May 2013 - Wine-devel - list.winehq.org

Re: [PATCH 4/5] wined3d: Explicitly pass the state to shader_load_constants
by Henri Verbeet 13 May '13

13 May '13

On 12 May 2013 15:30, Stefan Dösinger <stefan(a)codeweavers.com> wrote: > @@ -806,7 +806,8 @@ struct wined3d_shader_backend_ops > void (*shader_deselect_depth_blt)(void *shader_priv, const struct wined3d_gl_info *gl_info); > void (*shader_update_float_vertex_constants)(struct wined3d_device *device, UINT start, UINT count); > void (*shader_update_float_pixel_constants)(struct wined3d_device *device, UINT start, UINT count); > - void (*shader_load_constants)(const struct wined3d_context *context, BOOL usePS, BOOL useVS); > + void (*shader_load_constants)(const struct wined3d_context *context, BOOL usePS, BOOL useVS, > + const struct wined3d_state *state, void *shader_priv); > void (*shader_load_np2fixup_constants)(void *shader_priv, const struct wined3d_gl_info *gl_info, > const struct wined3d_state *state); > void (*shader_destroy)(struct wined3d_shader *shader); > I think we'll want shader_priv to be the first argument, for consistency with the other calls. I don't think usePS and useVS are needed anymore.

1 0

Re: [PATCH 3/3] iphlpapi: Implement CancelIPChangeNotify.
by Hans Leidekker 13 May '13

13 May '13

On Mon, 2013-05-06 at 19:23 -0600, Erich E. Hoover wrote: > --- a/dlls/iphlpapi/iphlpapi_main.c > +++ b/dlls/iphlpapi/iphlpapi_main.c > @@ -210,8 +210,14 @@ DWORD WINAPI AllocateAndGetIpAddrTableFromStack(PMIB_IPADDRTABLE *ppIpAddrTable, > */ > BOOL WINAPI CancelIPChangeNotify(LPOVERLAPPED overlapped) > { > - FIXME("(overlapped %p): stub\n", overlapped); > - return FALSE; > + HANDLE handle; > + > + TRACE("(overlapped %p)\n", overlapped); > + > + if (!overlapped) > + return FALSE; > + handle = (HANDLE) overlapped->u.Pointer; > + return TerminateThread(handle, STATUS_CANCELLED); > } TerminateThread is not the right tool here. If the thread is blocked on the recv call while being terminated it will leak the netlink socket. You'd need a way to synchronize the threads, but an overhead of one thread per caller is rather costly in the first place. This suggests that the server should poll the socket.

3 9

Re: [2/2] server: Remove a not necessary check for 'current' being not NULL.
by Alexandre Julliard 13 May '13

13 May '13

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > 'current' is already dereferenced twice before this point. > --- > server/request.c | 23 ++++++++++------------- > 1 file changed, 10 insertions(+), 13 deletions(-) You can't do that. -- Alexandre Julliard julliard(a)winehq.org

2 3

Re: [website] Simplified Chinese translation for release 1.5.30
by Qian Hong 13 May '13

13 May '13

2013/5/11 Jactry Zeng <jactry92(a)gmail.com>: > + <li>WinInet 对代理旁路的支持。</li> This translation seems suspect, but I don't have a very good alternative yet. In 'Internet Options' setting of MSIE7, 'Bypass proxy server for local address' is translated to '跳过本地地址的代理服务器' Thanks for the work. -- Regards, Qian Hong - http://www.winehq.org

2 2

IPC between linux processes and wine processes
by Alexandre Bique 12 May '13

12 May '13

Hi, I plan to write a Linux VST bridge to Windows VST. This could improve windows VST support in our native DAW. I wonder what is the best strategy and what do you recommend? - spawn a wine vst host process from the linux vst and then do IPC through unix socket - spawn a wine vst host process from the linux vst and then do IPC through tcp socket - spawn a wine vst host process from the linux vst and then do IPC through shared memory - load the windows vst into the Linux process and directly call its functions (with appropriate calling convention). Thanks a lot! Regards, -- Alexandre Bique

