April 2014 - Wine-Devel - list.winehq.org

Re: [PATCH 2/2] user32: add tests for valid cbSize in GetMonitorInfoA

by Alexandre Julliard

Patrick Rudolph <siro(a)das-labor.org> writes: > @@ -362,6 +363,21 @@ static void test_monitors(void) > monitor = pMonitorFromRect( &rc, MONITOR_DEFAULTTONULL ); > } > > + /* sanity check */ > + monitor = pMonitorFromRect( &rc, MONITOR_DEFAULTTONEAREST ); > + mi.cbSize = sizeof(mi); > + ret = pGetMonitorInfoA( monitor, &mi ); > + ok( ret, "GetMonitorInfo failed\n" ); > + > + miex.cbSize = sizeof(miex); > + ret = pGetMonitorInfoA( monitor, (LPMONITORINFO)&miex ); > + ok( ret, "GetMonitorInfo failed\n" ); > + > + /* specify invalid cbSize */ > + mi.cbSize = 0xdeadbeef; > + ret = pGetMonitorInfoA( monitor, &mi ); > + ok( !ret, "GetMonitorInfo succeded\n" ); That's not a very convincing test. -- Alexandre Julliard julliard(a)winehq.org

April 18, 2014

Re: [PATCH 1/6] d3d9: Retrieve the volume description from the container.

by Henri Verbeet

On 17 April 2014 15:47, Stefan Dösinger <stefan(a)codeweavers.com> wrote: > --- > dlls/d3d9/texture.c | 20 +++++++++++++------- > 1 file changed, 13 insertions(+), 7 deletions(-) > I suppose this works, but I don't think it's really necessary. At some point wined3d_texture_get_sub_resource() will go away of course, but I don't think there's a reason we can't e.g. have a wined3d_resource_get_sub_resource_desc(), or alternatively just store the information on d3d9 volume creation.

April 18, 2014

Re: [PATCH 2/6] wined3d: Pass the volume texture level to the client libraries.

by Henri Verbeet

On 17 April 2014 15:47, Stefan Dösinger <stefan(a)codeweavers.com> wrote: > The goal is to remove volumes and surfaces from the wined3d public API. > This extra information is needed for the transition. We got rid of this not too long ago, adding it back doesn't sound very attractive. If you really need a way to find out a volume's level, I'd prefer wined3d_volume_get_level() instead, mostly because it doesn't need changes in every D3D library. I'm not convinced this is really necessary though, or that starting in d3d9 as opposed to wined3d with this is necessarily the right approach.

April 18, 2014

Re: shell32: Use CREATE_NEW_CONSOLE when SEE_MASK_NOCONSOLE is omitted.

by Alistair Leslie-Hughes

Hi Vincent, On 18/04/2014 8:30 AM, Vincent Povirk wrote: >>> There's no way around this. >> > If I fix that bug and use SEE_MASK_NO_CONSOLE when calling > ShellExecuteEx from the tests, I still get some console windows. If I > also hack ShellExecuteA to use SEE_MASK_NO_CONSOLE, I don't get any > console windows, but my testing on Windows has shown that to be > incorrect. So, there are a few things I can do to mitigate this, and I > will send patches for them, but I can't prevent all consoles from > being created. > That's what I saw as well but didn't have time to completely test my solution. Alistair.

April 17, 2014

Re: ole32/tests: Fix crashes in PropVariantClear tests under win2000 and XP.

by Marvin

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at https://testbot.winehq.org/JobDetails.pl?Key=6384 Your paranoid android. === w2008s64 (64 bit propvariant) === propvariant: unhandled exception c0000005 at 000007FEFF6093CE

April 17, 2014

Re: shell32: Use CREATE_NEW_CONSOLE when SEE_MASK_NOCONSOLE is omitted.

by Vincent Povirk

>> There's no way around this. > > Spoke too soon. I'll get back to you after further investigation. There's another Wine bug where invoking a .lnk file with SEE_MASK_NO_CONSOLE will create a console. That doesn't happen on Windows. I'll send a patch for that. If I fix that bug and use SEE_MASK_NO_CONSOLE when calling ShellExecuteEx from the tests, I still get some console windows. If I also hack ShellExecuteA to use SEE_MASK_NO_CONSOLE, I don't get any console windows, but my testing on Windows has shown that to be incorrect. So, there are a few things I can do to mitigate this, and I will send patches for them, but I can't prevent all consoles from being created.

April 17, 2014

Re: shell32: Use CREATE_NEW_CONSOLE when SEE_MASK_NOCONSOLE is omitted.

by Vincent Povirk

> There's no way around this. Spoke too soon. I'll get back to you after further investigation.

