November 2006 - Wine-Devel - list.winehq.org

Re: [AppDB] Allow maintainer request along with app submission

by Chris Morgan

On Wednesday 29 November 2006 10:08 am, Alexander Nicolaysen Sørnes wrote: > Allow the user to submit a maintainer request along with an application, by > ticking a checkbox. > If the application is accepted, the maintainer request is accepted as well; > and if it is deleted or marked as a duplicate, the request is deleted. > These maintainer requests are not displayed in the queue. > > Also, maintainer requests submitted by an administrator are now > automatically approved. > > > Regards, > > Alexander N. Sørnes + if($hResultMaint) + { + $oMaintainer = mysql_fetch_row($hResultMaint); + $oMaintainer = new Maintainer($oMaintainer[0]); + } We don't want to overload the $oMaintainer variable here. Also, we want to use the name of the field you are accessing and not the index so we aren't asking ourselves what $oMaintainer[0] means. Also you'll want to check that mysql_fetch_row() actually returns a row prior to creating a new maintainer based upon it. This logic isn't quite right: + /* Reject super maintainer request if app is deleted or a duplicate */ + if(($aClean['sSub'] == "Reject" || $aClean['sSub'] == "Delete") && $hResultMaint) + { + $oMaintainter->reject("Deleting super maintainer request along with application entry."); + } + Rejecting an application doesn't mean deleting it so we don't want to delete or reject the maintainer request right there. Users can modify and resubmit applications after they have been rejected. The application class already handles the case where an application is deleted by calling Maintainer::deleteMaintainersForApplication() from Application::Delete(). In application.php: + if(makesafe($_REQUEST['bSmaintainerRequest'])) We should make the variable explicit, bSuperMaintainerRequest makes the purpose of this variable more clear. It's probably better to add a new variable to the class that we document as being used during application creation. That way we can modify Application::GetOutputEditorValues() to retrieve the value from the $aValues array passed in. This avoids having to makesafe() a $_REQUEST value, something that we really don't want to do anymore because eventually I think we'll want to clear out all non-standard entries in $_REQUEST as an additional protection against attacks from poorly checked input. The text here in the queries is too specific: "AND queued = '?' AND maintainReason != 'This user submitted the application; auto-queued.' ORDER by submitTime"; The query will fail if the text changes and we should assume it will. We need to do something like: select count(*) as queued_maintainers from appMaintainers, appFamily where appFamily.queued = 'true' and appMaintainers.queued = 'true' and appMaintainers.appId = appFamily.appId and appMaintainers.iSubmitterId != appMaintainers.iUserId; We also need to add a comment to getQueuedMaintainerCount() and to the sql to note that we are excluding queued maintainers that have applications queued as we will automatically submit them when the application is submitted. Chris

Nov. 29, 2006

Re: mlang: Make user32 a delay load import

by Alexandre Julliard

Dmitry Timoshkov <dmitry(a)codeweavers.com> writes: > the only import from user32 which mlang uses is wsprintfW, and it's only used > by the DllRegisterServer code, so it's reasonable to move user32 to delay load > imports. As long as mlang imports ole32 which imports user32, there is nothing to gain by not importing user32 directly. -- Alexandre Julliard julliard(a)winehq.org

Nov. 29, 2006

Re: d3d [1]: Remove AddRef from IWineD3DDevice_GetVertexDeclaration.

by Alexandre Julliard

Stefan Dösinger <stefandoesinger(a)gmx.at> writes: > Well, GetParent isn't part of IUnknown(its a wined3d invention), so from the > COM point of view we're fine to do whatever we want with it. (Just > QueryInterface has the rule that it addrefs the returned interface) The COM rules are not limited to QueryInterface, any call that returns an interface pointer needs to addref it. -- Alexandre Julliard julliard(a)winehq.org

Nov. 29, 2006

Re: d3d [1]: Remove AddRef from IWineD3DDevice_GetVertexDeclaration.

by Stefan Dösinger

Am Dienstag 28 November 2006 22:05 schrieb Markus Amsler: > Stefan Dösinger wrote: > > I proposed getting rid of COM some time ago, but I didn't consider that > > we need some features like inheritance. > > > > I think we should keep COM, but get rid of the D3D-Specific Addrefs > > because they are different between the various versions d3d versions. So > > keep the AddRef in IUnknown methods(QueryInterface namely), and > > GetParent, but do not AddRef in d3d methods(SetTexture / GetTexture, ...) > > > > Opinions? > > From my point of view (solving d3dx implicit surface refcounts) the > GetParent AddRef is ugly. It spreads d3dx refcounting across d3dx and > wined3d. It would be easier to only do d3dx refcounting in d3dx. > > But that's only my pragmatic view, without further COM knowledge. Well, GetParent isn't part of IUnknown(its a wined3d invention), so from the COM point of view we're fine to do whatever we want with it. (Just QueryInterface has the rule that it addrefs the returned interface)

