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

Re: wininet: Make resolving hostnames thread safe.
by Rob Shearman May 24, 2008

May 24, 2008

2008/5/19 Hans Leidekker <hans(a)it.vu.nl>: > @@ -151,7 +158,21 @@ BOOL GetAddress(LPCWSTR lpszServerName, INTERNET_PORT nServerPort, > name = HeapAlloc(GetProcessHeap(), 0, sz+1); > WideCharToMultiByte( CP_UNIXCP, 0, lpszServerName, len, name, sz, NULL, NULL ); > name[sz] = 0; > + > +#ifdef HAVE_LINUX_GETHOSTBYNAME_R_6 > + phe = NULL; > + buf = HeapAlloc(GetProcessHeap(), 0, bufsize); > + while (buf) > + { > + int res = gethostbyname_r(name, &hostentry, buf, bufsize, &phe, &locerr); > + if (res != ERANGE) break; > + bufsize *= 2; > + buf = HeapReAlloc(GetProcessHeap(), 0, buf, bufsize); > + } > + HeapFree(GetProcessHeap(), 0, buf); > +#else > phe = gethostbyname(name); > +#endif In the long term it would be better to use getaddrinfo instead of gethostbyname/gethostbyname_r, but that's too big a change for 1.0. Rob

2 1

Re: ddraw: [3/3] Fix for regression bug #13277
by Stefan Dösinger May 24, 2008

May 24, 2008

> + ref = getRefcount(lpDD); > + ok(ref == 1, "Got refcount %ld, expected 2\n", ref); > + > + IDirectDraw_Release(lpDD); > + > + ref = getRefcount(lpDD); > + ok(ref == 0, "Got refcount %ld, expected 1\n", ref); I think there's a copypaste error, it seems you forgot to adjust the error string. Your test shows the correctness of the patch pretty well, although out of curiosity, can you try this: Between the IDirectDraw_Release(), which reduces the refcount to 0 and the ref = getRefcount(), can you try to HeapAlloc() small portions(e.g. 512 bytes) of memory with HEAP_ZERO_MEMORY until you get an out of memory error. This should make it less likely that the lpDD object survives by luck. It might be destroyed by windows, but the memory is not yet overwritten by something else, so the getRefcount works by luck. If this does not lead to a crash, try the memory allocation + getRefcount again after the Surface::Release call below, there it should cause a crash. > @@ -1998,6 +2026,13 @@ IDirectDrawImpl_CreateNewSurface(IDirectDrawImpl *This, > InterlockedIncrement(&This->surfaces); > list_add_head(&This->surface_list, &(*ppSurf)->surface_list_entry); > > + IDirectDrawInternal_AddRef(This); > @@ -424,6 +424,9 @@ IDirectDrawSurfaceImpl_Release(IDirectDrawSurface7 *iface) > /* Reduce the ddraw refcount */ > if(ifaceToRelease) IUnknown_Release(ifaceToRelease); > LeaveCriticalSection(&ddraw_cs); > + > + /* Surface does not need ddraw anymore, release it's reference */ > + IDirectDrawInternal_Release(ddraw); The counterpart to IDirectDrawImpl_CreateNewSurface is IDirectDrawSurfaceImpl_Destroy, not _Release. CreateNewSurface is called for each complex sublevel surface(e.g. front+backbuffer if they are created via DDSC_COMPLEX), while IDirectDrawSurface::Release must be called only for the root surface(ie, the pointer returned from IDirectDraw::CreateSurface). That means you have to move the IDirectDrawInternal_Release into IDirectDrawSurfaceImpl_Destroy, otherwise there's a reference leak.

2 1

Re: rpcrt4: Skip client tests if the server isn't listening
by Rob Shearman May 24, 2008

May 24, 2008

2008/5/23 Maarten Lankhorst <m.b.lankhorst(a)gmail.com>: > --- > > *Don't apply, breaks tests* > > Rob: Is there anything wrong with this patch apart from the fact it > stops the tests from running in wine due to a failure? Would it > prevent windows failures? > > Ideally I would make our RpcMgmtIsServerListening return RPC_S_OK, but > I don't know the code so I'm afraid to make changes like this. Please give me links to test data that shows this happening and I'll try to figure out why the client isn't able to talk to the client. Rob

1 0

RFC: comctl32 indentations
by Andrew Talbot May 24, 2008

May 24, 2008

Hi, In toolbar.c:TOOLBAR_Destroy(), should the "if" statement at line 5439 be compound to match the indentation, or should the three invocations of TOOLBAR_DeleteImageList() be outdented? treeview.c: In TREEVIEW_DeleteItem(), how conditional is the call of TREEVIEW_SetFirstVisible() at line 1539? I'm guessing that the simple "if" statement at line 3045 just requires indenting(?) If the forum is the wrong place to raise this sort of query, please forgive and advise. :) Thanks, -- Andy.

