December 2005 - Wine-Devel - list.winehq.org

Re: Latest winetest (20051212) issue on WinXP

by Joris Huizer

Paul Vriens wrote: > Hi, > > I've just noticed that during the running of winetest.exe windows > creates a nice desktop stating that 'Active Desktop' is stopped because > an unexpected error occurred. > > Anyone? > > Paul. > Send them a bugreport ;)

Dec. 13, 2005

Re: WINED3D / D3D9: SetVertexShader & SetPixelShader should not change the shader's refcount

by H. Verbeet

> The best thing to do is add an internal reference instead of referencing the parent. e.g. > if (pShader != NULL) { > IWineD3DVertexShader_AddRef(pShader); > } > if (oldShader != NULL) { > IWineD3DVertexShader_Release(oldShader); > } Is that reference really needed?

Dec. 13, 2005

Re: WINED3D / D3D9: SetVertexShader & SetPixelShader should not change the shader's refcount

by Oliver Stieber

--- "H. Verbeet" <hverbeet(a)gmail.com> wrote: > SetVertexShader & SetPixelShader should not change the shader's > refcount. Verified on win2k, DirectX 9.0c. > > Changelog: > - Don't change the shader's refcount in SetVertexShader & SetPixelShader > > 26c6ae144caaaa6f7e40f7a7c6e3059851819ecf > diff --git a/dlls/wined3d/device.c b/dlls/wined3d/device.c > index c3a7f9c..a9c253f 100644 > --- a/dlls/wined3d/device.c > +++ b/dlls/wined3d/device.c > @@ -3761,7 +3761,6 @@ HRESULT WINAPI IWineD3DDeviceImpl_GetVer > > HRESULT WINAPI IWineD3DDeviceImpl_SetVertexShader(IWineD3DDevice *iface, IWineD3DVertexShader* > pShader) { > IWineD3DDeviceImpl *This = (IWineD3DDeviceImpl *)iface; > - IWineD3DVertexShader *oldShader = This->updateStateBlock->vertexShader; > > This->updateStateBlock->vertexShader = pShader; > This->updateStateBlock->changed.vertexShader = TRUE; > @@ -3772,20 +3771,7 @@ HRESULT WINAPI IWineD3DDeviceImpl_SetVer > return D3D_OK; > } > > - if (pShader != NULL) { > - IUnknown *newVertexShaderParent; > - /* GetParent will add a ref, so leave it hanging until the vertex buffer is cleared */ > - TRACE("(%p) : setting pShader(%p)\n", This, pShader); > - IWineD3DVertexShader_GetParent(pShader, &newVertexShaderParent); > - } else { > - TRACE("Clear down the shader\n"); > - } > - if (oldShader != NULL) { > - IUnknown *oldVertexShaderParent; > - IWineD3DVertexShader_GetParent(oldShader, &oldVertexShaderParent); > - IUnknown_Release(oldVertexShaderParent); > - IUnknown_Release(oldVertexShaderParent); > - } > + TRACE("(%p) : setting pShader(%p)\n", This, pShader); The best thing to do is add an internal reference instead of referencing the parent. e.g. if (pShader != NULL) { IWineD3DVertexShader_AddRef(pShader); } if (oldShader != NULL) { IWineD3DVertexShader_Release(oldShader); } ___________________________________________________________ NEW Yahoo! Cars - sell your car and browse thousands of new and used cars online! http://uk.cars.yahoo.com/

Dec. 13, 2005

Re: WINED3D / D3D9: SetVertexShader & SetPixelShader should not change the shader's refcount

by Oliver Stieber

--- "H. Verbeet" <hverbeet(a)gmail.com> wrote: > SetVertexShader & SetPixelShader should not change the shader's > refcount. Verified on win2k, DirectX 9.0c. > > Changelog: > - Don't change the shader's refcount in SetVertexShader & SetPixelShader > > 26c6ae144caaaa6f7e40f7a7c6e3059851819ecf > diff --git a/dlls/wined3d/device.c b/dlls/wined3d/device.c > index c3a7f9c..a9c253f 100644 > --- a/dlls/wined3d/device.c > +++ b/dlls/wined3d/device.c > @@ -3761,7 +3761,6 @@ HRESULT WINAPI IWineD3DDeviceImpl_GetVer > > HRESULT WINAPI IWineD3DDeviceImpl_SetVertexShader(IWineD3DDevice *iface, IWineD3DVertexShader* > pShader) { > IWineD3DDeviceImpl *This = (IWineD3DDeviceImpl *)iface; > - IWineD3DVertexShader *oldShader = This->updateStateBlock->vertexShader; > > This->updateStateBlock->vertexShader = pShader; > This->updateStateBlock->changed.vertexShader = TRUE; > @@ -3772,20 +3771,7 @@ HRESULT WINAPI IWineD3DDeviceImpl_SetVer > return D3D_OK; > } > > - if (pShader != NULL) { > - IUnknown *newVertexShaderParent; > - /* GetParent will add a ref, so leave it hanging until the vertex buffer is cleared */ > - TRACE("(%p) : setting pShader(%p)\n", This, pShader); > - IWineD3DVertexShader_GetParent(pShader, &newVertexShaderParent); > - } else { > - TRACE("Clear down the shader\n"); > - } > - if (oldShader != NULL) { > - IUnknown *oldVertexShaderParent; > - IWineD3DVertexShader_GetParent(oldShader, &oldVertexShaderParent); > - IUnknown_Release(oldVertexShaderParent); > - IUnknown_Release(oldVertexShaderParent); > - } > + TRACE("(%p) : setting pShader(%p)\n", This, pShader); The best thing to do is add an internal reference e.g. if (pShader != NULL) { IWineD3DVertexShader_AddRef(pShader) IWineD3DVertexShader_Release(oldShader) ___________________________________________________________ NEW Yahoo! Cars - sell your car and browse thousands of new and used cars online! http://uk.cars.yahoo.com/

