August 2009 - Wine-Devel - list.winehq.org

Re: winelib include problem
by Maarten Lankhorst Aug. 19, 2009

Aug. 19, 2009

Hi Jesse, Jesse Allen schreef: > Maartan, > > In a patch committed on 10 Jul 2008, includes: Add video mixing > renderer header, it included a construct that autogenerates this piece > of code for strmif.h > > " > typedef struct tagVMRGUID { > GUID *pGUID; > GUID GUID; > } VMRGUID; > " > > Unfortunately any C++ file that includes that header seems to miscompile. > > " > #include <strmif.h> > " > > produces with wineg++ > > In file included from blarg.c:1: > /home/jesse/build/wine/include/strmif.h:8972: error: declaration of > 'GUID tagVMRGUID::GUID' > /home/jesse/build/wine/include/guiddef.h:31: error: changes meaning of > 'GUID' from 'typedef struct _GUID GUID' > winegcc: g++ failed > > > This is also a problem in the DXSDK when compiling with g++, I > suppose. This is explained here: > http://home.clara.net/raoulgough/vidmodem/dshow.html > > I tried patching the vmrender.idl similar to this: > " > typedef struct tagVMRGUID > { > #if defined(__cplusplus) > GUID *pGUID, ::GUID; > #else > GUID *pGUID, GUID; > #endif > } VMRGUID; > " > Judging by that link you sent me, shouldn't it be: typedef struct tagVMRGUID { #if defined(__cplusplus) ::GUID *pGUID, GUID; #else GUID *pGUID, GUID; #endif } VMRGUID; Or am i reading that link wrong? > But no go. It still produces the same problem. Obviously you can't > have the ::GUID prefixed and work on gcc as much as you can have the > GUID GUID in g++. But I think we will need it, otherwise I will just > have to hack the strmif.h everytime. > > Can you take a look at the problem? I think we will want this to work > for winelib on g++ as it's pretty well dependent. > Main problem is of course the msvc compiler doesn't have that problem, but if that ifdef works perhaps it's possible to have it like that. Cheers, Maarten.

2 3

Re: [PATCH 5/5] wined3d: Hide WINED3DFMT_R16G16B16A16_UNORM again.
by Stefan Dösinger Aug. 19, 2009

Aug. 19, 2009

Am Wednesday 19 August 2009 10:55:39 schrieb Henri Verbeet: > This format is broken on some cards. Hide it until we figure out a reliable > way to deal with it. Didn't you consider breaking something the best way to get the driver fixed? :-) Fwiw, I reported this to linux-bugs(a)nvidia.com, together with a testcase. No reply yet. Feel free to disable it if there are too many issues though, but maybe add a comment why it is disabled.

1 0

Re: (resend, 2nd) shell32: fix program search in ShellExecuteEx, quote program name if it contains spaces [2nd]
by Alexandre Julliard Aug. 19, 2009

Aug. 19, 2009

Stefan Leichter <Stefan.Leichter(a)camline.com> writes: > @@ -1711,28 +1713,29 @@ > } > else > { > - /* If the executable name is not quoted, we have to use this search loop here, > + /* FIXME: what versions support this? Fails on 2000/XP > + If the executable name is not quoted, we have to use this search loop here, > that in CreateProcess() is not sufficient because it does not handle shell links. */ > WCHAR buffer[MAX_PATH], xlpFile[MAX_PATH]; > - LPWSTR space, s; > + LPWSTR space; > > - LPWSTR beg = wszApplicationName/*sei_tmp.lpFile*/; > - for(s=beg; (space=strchrW(s, ' ')); s=space+1) { > - int idx = space-sei_tmp.lpFile; > - memcpy(buffer, sei_tmp.lpFile, idx * sizeof(WCHAR)); > + lstrcpynW(buffer, wszApplicationName, sizeof(buffer)/sizeof(WCHAR)); > + while((space=strchrW(buffer, ' '))) > + { > + int idx = space-buffer; > buffer[idx] = '\0'; This will break the loop since you don't reset the buffer contents. Also the indentation is messed up, please fix your editor setup. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [1/18] d3dx9: Add shader assembler private definitions
by Vincent Povirk Aug. 18, 2009

Aug. 18, 2009

Don't send patches that only add unused code. On Sun, Aug 16, 2009 at 12:50 PM, Matteo Bruni<matteo.mystral(a)gmail.com> wrote: > This is a long patch series that implements the D3D shader assembler > and D3DXAssembleShader function. The first part has already passed a > couple of rounds of review in wine-devel. > > > > -- Vincent Povirk

4 6

Re: [PATCH 1/3] tools/widl: A structure containing a BOGUS_ARRAY should be a BOGUS_STRUCT.
by Alexandre Julliard Aug. 18, 2009