April 17, 2014

Re: shell32: Use CREATE_NEW_CONSOLE when SEE_MASK_NOCONSOLE is omitted.

by Vincent Povirk

> I submitted a patch to do this but was reject since it causes the test to > behave exactly like window. :( > > Look for "Correct create flags when SEE_MASK_NO_CONSOLE isn't specified" in > January 2014 mailing lists. There's no way around this. We can have the tests call ShellExecuteEx with SEE_MASK_NO_CONSOLE, and that helps somewhat, but ShellExecute behaves as if SEE_MASK_NO_CONSOLE is omitted.

April 17, 2014

Re: shell32: Use CREATE_NEW_CONSOLE when SEE_MASK_NOCONSOLE is omitted.

by Vincent Povirk

> I submitted a patch to do this but was reject since it causes the test to > behave exactly like window. :( Which test? And why is this a problem?

April 17, 2014

Re: [PATCH 6/6] d3d9: Decouple d3d9 volumes from wined3d volumes.

by Stefan Dösinger

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Fyi, doing the same for surfaces will take a while because a number of surface-related functions have to be moved out of the way first. If having a difference between volumes and surfaces is an issue skip this patch for now. Am 2014-04-17 15:47, schrieb Stefan Dösinger: > Releasing the partially created texture in case of an out of memory > error is a problem, see volumetexture_init. > > My choice to handle this would be to move the memory allocation into the > init function (and rename it), but a few years ago Henri has changed it > into the opposite direction everywhere. I'm open to other suggestions. > (And no, I don't like the idea of calculating the mipmap levels before > calling wined3d_texture_create.) > --- > dlls/d3d9/d3d9_private.h | 10 +++-- > dlls/d3d9/device.c | 15 +------ > dlls/d3d9/texture.c | 108 ++++++++++++++++++++++++++++++++--------------- > dlls/d3d9/volume.c | 30 ++----------- > 4 files changed, 85 insertions(+), 78 deletions(-) > > diff --git a/dlls/d3d9/d3d9_private.h b/dlls/d3d9/d3d9_private.h > index 389cc95..2f59a2f 100644 > --- a/dlls/d3d9/d3d9_private.h > +++ b/dlls/d3d9/d3d9_private.h > @@ -184,14 +184,13 @@ struct d3d9_volume > { > IDirect3DVolume9 IDirect3DVolume9_iface; > struct d3d9_resource resource; > - struct wined3d_volume *wined3d_volume; > IUnknown *container; > IUnknown *forwardReference; > UINT level; > }; > > -void volume_init(struct d3d9_volume *volume, struct wined3d_volume *wined3d_volume, > - UINT level, const struct wined3d_parent_ops **parent_ops) DECLSPEC_HIDDEN; > +void d3d9_volume_init(struct d3d9_volume *volume, UINT level) DECLSPEC_HIDDEN; > +void d3d9_volume_destroy(void *sub_resource) DECLSPEC_HIDDEN; > > struct d3d9_swapchain > { > @@ -251,6 +250,11 @@ struct d3d9_texture > struct d3d9_resource resource; > struct wined3d_texture *wined3d_texture; > IDirect3DDevice9Ex *parent_device; > + > + void **sub_resource; > + void (*sub_resource_destroy)(void *sub_resource); > + UINT level_count, sub_resource_count; > + BOOL error_cleanup; > }; > > HRESULT cubetexture_init(struct d3d9_texture *texture, struct d3d9_device *device, > diff --git a/dlls/d3d9/device.c b/dlls/d3d9/device.c > index 84184ec..11e13c5 100644 > --- a/dlls/d3d9/device.c > +++ b/dlls/d3d9/device.c > @@ -3448,22 +3448,11 @@ static HRESULT CDECL device_parent_volume_created(struct wined3d_device_parent * > void *container_parent, struct wined3d_volume *volume, UINT level, void **parent, > const struct wined3d_parent_ops **parent_ops) > { > - struct d3d9_volume *d3d_volume; > - > TRACE("device_parent %p, container_parent %p, volume %p, parent %p, parent_ops %p.