Nov. 29, 2006

Re: Mozilla Sunbird install - Mystery bug #6613

by Marcus Meissner

On Wed, Nov 29, 2006 at 01:39:22PM +0900, Mike McCormack wrote: > > Anybody have time to help triage an install problem? > > Sunbird (and FireFox 2) installs run in an endless loop for some people. > If you have time to try the install out and see if it works on your > system, then comment on bug 6613, it would help out. > > http://bugs.winehq.org/show_bug.cgi?id=6613 > > I'm interested in which distro you're using, which filesystem, free disk > space, locale, timezone, CPU (64bit or 32 bit?), etc. I see a nice crash in XCheckIfEvent, however in some kind of "crash handling" program? Wine-dbg>bt Backtrace: 1 0x60dddad7 XCheckIfEvent+0x17() in libx11.so.6 (0x0034e264) =>2 0x60d45cac process_events+0x8c(display=0x37, mask=0x40) [/home/marcus/projects/wine/dlls/winex11.drv/event.c:247] in winex11 (0x0034e324) 3 0x60d45f96 X11DRV_MsgWaitForMultipleObjectsEx+0xb6(count=0x3, handles=0x34e494, timeout=0x7530, mask=0x40, flags=<register not in topmost frame>) [/home/marcus/projects/wine/dlls/winex11.drv/event.c:296] in winex11 (0x0034e464) 4 0x605d4129 MsgWaitForMultipleObjectsEx+0xf9(count=0x2, pHandles=0x34e6e0, timeout=0x7530, mask=0x40, flags=0x0) [/home/marcus/projects/wine/dlls/user32/message.c:3195] in user32 (0x0034e614) 5 0x605d418f MsgWaitForMultipleObjects+0x3f(count=0x2, handles=0x34e6e0, wait_all=0x0, timeout=0x7530, mask=0x40) [/home/marcus/projects/wine/dlls/user32/message.c:3209] in user32 (0x0034e634) 6 0x605dad36 WaitForInputIdle+0xe6(hProcess=0x118, dwTimeOut=0x7530) [/home/marcus/projects/wine/dlls/user32/message.c:3236] in user32 (0x0034e6f4) 7 0x7b86e257 wait_input_idle+0x57() in kernel32 (0x0034e714) 8 0x7b872fde WinExec+0x13e() in kernel32 (0x0034e7b4) 9 0x013299e3 in fullsoft (+0x99e3) (0x0034eae0) 10 0x01321ca1 in fullsoft (+0x1ca1) (0x0034eb48) 11 0x01321b33 in fullsoft (+0x1b33) (0x0034eb78) 12 0x01321b7a in fullsoft (+0x1b7a) (0x0034eb88) 13 0x7b8409d3 UnhandledExceptionFilter+0x53() in kernel32 (0x0034eeb8) 14 0x00b18b5e in msvcr80 (+0x8b5e) (0x0034eed8) 15 0x008c82a6 in sunbird (+0x4c82a6) (0x0034ff08) 16 0x7b870fee start_process+0xee() in kernel32 (0x0034ffe8) 17 0x6001e697 wine_switch_to_stack+0x17() in libwine.so.1 (0x00000000) Wine-dbg>x /i $eip 0x60dddad7 XCheckIfEvent+0x17 in libx11.so.6: movl 0x4d0(%edx),%eax Wine-dbg>print $edx 55 Wine-dbg> display is obviously bad. Likely a different problem, but not good anyway. Ciao, Marcus

Nov. 29, 2006

Mozilla Sunbird install - Mystery bug #6613

by Mike McCormack

Anybody have time to help triage an install problem? Sunbird (and FireFox 2) installs run in an endless loop for some people. If you have time to try the install out and see if it works on your system, then comment on bug 6613, it would help out. http://bugs.winehq.org/show_bug.cgi?id=6613 I'm interested in which distro you're using, which filesystem, free disk space, locale, timezone, CPU (64bit or 32 bit?), etc. Mike

Nov. 29, 2006

Re: d3d [1]: Remove AddRef from IWineD3DDevice_GetVertexDeclaration.

by Markus Amsler

Stefan Dösinger wrote: > I proposed getting rid of COM some time ago, but I didn't consider that we > need some features like inheritance. > > I think we should keep COM, but get rid of the D3D-Specific Addrefs because > they are different between the various versions d3d versions. So keep the > AddRef in IUnknown methods(QueryInterface namely), and GetParent, but do not > AddRef in d3d methods(SetTexture / GetTexture, ...) > > Opinions? From my point of view (solving d3dx implicit surface refcounts) the GetParent AddRef is ugly. It spreads d3dx refcounting across d3dx and wined3d. It would be easier to only do d3dx refcounting in d3dx. But that's only my pragmatic view, without further COM knowledge.