Dec. 13, 2005

Re: gethostbyname(my_name) and IL2-Sturmovik

by David Nolden

@Peter I'm attaching the patch I've written again, so everyone can find it. It tries to find out the public IP by walkind the Interfaces available through the Kernel. @Pavel I think that such a patch should be applied, Peter just said what I wanted to hear :-) Windows just acts different in this, and wine should act exactly the same. You can't expect every Linux-user to specially hand-write his /etc/hosts so that his windows-programs/games work. :) greetings Am Dienstag, 13. Dezember 2005 16:57 schrieb Peter Åstrand: > On Tue, 13 Dec 2005, Pavel Troller wrote: > >> 127.0.0.1 localhost.localdomain localhost Zwirch > > > > Yes, I know that *MANY* well-known distros are wrong with /etc/hosts and > > similar files.. > > > > I think that it is NOT GOOD to patch wine to fix this case; it behaves > > absolutely correctly and all the problem should be solved by correctly > > configuring your Linux networking. Please see any good document about > > naming principles of network interfaces. > > I haven't followed the whole thread, so sorry if I've missed something > obvious, but: > > I think you are wrong. POSIX does not say that it's wrong to have the > computer name associated with 127.0.0.1 in /etc/hosts, and it doesn't say > that gethostbyname(my_name) should return a "real" IP. If I remember > correctly, though, Microsoft, specifies that gethostbyname(my_name) > *should* return real IP. > > This subtle difference means that Wine must take extra care. Fetching a > real IP on a UNIX host is not trivial, but doable. > > Instead of just fetching the IP by looking up the hostname, I suggest > using the approach below instead. It will fail if there is no route to > 192.168.1.1 (typically this means that you don't have a default gateway). > When this happens, you can resolv the hostname as a fallback. > > #include <stdio.h> > #include <stdlib.h> > #include <sys/socket.h> > #include <netinet/in.h> > #include <arpa/inet.h> > > int main() > { > int s, c, got_sockname = 0; > struct sockaddr_in serv_addr; > struct sockaddr_in my_addr; > char *my_ip; > > serv_addr.sin_family = AF_INET; > serv_addr.sin_port = htons(1); > /* Even better: Use the server you will connect to */ > inet_aton("192.168.1.1", &serv_addr.sin_addr); > > s = socket (PF_INET, SOCK_DGRAM, 0); > if (s < 0) { > perror("socket"); > exit(EXIT_FAILURE); > } > > /* connect fails if if server is unreachable, for example */ > if ((c = connect(s, (struct sockaddr*) &serv_addr, sizeof(serv_addr))) > >= 0) { socklen_t my_addrlen; > my_addrlen = sizeof(my_addr); > if (getsockname(s, (struct sockaddr*)&my_addr, &my_addrlen) >= 0) > { got_sockname = 1; > } > } > > if (!got_sockname) > /* Use 127.0.0.1 as a fallback. Even better: Resolv the > hostname. */ > inet_aton("127.0.0.1", &my_addr.sin_addr); > > my_ip = inet_ntoa(my_addr.sin_addr); > printf("%s\n", my_ip); > > return 0; > } > > Regards,

Dec. 13, 2005

Latest winetest (20051212) issue on WinXP

by Paul Vriens

Hi, I've just noticed that during the running of winetest.exe windows creates a nice desktop stating that 'Active Desktop' is stopped because an unexpected error occurred. Anyone? Paul.