\n", > device_parent, container_parent, volume, parent, parent_ops); > > - if (!(d3d_volume = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, sizeof(*d3d_volume)))) > - return E_OUTOFMEMORY; > - > - volume_init(d3d_volume, volume, level, parent_ops); > - *parent = d3d_volume; > - TRACE("Created volume %p.\n", d3d_volume); > - > - d3d_volume->container = container_parent; > - > - IDirect3DVolume9_Release(&d3d_volume->IDirect3DVolume9_iface); > - d3d_volume->forwardReference = container_parent; > + *parent_ops = &d3d9_null_wined3d_parent_ops; > + *parent = container_parent; > > return D3D_OK; > } > diff --git a/dlls/d3d9/texture.c b/dlls/d3d9/texture.c > index 7e8dd68..03c75fa 100644 > --- a/dlls/d3d9/texture.c > +++ b/dlls/d3d9/texture.c > @@ -1018,15 +1018,9 @@ static DWORD WINAPI d3d9_texture_3d_GetLOD(IDirect3DVolumeTexture9 *iface) > static DWORD WINAPI d3d9_texture_3d_GetLevelCount(IDirect3DVolumeTexture9 *iface) > { > struct d3d9_texture *texture = impl_from_IDirect3DVolumeTexture9(iface); > - DWORD ret; > > TRACE("iface %p.\n", iface); > - > - wined3d_mutex_lock(); > - ret = wined3d_texture_get_level_count(texture->wined3d_texture); > - wined3d_mutex_unlock(); > - > - return ret; > + return texture->level_count; > } > > static HRESULT WINAPI d3d9_texture_3d_SetAutoGenFilterType(IDirect3DVolumeTexture9 *iface, > @@ -1073,55 +1067,48 @@ static void WINAPI d3d9_texture_3d_GenerateMipSubLevels(IDirect3DVolumeTexture9 > static HRESULT WINAPI d3d9_texture_3d_GetLevelDesc(IDirect3DVolumeTexture9 *iface, UINT level, D3DVOLUME_DESC *desc) > { > struct d3d9_texture *texture = impl_from_IDirect3DVolumeTexture9(iface); > - HRESULT hr = D3D_OK; > - DWORD level_count; > + struct wined3d_resource *resource = wined3d_texture_get_resource(texture->wined3d_texture); > + struct wined3d_resource_desc wined3d_desc; > > TRACE("iface %p, level %u, desc %p.\n", iface, level, desc); > > - wined3d_mutex_lock(); > - level_count = wined3d_texture_get_level_count(texture->wined3d_texture); > - if (level >= level_count) > + if (level >= texture->level_count) > { > WARN("Invalid mipmap level specified, returning D3DERR_INVALIDCALL.\n"); > - hr = D3DERR_INVALIDCALL; > + return D3DERR_INVALIDCALL; > } > - else > - { > - struct wined3d_resource *resource = wined3d_texture_get_resource(texture->wined3d_texture); > - struct wined3d_resource_desc wined3d_desc; > > - wined3d_resource_get_desc(resource, &wined3d_desc); > - desc->Format = d3dformat_from_wined3dformat(wined3d_desc.format); > - desc->Type = D3DRTYPE_VOLUME; > - desc->Usage = wined3d_desc.usage & WINED3DUSAGE_MASK; > - desc->Pool = wined3d_desc.pool; > - desc->Width = max(wined3d_desc.width >> level, 1); > - desc->Height = max(wined3d_desc.height >> level, 1); > - desc->Depth = max(wined3d_desc.depth >> level, 1); > - hr = D3D_OK; > - } > + wined3d_mutex_lock(); > + wined3d_resource_get_desc(resource, &wined3d_desc); > + desc->Format = d3dformat_from_wined3dformat(wined3d_desc.format); > + desc->Type = D3DRTYPE_VOLUME; > + desc->Usage = wined3d_desc.usage & WINED3DUSAGE_MASK; > + desc->Pool = wined3d_desc.pool; > + desc->Width = max(wined3d_desc.width >> level, 1); > + desc->Height = max(wined3d_desc.height >> level, 1); > + desc->Depth = max(wined3d_desc.depth >> level, 1); > wined3d_mutex_unlock(); > > - return hr; > + return D3D_OK; > } > > static HRESULT WINAPI d3d9_texture_3d_GetVolumeLevel(IDirect3DVolumeTexture9 *iface, > UINT level, IDirect3DVolume9 **volume) > { > struct d3d9_texture *texture = impl_from_IDirect3DVolumeTexture9(iface); > - struct wined3d_resource *sub_resource; > struct d3d9_volume *volume_impl; > > TRACE("iface %p, level %u, volume %p.