2 2

Can I do this in WINE?
by Bret Comstock Waldow May 24, 2008

May 24, 2008

I have a Tablet PC, running Kubuntu, and I have just released a project on sourceforge to allow me to use the Microsoft hand writing recognition. The current approach uses a .NET server in IIS running on Tablet XP to provide the recognition. I would prefer to write a classic Windows program that accepts the polyline ink strokes, accesses the Tablet SDK & .dlls to recognize ink, and run that program in WINE directly in Linux without IIS and network connections. I'm thinking of my legal dual-boot installation, and reaching over with WINE to run the app just as I run Notepad from the Windows install now. I'm a Java programmer by day, and I haven't programmed Windows COM before (much). I'd like comments and advice (even help) on doing this. All of this message is written by hand in Linux using the SHIP project (except copying and pasting this URL): https://sourceforge.net/projects/ship-project/ Thanks in advance for information, bret

5 14

New winetricks 20080523: directx9, bugfixes
by Dan Kegel May 23, 2008

May 23, 2008

Another week, another winetricks. Changes since May 19's wine-devel annoucement: r52: - rotate sourceforge mirrors to avoid getting stuck on bad one - cope with users who rename c:/windows/Fonts to c:/windows/fonts - more conservative advice on sha1sum mismatches - handle bad exit status from /usr/bin/which on mac - cope with redhat.com sending decompressed files - use win2K when installing directx9 so it installs more stuff r50: - Added directx9 verb Online as always at http://kegel.com/wine/winetricks or http://winezeug.googlecode.com

1 0

Proposed indentation fixes
by Andrew Talbot May 23, 2008

May 23, 2008

Hi All, I'm intending to "correct" some indentation anomalies, and I propose to do this in an "indiscriminate" way. In other words, I would set the indentation to match what the code does now, without any interpretation of what may have been intended. This means that something like if (a) b; c; would become 1) if (a) b; c; rather than 2) if (a) { b; c; } The advantages are that the code is "honest" and the existing behaviour is preserved. The disadvantages are that if (2) were intended, it won't be achieved here, and the corrected version will no longer "suggest" what was really intended. Is this "automatic" method still the way to go? Thanks, -- Andy.

2 2

unknown device issues after running winetest
by Paul Vriens May 23, 2008

May 23, 2008

Hi, On several windows boxes (VMware) I receive this nice message that Windows will try to find a driver for an unknown device. This device is added by the setupapi/devinst test and is not (or can't be?) properly removed after the tests. On NT/W2K I'm able to delete all the relevant registry keys (regedt32.exe) but on Vista I don't seem to be able to delete them Any idea? Do others have the same issue(s) ? -- Cheers, Paul.

5 30

Re: ddraw: Fixed in part regression bug #13277, retry
by Stefan Dösinger May 23, 2008

May 23, 2008

Please send test and patch in one patch file, or send them in different emails if a combined patch is too big(which surely isn't the case here)

1 0

rpcrt4 test failures
by Maarten Lankhorst May 23, 2008

May 23, 2008

Hi Rob, I roughly found the reason of the rpcrt4:server test crashing for me. It crashes when it's trying to connect to localhost through the rpc 'tcp' transport. The crash is likely caused because it can't connect to the locally hosted test server the tcp way, and decides to bomb out instead. However it won't say it failed to connect, instead it crashes on the first test-case that uses rpc. I would rather not have the test crash, that would also obscure the possibility of running the other tests in that file. However I don't know why it fails to connect over tcp (netbios works fine). I disabled my firewall entirely, so it shouldn't be the cause of this failure. Could there be other restrictions in place? Cheers, Maarten

1 0