, and sometimes an action to find edges, feather and apply unsharp mask on that), but being able to access and apply this functionality swiftly is an absolute must because of the volume of photos I deal with. Photoshop is optimized to perfection to allow a swift workflow, while the gimp seems optimized to perfection to hinder it. Focus is never where I need it, shortcuts to access tools don't work depending on which sub-window has focus, etc. So yes, I really need Photoshop. I last tried Photoshop 7 under wine about a year ago. It was functional to an amazing degree (for someone who'd never seen or used wine before), but the rough edges were slightly too rough for me to be able to switch to Linux fulltime. I could trigger a dozen crashes in Photoshop at will just by resizing panels and doing other simple things like that, the program didn't feel native (alt-tabbing would keep the panels in the foreground, obscuring other programs), and focus sometimes strayed, amongst other lesser (but still annoyingly noticeable) issues. I just tried the latest wine with these Google sponsored improvements, and wow. This is an amazing difference. Every single issue I saw a year ago is gone. Photoshop feels as responsive as it does under Windows (perhaps even more so), and I went through an hour long editing session without being slowed down or annoyed even once. As far as I'm concerned, Linux is now ready to become my main OS. Google: I don't like your lack of respect for my privacy, but for this work on Wine, I can say from the bottom of my heart: Thank you! --- snip --- Ah. Music to my ears :-) - Dan

1 0

Re: Adding devices in setupapi
by Juan Lang Feb. 20, 2008

Feb. 20, 2008

> How hard is it to create a stub forwarder dll for setupapi and cfgmgr32? Not sure what you mean exactly. cfgmgr32 should forward to setupapi already, and I sent a patch that does that for the few stub functions that it had. setupapi's CM_* stubs could be improved, but I don't think you'll be able to forward them to any native code to try out: I expect the Windows configuration manager is rather too low level to run under Wine. You'll probably want to create a hacky version of setupapi that creates a fake device tree, with a root, USB hub, and iPod device, just to see if that makes iTunes happy. Does that answer your question? --Juan

2 1

Re: James Hawkins : msi: Forward MsiSourceListEnumSourcesA to a stub implementation of MsiSourceListEnumSourcesW .
by Detlef Riekenberg Feb. 20, 2008

Feb. 20, 2008

On Mi, 2008-02-20 at 06:46 -0600, > +/****************************************************************** > + * MsiSourceListEnumSourcesA (MSI.@) > + */ > +UINT WINAPI MsiSourceListEnumSourcesW( Greetings from Copy & Paste :-) BTW: Congratulations James for you new Address -- By by ... Detlef

1 0

bugzilla on the fritz again
by Dan Kegel Feb. 20, 2008

Feb. 20, 2008

Just started getting this: Software error: Can't connect to the database. Error: Too many connections Is your database installed and up and running? Do you have the correct username and password selected in localconfig?

2 1

Adding devices to registry with the help of hal
by Christian Gmeiner Feb. 20, 2008

Feb. 20, 2008

Hi all, I want you to present my current proof-of-conecpt. It shows how we can use hal.c of explorer to create registry keys for devices. Keep in mind that it only creates one registry key for each usb device at the moment. Before I investigate more time in this, I want to get some feedback about my method. If you are happy with it, I will finish this suftt :) Greets, Christian

2 1