\n", iface, level, volume); > > wined3d_mutex_lock(); > - if (!(sub_resource = wined3d_texture_get_sub_resource(texture->wined3d_texture, level))) > + if (level >= texture->level_count) > { > + WARN("Invalid mipmap level specified, returning D3DERR_INVALIDCALL.\n"); > wined3d_mutex_unlock(); > return D3DERR_INVALIDCALL; > } > > - volume_impl = wined3d_resource_get_parent(sub_resource); > + volume_impl = texture->sub_resource[level]; > *volume = &volume_impl->IDirect3DVolume9_iface; > IDirect3DVolume9_AddRef(*volume); > wined3d_mutex_unlock(); > @@ -1234,9 +1221,18 @@ struct d3d9_texture *unsafe_impl_from_IDirect3DBaseTexture9(IDirect3DBaseTexture > > static void STDMETHODCALLTYPE d3d9_texture_wined3d_object_destroyed(void *parent) > { > + UINT i; > struct d3d9_texture *texture = parent; > d3d9_resource_cleanup(&texture->resource); > - HeapFree(GetProcessHeap(), 0, texture); > + > + if (texture->sub_resource) > + { > + for (i = 0; i < texture->sub_resource_count && texture->sub_resource[i]; i++) > + texture->sub_resource_destroy(texture->sub_resource[i]); > + HeapFree(GetProcessHeap(), 0, texture->sub_resource); > + } > + if (!texture->error_cleanup) > + HeapFree(GetProcessHeap(), 0, texture); > } > > static const struct wined3d_parent_ops d3d9_texture_wined3d_parent_ops = > @@ -1331,6 +1327,7 @@ HRESULT volumetexture_init(struct d3d9_texture *texture, struct d3d9_device *dev > { > struct wined3d_resource_desc desc; > HRESULT hr; > + DWORD i; > > texture->IDirect3DBaseTexture9_iface.lpVtbl = (const IDirect3DBaseTexture9Vtbl *)&d3d9_texture_3d_vtbl; > d3d9_resource_init(&texture->resource); > @@ -1350,15 +1347,56 @@ HRESULT volumetexture_init(struct d3d9_texture *texture, struct d3d9_device *dev > wined3d_mutex_lock(); > hr = wined3d_texture_create(device->wined3d_device, &desc, levels, 0, > texture, &d3d9_texture_wined3d_parent_ops, &texture->wined3d_texture); > - wined3d_mutex_unlock(); > if (FAILED(hr)) > { > WARN("Failed to create wined3d volume texture, hr %#x.\n", hr); > - return hr; > + goto error; > + } > + > + texture->level_count = wined3d_texture_get_level_count(texture->wined3d_texture); > + texture->sub_resource_count = texture->level_count; > + > + texture->sub_resource_destroy = d3d9_volume_destroy; > + texture->sub_resource = HeapAlloc(GetProcessHeap(), 0, > + sizeof(*texture->sub_resource) * texture->sub_resource_count); > + if (!texture->sub_resource) > + { > + hr = E_OUTOFMEMORY; > + goto error; > + } > + > + for (i = 0; i < texture->level_count; i++) > + { > + struct d3d9_volume *d3d9_volume; > + > + if (!(d3d9_volume = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, sizeof(*d3d9_volume)))) > + { > + hr = E_OUTOFMEMORY; > + goto error; > + } > + d3d9_volume_init(d3d9_volume, i); > + TRACE("Created volume %p.