Nov. 28, 2006

Re: ws2_32: Implement SO_GET_MAX_MSG_SIZE for getsockopt.

by Kai Blin

On Tuesday 28 November 2006 17:05, Dmitry Timoshkov wrote: > "Kai Blin" <kai.blin(a)gmail.com> wrote: > > +static void test_extendedSocketOptions() > > +{ > > + WSADATA wsa; > > + SOCKET sock; > > + struct sockaddr_in sa; > > + int sa_len = sizeof(struct sockaddr_in); > > + ULONG optval, optlen = sizeof(ULONG); > > Why to use ULONG if what you later cast it to is 'int *', and according to > the docs (and your implementation) SO_MAX_MSG_SIZE expects sizeof(int) for > the optval? Because another part of the docs says SO_MAX_MSG_SIZE expects ULONG. > > ... > > > + if(bind(sock, (struct sockaddr *) &sa, (int)sa_len) < 0){ > > There is not need for a cast, sa_len has the type 'int'. Whoops, copy&paste error from my linux test program, where sa_len was a size_t. > > + trace("Failed to bind socket: 0x%08x\n", WSAGetLastError()); > > + closesocket(sock); > > + WSACleanup(); > > + exit(1); > > This kills the whole process instead of terminating gracefully just this > particular test. Yeah, forgot to convert that from the standalone test I wrote first. Reading all the code in socket.c must have turned my brain all mushy. Thanks for catching those. I'll fix it and resubmit. Kai -- Kai Blin, <kai Dot blin At gmail Dot com> WorldForge developer http://www.worldforge.org/ Wine developer http://wiki.winehq.org/KaiBlin/ -- Will code for cotton.

Nov. 28, 2006

Re: d3d [1]: Remove AddRef from IWineD3DDevice_GetVertexDeclaration.

by Stefan Dösinger

Am Dienstag 28 November 2006 16:55 schrieb Alexandre Julliard: > Markus Amsler <markus.amsler(a)oribi.org> writes: > > I wasn't fully aware of the importance of the COM refcount rules, just > > saw the opportunity for simplification. For removing COM it would be > > probably better to start with removing the lpVtbl's. I agree it's > > probably best to revert them for consistency. Should I send patches? > > If we are not going further with the COM removal, then yes if you > could send patches to revert the changes that would be good. I proposed getting rid of COM some time ago, but I didn't consider that we need some features like inheritance. I think we should keep COM, but get rid of the D3D-Specific Addrefs because they are different between the various versions d3d versions. So keep the AddRef in IUnknown methods(QueryInterface namely), and GetParent, but do not AddRef in d3d methods(SetTexture / GetTexture, ...) Opinions?

Nov. 28, 2006

Re: ws2_32: Implement SO_GET_MAX_MSG_SIZE for getsockopt.

by Dmitry Timoshkov

"Kai Blin" <kai.blin(a)gmail.com> wrote: > +static void test_extendedSocketOptions() > +{ > + WSADATA wsa; > + SOCKET sock; > + struct sockaddr_in sa; > + int sa_len = sizeof(struct sockaddr_in); > + ULONG optval, optlen = sizeof(ULONG); Why to use ULONG if what you later cast it to is 'int *', and according to the docs (and your implementation) SO_MAX_MSG_SIZE expects sizeof(int) for the optval? ... > + if(bind(sock, (struct sockaddr *) &sa, (int)sa_len) < 0){ There is not need for a cast, sa_len has the type 'int'. > + trace("Failed to bind socket: 0x%08x\n", WSAGetLastError()); > + closesocket(sock); > + WSACleanup(); > + exit(1); This kills the whole process instead of terminating gracefully just this particular test. > + } > + > + ret = getsockopt(sock, SOL_SOCKET, SO_MAX_MSG_SIZE, (char *)&optval, (int *)&optlen); -- Dmitry.

Nov. 28, 2006