Aug. 18, 2009

Huw Davies <huw(a)codeweavers.com> writes: > @@ -280,9 +282,10 @@ unsigned char get_struct_fc(const type_t *type) > if (get_enum_fc(t) == RPC_FC_ENUM16) > return RPC_FC_BOGUS_STRUCT; > break; > - case TGT_POINTER: > case TGT_ARRAY: > - if (get_pointer_fc(t, field->attrs, FALSE) == RPC_FC_RP || pointer_size != 4) > + if(get_array_fc(t) == RPC_FC_BOGUS_ARRAY) return RPC_FC_BOGUS_STRUCT; > + case TGT_POINTER: > + if(get_pointer_fc(t, field->attrs, FALSE) == RPC_FC_RP || pointer_size != 4) This doesn't seem to match the midl behavior. For instance in the server.idl test puints_t becomes a bogus struct, but it's a simple struct with midl. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: advapi32: SetEntriesInAclW() should accept account name "CURRENT_USER".
by Rob Shearman Aug. 18, 2009

Aug. 18, 2009

2009/8/16 Rein Klazes <wijn(a)online.nl>: > @@ -3510,7 +3511,8 @@ DWORD WINAPI SetEntriesInAclW( ULONG count, PEXPLICIT_ACCESSW pEntries, > DWORD sid_size = FIELD_OFFSET(SID, SubAuthority[SID_MAX_SUB_AUTHORITIES]); > DWORD domain_size = MAX_COMPUTERNAME_LENGTH + 1; > SID_NAME_USE use; > - if (!LookupAccountNameW(NULL, pEntries[i].Trustee.ptstrName, ppsid[i], &sid_size, NULL, &domain_size, &use)) > + if ( strcmpW( pEntries[i].Trustee.ptstrName, CURRENT_USER ) && What will the memory pointed to by ppsid[i] be set to in this case? Hint: whatever the memory was last used for, not the SID of the current user which is what is desired. > + !LookupAccountNameW(NULL, pEntries[i].Trustee.ptstrName, ppsid[i], &sid_size, NULL, &domain_size, &use)) > { > WARN("bad user name %s for trustee %d\n", debugstr_w(pEntries[i].Trustee.ptstrName), i); > ret = ERROR_INVALID_PARAMETER; -- Rob Shearman

1 0

Re: msxml3: Support loading windows file path under wine (try 2 - Resend)
by Alexandre Julliard Aug. 18, 2009

Aug. 18, 2009

Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> writes: > Hi, > Stopped converting the filename to Unicode. Obviously you still need to convert for opening it. -- Alexandre Julliard julliard(a)winehq.org

1 0

mp3 update
by Aric Stewart Aug. 18, 2009

Aug. 18, 2009

Hello all, I am working on updating our winemp3.acm implementation to a modern libmpg123. We where at about 0.59r and I have 1.8.1 working very well. I have a few questions. Right now I have the smallest subset of files from libmpg123 that are needed to compile and work in the wine framework (mostly) unmodified. I have replaced all the malloc/free calls with HeapAlloc/HeapFree and worked in all the wine debugging code instead of the libmpg123 debugging code. What I am wondering about is that there is still a lot of code in these files we have no interest in. The file reader and http reader and other things that we do not use (it is accessed just by the direct stream methods) and those things dependencies. Should I leave all this code present in our version of the modules and mostly preserve the libmpg123 work to make it easier to keep updating forward? Or should i remove all the unneeded code so that the modules present in winehq are leaner and more targeted to just the things the wine needs, but then they deviate much more from the original libmpg123 source? With what I have now the mp3 source material this game i am testing is using sounds much much better. So it should be a great improvement. Thanks, -aric

11 27

add new application attrib.exe
by EA Durbin Aug. 18, 2009

Aug. 18, 2009

This fixes bug 18059. If this one is not acceptable, feel free to take it and clean it up, I am abandoning this effort from here forward. _________________________________________________________________ Windows Live: Keep your friends up to date with what you do online. http://windowslive.com/Campaign/SocialNetworking?ocid=PID23285::T:WLMTAGL:O…

1 0

Re: [7/10] windowscodecs: implement CopyPixels for GIF decoder (resend 2)
by Vincent Povirk Aug. 17, 2009

Aug. 17, 2009

Wait, no, this one's wrong too. Don't commit it. It's correct for odd heights but not even heights. To calculate the number of rows divisible by N, I have to divide the 0-based index of the maximum row by N and add 1 (for the ever-present row 0). However, I divided the height, which is maximum row+1. To do this correctly, I would have to do (height-1)/N+1 or the equivalent (height+(N-1))/N. I'll fix it tomorow. On Mon, Aug 17, 2009 at 3:49 PM, Vincent Povirk<vincent(a)codeweavers.com> wrote: > This time with correct interlacing code in the patch that is attached. > > > > -- Vincent Povirk

1 0