\n", d3d9_volume); > + > + d3d9_volume->container = (IUnknown *)&texture->IDirect3DBaseTexture9_iface; > + IDirect3DVolume9_Release(&d3d9_volume->IDirect3DVolume9_iface); > + d3d9_volume->forwardReference = d3d9_volume->container; > + texture->sub_resource[i] = d3d9_volume; > } > > texture->parent_device = &device->IDirect3DDevice9Ex_iface; > IDirect3DDevice9Ex_AddRef(texture->parent_device); > - > + wined3d_mutex_unlock(); > return D3D_OK; > + > +error: > + /* Once the wined3d texture is created there is no way to destroy it without running > + * d3d9_texture_wined3d_object_destroyed. The caller expects volumetexture_init not > + * to free texture in case of an error. */ > + if (texture->wined3d_texture) > + { > + texture->error_cleanup = TRUE; > + wined3d_texture_decref(texture->wined3d_texture); > + } > + wined3d_mutex_unlock(); > + return hr; > } > diff --git a/dlls/d3d9/volume.c b/dlls/d3d9/volume.c > index 07f15ca..916a949 100644 > --- a/dlls/d3d9/volume.c > +++ b/dlls/d3d9/volume.c > @@ -63,13 +63,6 @@ static ULONG WINAPI d3d9_volume_AddRef(IDirect3DVolume9 *iface) > refcount = InterlockedIncrement(&volume->resource.refcount); > TRACE("%p increasing refcount to %u.\n", iface, refcount); > > - if (refcount == 1) > - { > - wined3d_mutex_lock(); > - wined3d_volume_incref(volume->wined3d_volume); > - wined3d_mutex_unlock(); > - } > - > return refcount; > } > > @@ -89,13 +82,6 @@ static ULONG WINAPI d3d9_volume_Release(IDirect3DVolume9 *iface) > refcount = InterlockedDecrement(&volume->resource.refcount); > TRACE("%p decreasing refcount to %u.\n", iface, refcount); > > - if (!refcount) > - { > - wined3d_mutex_lock(); > - wined3d_volume_decref(volume->wined3d_volume); > - wined3d_mutex_unlock(); > - } > - > return refcount; > } > > @@ -213,26 +199,16 @@ static const struct IDirect3DVolume9Vtbl d3d9_volume_vtbl = > d3d9_volume_UnlockBox, > }; > > -static void STDMETHODCALLTYPE volume_wined3d_object_destroyed(void *parent) > +void d3d9_volume_destroy(void *sub_resource) > { > - struct d3d9_volume *volume = parent; > + struct d3d9_volume *volume = sub_resource; > d3d9_resource_cleanup(&volume->resource); > HeapFree(GetProcessHeap(), 0, volume); > } > > -static const struct wined3d_parent_ops d3d9_volume_wined3d_parent_ops = > -{ > - volume_wined3d_object_destroyed, > -}; > - > -void volume_init(struct d3d9_volume *volume, struct wined3d_volume *wined3d_volume, > - UINT level, const struct wined3d_parent_ops **parent_ops) > +void d3d9_volume_init(struct d3d9_volume *volume, UINT level) > { > volume->IDirect3DVolume9_iface.lpVtbl = &d3d9_volume_vtbl; > d3d9_resource_init(&volume->resource); > - wined3d_volume_incref(wined3d_volume); > - volume->wined3d_volume = wined3d_volume; > volume->level = level; > - > - *parent_ops = &d3d9_volume_wined3d_parent_ops; > } > -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/ iQIcBAEBAgAGBQJTT92RAAoJEN0/YqbEcdMwAe0P/2Nt/ZATqANNkVaY6o6BdknG qfw6wSQemhF3OvKX84aKSqdEdjOkHNqtQb9wHE1jEoGufkSO9VLINIOJRqa8M962 8cwg2fceNVlU148pwvdCYbkbvJofN0zqDT91IRTBfkKsoAo4Fcwo9fSzKp28gQQv KLBvcJksQ0DZNZAxHrjbGpu16xVinfTypVIa8KLrar4Ns+XPoYDikghKh7s+kSb6 Mvo5ZwmNTLYi6dLFblzjRV8ftLriN8gI6r4BxYoAotVTuN70FRG4ce+qs97Nwwe9 nAEfcSp752N5DQ2qLWiBZdaa/ZjX3y389hkaHB2uYz8RZH91hEecMgIV+6SzF4el SU5apNTF214x2SBKvGwL0jLRQ0J1QoEVbw2siwwyIS02id3yp19yekiEdOsfmq+4 CMaYdZSHHOBntcFNwz/TMli1y46ugTHWVPimVz1WGqNfWdxN1xyJX5YdIyRpIPC8 lzuzsobX9DyULFHAtrDz3RrBCMsztTFbag7bIAZ3ZbLLxEEMm0926eEQVBsGY1fZ uS5V6AKA8JLQ+OZfNCoZeiMASR70PfICoBpahld1rNo5NfjsB56P+h+kWWkP8l6q OOeXJgYYn3n40LndHl0EPk6EjXUqyUkZ9s1sPxcdgL6GD1iE9oRPHBmm2OC19+tF rswC/IHRuFD+4KHlR5OR =dY9X -----END PGP SIGNATURE-----