Dec. 13, 2005

Re: Can Wine run "sv.net classic"? Not at the moment :-(

by Dan Kegel

On 12/13/05, Marcus Meissner <marcus(a)jet.franken.de> wrote: > MDAC also requires WGA, so we might need to reimplement it anyway first. MDAC-2.7sp1 is plenty for (most?) VB6 apps, and we definitely want Wine to be able to run them sooner rather than later. So I think this deserves fixing, and there's no need to delay. - Dan -- Wine for Windows ISVs: http://kegel.com/wine/isv

Dec. 13, 2005

Re: gethostbyname(my_name) and IL2-Sturmovik

by Peter Åstrand

On Tue, 13 Dec 2005, Pavel Troller wrote: >> 127.0.0.1 localhost.localdomain localhost Zwirch > Yes, I know that *MANY* well-known distros are wrong with /etc/hosts and > similar files.. > I think that it is NOT GOOD to patch wine to fix this case; it behaves > absolutely correctly and all the problem should be solved by correctly > configuring your Linux networking. Please see any good document about > naming principles of network interfaces. I haven't followed the whole thread, so sorry if I've missed something obvious, but: I think you are wrong. POSIX does not say that it's wrong to have the computer name associated with 127.0.0.1 in /etc/hosts, and it doesn't say that gethostbyname(my_name) should return a "real" IP. If I remember correctly, though, Microsoft, specifies that gethostbyname(my_name) *should* return real IP. This subtle difference means that Wine must take extra care. Fetching a real IP on a UNIX host is not trivial, but doable. Instead of just fetching the IP by looking up the hostname, I suggest using the approach below instead. It will fail if there is no route to 192.168.1.1 (typically this means that you don't have a default gateway). When this happens, you can resolv the hostname as a fallback. #include <stdio.h> #include <stdlib.h> #include <sys/socket.h> #include <netinet/in.h> #include <arpa/inet.h> int main() { int s, c, got_sockname = 0; struct sockaddr_in serv_addr; struct sockaddr_in my_addr; char *my_ip; serv_addr.sin_family = AF_INET; serv_addr.sin_port = htons(1); /* Even better: Use the server you will connect to */ inet_aton("192.168.1.1", &serv_addr.sin_addr); s = socket (PF_INET, SOCK_DGRAM, 0); if (s < 0) { perror("socket"); exit(EXIT_FAILURE); } /* connect fails if if server is unreachable, for example */ if ((c = connect(s, (struct sockaddr*) &serv_addr, sizeof(serv_addr))) >= 0) { socklen_t my_addrlen; my_addrlen = sizeof(my_addr); if (getsockname(s, (struct sockaddr*)&my_addr, &my_addrlen) >= 0) { got_sockname = 1; } } if (!got_sockname) /* Use 127.0.0.1 as a fallback. Even better: Resolv the hostname. */ inet_aton("127.0.0.1", &my_addr.sin_addr); my_ip = inet_ntoa(my_addr.sin_addr); printf("%s\n", my_ip); return 0; } Regards, -- Peter Åstrand Chief Developer Cendio www.thinlinc.com Teknikringen 3 www.cendio.se 583 30 Linköping Phone: +46-13-21 46 00

Dec. 13, 2005

Re: gethostbyname(my_name) and IL2-Sturmovik

by Pavel Troller

> On 12/13/05, David Nolden <david.nolden.wine(a)art-master.de> wrote: > > @Pavel Troller > > I've checkt the /etc/hosts now. You were right, it says > > > > 127.0.0.1 localhost.localdomain localhost Zwirch > > > > (With Zwirch the Name of this Computer) > > > > But I can't just statically rewrite it, because I use DHCP. Also this is a > > fresh Ubuntu-Install, so I'm surely not the only > > Wine-User with such an /etc/hosts.. > > > > This link may be of interest: > http://lists.debian.org/debian-devel/2005/10/msg00387.html > > Tom > Hi! Yes, it uncovers even more problems with the /etc/hosts file on many distributions (hey, I'm sooooo happy that I'm maintaining my own private distro, independently of all the "big players" on the market :-) ). However, your link talks mainly about localhost.localdomain. I am even more strict and I'm saying that the host name also should not be there. And because I know that the host name should be assigned to EXACTLY ONE network interface, it is a reason that in my distro, when there are no public if's active, a dummy0 comes up and the host name is assigned to it. This if holds an address DIFFERENT from localhost (192.168.0.1 is used as a preconfigured default on a fresh install). Once more, please don't patch wine, it's evidently not its problem at all! With regards, Pavel Troller

Dec. 13, 2005

Re: gethostbyname(my_name) and IL2-Sturmovik

by Tom Wickline

On 12/13/05, David Nolden <david.nolden.wine(a)art-master.de> wrote: > @Pavel Troller > I've checkt the /etc/hosts now. You were right, it says > > 127.0.0.1 localhost.localdomain localhost Zwirch > > (With Zwirch the Name of this Computer) > > But I can't just statically rewrite it, because I use DHCP. Also this is a > fresh Ubuntu-Install, so I'm surely not the only > Wine-User with such an /etc/hosts.. > This link may be of interest: http://lists.debian.org/debian-devel/2005/10/msg00387.html Tom

Dec. 13, 2005