3 4

Re : Full software of Wine version 1.4
by Nick Kucan 11 May '13

11 May '13

Dear Sir/madam, My name is Nick Kucan and I have ubuntu 11.04 at home - In the last three month I had a major issue with my the operating system (which I am currently investigating the source error) - Each time the system crashes (fails to load and screen says "input signal out of range" and does not load Ubuntu 11.04). I have tried Ubuntu 12.10 but unfortunatly there is a major issue with my modem - Every time i tried to connect - less than a second later - it cuts it right off. I have tried get help with the ubuntu website and provider - no such luck at this point of time. There is no such issue with version 11.04. I am forced to re-install and restore from backup - the wine program that I backed up wont even load so I am forced to download at least 170mb (and at least 1 hour to download from your headquaters website which I am tired of downloading the last 30 times - that cost me a lot of money to from my pre-paid account. I wish to download that .deb package from another source and burn into a dvd for my personal use only instead of downloading every time from your website each time and restore everything from its point. Can provide me a good link that I can directly download that file and I need to burn only one copy for my home pc so it can make my life a little easy - this copy of windows wine will NOT be distributed to anyone else except for my own stand alone pc in an emergency so that I can restore it back the way it is please. I have backed up the firefox and adobe flashplayer on the backup dvd - it works fine - the only issue is that the wine program full file cant get it working so its forced me to go back to the software source and download that painfully long software again!. Can u help me please.... Nick Kucan.

2 1

Re: [1/2] winex11: Cache XRandR 1.0 display mode. (try 2)
by Ruslan Kabatsayev 10 May '13

10 May '13

Hi, Won't such code fail if the mode is changed outside wine? Regards, Ruslan On Fri, May 10, 2013 at 8:07 PM, Sam Edwards <cfsworks(a)gmail.com> wrote: > > > >

2 1

Re: [1/3] winex11: Select for XRandR screen change events and implement an event handler.
by Alexandre Julliard 10 May '13

10 May '13

Sam Edwards <cfsworks(a)gmail.com> writes: > This patch adds an event handler to xrandr.c to process > RRScreenChangeNotify events, so that when non-Wine applications change > the display resolution, Wine can find out about it. > > Two main discussion points: > > - Previously, there was nothing pulling events out of gdi_display's > queue, thus necessitating the need for a new event handler. I opted to > put it in xrandr.c since this handler only manages XRandR events, but > it might be cleaner to move it over to event.c. > > - Are we okay with having to call the event handler, or would this > work better in a continuously-running background thread instead? It's okay only because you are not actually calling it ;-) But no, you can't select for events on the gdi display; actually I don't think you want to do it this way at all. Handling external resizes should be done in the desktop process, and most likely involves the wineserver too. It certainly won't be an easy task. -- Alexandre Julliard julliard(a)winehq.org

3 4

Re: [2/3] kernel32: Terminate threads before unloading DLLs on process exit.
by Alexandre Julliard 10 May '13

10 May '13

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > If it's not yet time to make this change please let me know what else needs > to be tested/fixed/investigated. The heap behavior would need testing, since freeing memory is probably the most common thing done in a detach routine. -- Alexandre Julliard julliard(a)winehq.org

2 3

Re: [PATCH] d3dcompiler: Add a LUT to find compilation targets info. (try 3)
by Alexandre Julliard 10 May '13

10 May '13

Christian Costa <titan.costa(a)gmail.com> writes: > - if (strlen(target) != 6 || target[1] != 's' || target[2] != '_' || target[4] != '_') > + for (i = 0; i < nb_targets; i++) > { > - FIXME("Unknown compilation target %s.\n", debugstr_a(target)); > - return D3DERR_INVALIDCALL; > + if (!strcmp(target, targets_info[i].name) || > + (targets_info[i].altname && !strcmp(target, targets_info[i].altname))) A binary search would be much better. -- Alexandre Julliard julliard(a)winehq.org

2 1