April 17, 2014

Re: riched20: Avoid releasing a non-existent interface.

by Jactry Zeng

2014-04-17 16:20 GMT+08:00 Nikolay Sivov <bunglehead(a)gmail.com>: > > It's possible that new instance will be returned after selection changed. This way you'll end up with two alive instances. > No. According these test: https://github.com/Jactry/wine/commit/4e0c5e8f2f169560ef8d8d7ae0d0620b8ddcd… GetSelection will not return a new instance after selection changed. > Another thing to test if ITextDocument responds to IID_ITextSelection query. > >> I will try some other ways to fix this bug. > First thing to do is to figure out what relations are between these interfaces. You are right, I need having more tests. -- Regards, Jactry Zeng

April 17, 2014

Invalid address for set breakpoint in winedbg

by Frédéric Delanoy

Hi. I'm setting a breakpoint in a wine test.so with winedbg, "br function_name" finds it, but when I hit (c)ontinue, I get an "Invalid address (0x... function_name) for breakpoint 1, disabling it". Any reason for this? FWIW, function_name is a non-static function I added in the testfile.c. Frédéric NB: winedbg in gdb mode doesn't give an error message, but otherwise exhibits the same behaviour

April 17, 2014

Re: [PATCH] atl: Added a typelib

by Jacek Caban

Hi Nikolay, On 04/17/14 11:22, Nikolay Sivov wrote: > +[ > + local, > + object, > + oleautomation, > + uuid(425b5af0-65f1-11d1-9611-0000f81e0d0d), > +] > +interface IDocHostUIHandlerDispatch : IDispatch Those interfaces should go to atliface.idl. Cheers, Jacek

April 17, 2014

Re: riched20: Avoid releasing a non-existent interface.

by Nikolay Sivov

On 4/17/2014 11:59, Jactry Zeng wrote: > > 2014-04-17 11:54 GMT+08:00 Nikolay Sivov <bunglehead(a)gmail.com > <mailto:bunglehead(a)gmail.com>>: > > > > What you're doing is a violation of refcount handling. The rule is > to release what you got, without relying on internals like that. In > this case GetSelection() returns interface pointer and you're > responsible in exactly one Release() on it. > > > > Interesting thing to test would be to check if GetSelection() > returns new instance every time it's called. If this is a case it will > justify some code changes to support this, right now patch is wrong. > > > > If it actually returns same interface pointer you can't protect from > use-after-free because I can grab multiple references with several > GetSelection() calls, and when I'll try to release them it will be > freed already by a loop like that. > > Yes, you are right. GetSelection() will not return a new pointer every > time it's called, I have checked it by some tests on Windows. Did you actually change selection range in between of these calls? It's possible that new instance will be returned after selection changed. This way you'll end up with two alive instances. Another thing to test if ITextDocument responds to IID_ITextSelection query. > I will try some other ways to fix this bug. First thing to do is to figure out what relations are between these interfaces. > > Thanks again! > > -- > Regards, > Jactry Zeng >

April 17, 2014

Re: shell32: Use CREATE_NEW_CONSOLE when SEE_MASK_NOCONSOLE is omitted.

by Alistair Leslie-Hughes

Hi Vincent, On 17/04/2014 7:11 AM, Vincent Povirk wrote: > MSDN specifically mentions that this flag is the opposite of > CREATE_NEW_CONSOLE: > http://msdn.microsoft.com/en-us/library/windows/desktop/bb759784%28v=vs.85%… > > I also wrote a test program (built with -mconsole), which uses > ShellExecuteEx to run cmd.exe. When running it from a command line on > windows, it creates a new console window for cmd.exe only when > SEE_MASK_NOCONSOLE is not in fMask. (If I compile with -mwindows, it > seems to create a new console window regardless of the value of the > flag or how I start it, which looks like another wine bug to me.) > I submitted a patch to do this but was reject since it causes the test to behave exactly like window. :( Look for "Correct create flags when SEE_MASK_NO_CONSOLE isn't specified" in January 2014 mailing lists. Best Regards Alistair Leslie-Hughes

April 17, 2014

Re: riched20: Avoid releasing a non-existent interface.

by Jactry Zeng

2014-04-17 11:54 GMT+08:00 Nikolay Sivov <bunglehead(a)gmail.com>: > > What you're doing is a violation of refcount handling. The rule is to release what you got, without relying on internals like that. In this case GetSelection() returns interface pointer and you're responsible in exactly one Release() on it. > > Interesting thing to test would be to check if GetSelection() returns new instance every time it's called. If this is a case it will justify some code changes to support this, right now patch is wrong. > > If it actually returns same interface pointer you can't protect from use-after-free because I can grab multiple references with several GetSelection() calls, and when I'll try to release them it will be freed already by a loop like that. Yes, you are right. GetSelection() will not return a new pointer every time it's called, I have checked it by some tests on Windows. I will try some other ways to fix this bug. Thanks again! -- Regards, Jactry Zeng

April 17, 2014

Re: riched20: Avoid releasing a non-existent interface.

by Nikolay Sivov

On 4/16/2014 17:37, Jactry Zeng wrote: > Hi Nikolay, > > Thanks for your review! > > The crash can be reproduced follow this: > - first release the ITextSelection or IOleClientSite > interfaces completely; > - release ITextDocument interface; > - try to release the IRichEditOle (crash happen) > > And this patch try to fix it. > > + /* hres = ITextDocument_GetSelection(txtDoc, &txtSel); */ > + /* ok(hres == S_OK, "ITextDocument_GetSelection\n"); */ > + /* while(ITextSelection_Release(txtSel)); */ What you're doing is a violation of refcount handling. The rule is to release what you got, without relying on internals like that. In this case GetSelection() returns interface pointer and you're responsible in exactly one Release() on it. Interesting thing to test would be to check if GetSelection() returns new instance every time it's called. If this is a case it will justify some code changes to support this, right now patch is wrong. If it actually returns same interface pointer you can't protect from use-after-free because I can grab multiple references with several GetSelection() calls, and when I'll try to release them it will be freed already by a loop like that.

April 16, 2014

Re: [4/5] user32: Our popup test window loses capture to itself.

by Marvin

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at https://testbot.winehq.org/JobDetails.pl?Key=6374 Your paranoid android. === w2000pro (32 bit win) === Failure running script in VM: network read timed out (wait2:ListSize:0/4) === w8 (32 bit win) === win.c:1159: Test failed: first call to SetShellWindow(hwnd1) win.c:1161: Test failed: wrong shell window: 000300F4 win.c:1198: Test failed: SetShellWindow(hwnd4) win.c:1200: Test failed: wrong shell window: 000300F4 - expected hwnd4 win.c:1203: Test failed: wrong next window for hwnd4: 00020080 - expected 0 win.c:1214: Test failed: wrong shell window: 000300F4 - expected hwnd4 win.c:1224: Test failed: wrong next window for hwnd4 after SetWindowPos(): 00020080 - expected 0

April 16, 2014

Re: [3/5] user32: Add a message test for popup menus on windows that have capture.

by Marvin

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at https://testbot.winehq.org/JobDetails.pl?Key=6373 Your paranoid android. === w2000pro (32 bit msg) === Failure running script in VM: network read timed out (wait2:ListSize:0/4)

April 16, 2014

Re: [2/5] user32: Add test calling EndMenu from WM_ENTERMENULOOP.

by Marvin

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at https://testbot.winehq.org/JobDetails.pl?Key=6372 Your paranoid android. === w2000pro (32 bit msg) === Failure running script in VM: network read timed out (wait2:ListSize:0/4)

April 16, 2014

Re: msvcp90/msvcp90.h: right-hand operand of comma expression has no effect

by Paul Chitescu

On Wed Apr 16, 2014 00:48:00 Gerald Pfeifer wrote: > Building git head with GCC development sources (soon to be GCC 4.9.0) > gives me the following warnings > > ../msvcp90/msvcp90.h:529:75: warning: right-hand operand of comma > expression has no effect [-Wunused-value] ../msvcp90/msvcp90.h:530:77: > warning: right-hand operand of comma expression has no effect > [-Wunused-value] > > several times. > > Digging into this in more detail I see > > ../msvcp90/ios.c: In function 'basic_streambuf_char__Xsgetn_s': > ../msvcp90/msvcp90.h:529:75: warning: right-hand operand of comma > expression has no effect [-Wunused-value] > #define memcpy_s( dst, size, src, count ) (memcpy( (dst), (src), (count) > ), 0) ^ ../msvcp90/ios.c:1218:13: note: in expansion of macro 'memcpy_s' > memcpy_s(ptr+copied, size, *this->prpos, chunk); > ^ > > Someone must have written that code intentionally, so I am wondering > how to best address this. > > Gerald memcpy_s() returns an errno_t so the author tried to force returning zero (no error). A simple replacement (still without any checks memcpy_s does) would be: #define memcpy_s( dst, size, src, count ) (!memcpy( (dst), (src), (count) )) Paul

April 16, 2014

Re: riched20: Avoid releasing a non-existent interface.

by Jactry Zeng

Hi Nikolay, Thanks for your review! The crash can be reproduced follow this: - first release the ITextSelection or IOleClientSite interfaces completely; - release ITextDocument interface; - try to release the IRichEditOle (crash happen) And this patch try to fix it. (tests in attachment can reproduce the crash.) 2014-04-16 18:59 GMT+08:00 Nikolay Sivov <bunglehead(a)gmail.com>: > - This->txtSel->reOle = NULL; >> - ITextSelection_Release(&This->txtSel->ITextSelection_iface); >> - IOleClientSite_Release(&This->clientSite->IOleClientSite_iface); >> + if(This->txtSel) >> + { >> + This->txtSel->reOle = NULL; >> + ITextSelection_Release(&This->txtSel->ITextSelection_iface); >> + } >> + if(This->clientSite) >> + { >> + This->clientSite->reOle = NULL; >> + IOleClientSite_Release(&This->clientSite->IOleClientSite_ >> iface); >> + } >> heap_fr >> > This can't happen. > > IOleClientSiteImpl *This = impl_from_IOleClientSite(iface); >> ULONG ref = InterlockedDecrement(&This->ref); >> if (ref == 0) >> + { >> + if(This->reOle) >> + This->reOle->clientSite = NULL; >> heap_free(This); >> + } >> return ref; >> > Why do you need this? > -- Regards, Jactry Zeng

April 16, 2014

Re: [PATCH 3/3] msvcr90: Share the source code with msvcrt (try2)

by Alexandre Julliard

Piotr Caban <piotr(a)codeweavers.com> writes: > --- > dlls/msvcr90/Makefile.in | 32 +- > dlls/msvcr90/msvcr90.c | 199 ---- > dlls/msvcr90/msvcr90.spec | 2418 > ++++++++++++++++++++++----------------------- > dlls/msvcrt/Makefile.in | 2 +- > dlls/msvcrt/main.c | 4 + > dlls/msvcrt/msvcrt.h | 1 + > 6 files changed, 1245 insertions(+), 1411 deletions(-) > delete mode 100644 dlls/msvcr90/msvcr90.c It's still broken: In file included from ../../../wine/dlls/msvcr90/../msvcrt/cpp.c:34:0: ../../../wine/dlls/msvcr90/../msvcrt/cxx.h:23:0: warning: "__thiscall" redefined [enabled by default] <built-in>:0:0: note: this is the location of the previous definition ../../../wine/dlls/msvcr90/../msvcrt/dir.c:1026:32: error: conflicting types for ‘_splitpath_s’ /usr/lib/gcc/i686-w64-mingw32/4.6/../../../../i686-w64-mingw32/include/sec_api/stdlib_s.h:48:70: note: previous declaration of ‘_splitpath_s’ was here ../../../wine/dlls/msvcr90/../msvcrt/dir.c:1126:32: error: conflicting types for ‘_wsplitpath_s’ /usr/lib/gcc/i686-w64-mingw32/4.6/../../../../i686-w64-mingw32/include/sec_api/stdlib_s.h:57:70: note: previous declaration of ‘_wsplitpath_s’ was here ../../../wine/dlls/msvcr90/../msvcrt/dir.c:1423:32: error: conflicting types for ‘_makepath_s’ /usr/lib/gcc/i686-w64-mingw32/4.6/../../../../i686-w64-mingw32/include/sec_api/stdlib_s.h:44:70: note: previous declaration of ‘_makepath_s’ was here ../../../wine/dlls/msvcr90/../msvcrt/dir.c:1527:32: error: conflicting types for ‘_wmakepath_s’ /usr/lib/gcc/i686-w64-mingw32/4.6/../../../../i686-w64-mingw32/include/sec_api/stdlib_s.h:53:70: note: previous declaration of ‘_wmakepath_s’ was here ../../../wine/dlls/msvcr90/../msvcrt/dir.c:1701:32: error: conflicting types for ‘_searchenv_s’ /usr/lib/gcc/i686-w64-mingw32/4.6/../../../../i686-w64-mingw32/include/sec_api/stdlib_s.h:46:70: note: previous declaration of ‘_searchenv_s’ was here ../../../wine/dlls/msvcr90/../msvcrt/dir.c:1832:32: error: conflicting types for ‘_wsearchenv_s’ /usr/lib/gcc/i686-w64-mingw32/4.6/../../../../i686-w64-mingw32/include/sec_api/stdlib_s.h:55:70: note: previous declaration of ‘_wsearchenv_s’ was here make[1]: *** [dir.o] Error 1 -- Alexandre Julliard julliard(a)winehq.org

April 16, 2014

Re: imagehlp: Implement (parts of) BindImageEx (try 3)

by Bernhard Reiter

*ping* -------- Ursprüngliche Nachricht -------- Von: Bernhard Reiter <ockham(a)raz.or.at> Datum: An: wine-patches(a)winehq.org Betreff: imagehlp: Implement (parts of) BindImageEx (try 3) One more fix: Pass DllPath argument on to MapAndLoad.

April 16, 2014

Re: typo fixes - rebased

by Marvin

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at https://testbot.winehq.org/JobDetails.pl?Key=6325 Your paranoid android. === w2000pro (32 bit shlfileop) === Failure running script in VM: network read timed out (wait2:ListSize:0/4)

April 16, 2014