December 2015 - Wine-devel - list.winehq.org

Re: winepulse.drv: Add a winediag message when buffer is too small.
by Andrew Eikum 10 Dec '15

10 Dec '15

On Thu, Dec 10, 2015 at 03:27:24PM +0100, Sebastian Lackner wrote: > Signed-off-by: Sebastian Lackner <sebastian(a)fds-team.de> > --- > > After pulseaudio was added, a lot of Debian people complained about missing sound. As it turns out, > this was the result of a packaging "mistake". Debian wrapper scripts set PULSE_LATENCY_MSEC to small > value by default, which worked fine with winealsa. > More than "worked fine," it actually made winealsa work better for some setups. So it was a good idea for winealsa, but is no longer needed and causes problems for winepulse. See also Bug 39744 and numerous references around the web to setting pulse's default-fragment-size-msec, which equates to about the same thing. > Even when the problem is fixed on Debian, the same problem might still occur when people have set > PULSE_LATENCY_MSEC manually on their machine, and the value is too small. http://ix.io/mJp would > be an alternative solution, but no longer allows users to control this environment variable manually. > Yeah, I definitely don't want to override the user's settings. > This patch adds at least a warning, so that users know which environment variable could be related > when sound is suddenly missing after an upgrade. Comments welcome, would like to know Andrews opinion > first how to proceed with this issue. > Adding a warning seems like a good idea, but I would do it in AudioClient_Initialize and remove the once-check. Around line 1340, warn the user if tlength is less than bufsize_bytes. Andrew

1 0

Re: [PATCH 1/2] d3dx9_36/tests: Add D3DXSaveSurfaceToFileInMemory D3DXIFF_DDS tests (try 3)
by Matteo Bruni 10 Dec '15

10 Dec '15

2015-12-09 5:35 GMT+01:00 Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com>: > Also updates the structure dds_header to be the same as d3dx9_36/surface.c. > > Signed-off-by: Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> > --- Thank you for improving the patches according to our comments. I've found more though. > dlls/d3dx9_36/tests/surface.c | 45 ++++++++++++++++++++++++++++++++++++++++++- > 1 file changed, 44 insertions(+), 1 deletion(-) > > diff --git a/dlls/d3dx9_36/tests/surface.c b/dlls/d3dx9_36/tests/surface.c > index 1697a03..eaab0f74 100644 > --- a/dlls/d3dx9_36/tests/surface.c > +++ b/dlls/d3dx9_36/tests/surface.c > @@ -211,7 +211,9 @@ struct dds_header > struct dds_pixel_format pixel_format; > DWORD caps; > DWORD caps2; > - DWORD reserved2[3]; > + DWORD caps3; > + DWORD caps4; > + DWORD reserved2; > }; > > /* fills dds_header with reasonable default values */ > @@ -1231,6 +1233,8 @@ static void test_D3DXSaveSurfaceToFileInMemory(IDirect3DDevice9 *device) > RECT rect; > ID3DXBuffer *buffer; > IDirect3DSurface9 *surface; > + struct dds_header *header; > +#define BROKEN_CAPS (DDS_CAPS_TEXTURE | 0x02) You don't need a define for this, I'd just put the expression in the ok() or maybe use a const DWORD variable for this kind of situation. FWIW 0x2 is probably DDSCAPS_ALPHA (see ddraw.h). Notice though that I always get 0x1002 in the caps field on Windows (7 and XP tested) so probably it isn't right to consider that result as broken. Did you actually get header->caps == DDS_CAPS_TEXTURE on any Windows version (and if so, which one)? If not I'd only accept (DDSCAPS_TEXTURE | DDSCAPS_ALPHA) as a valid result and add a todo_wine to the relevant ok() call. You don't need a define or const for the broken caps then and you can just put the expected result in the ok(). > > hr = IDirect3DDevice9_CreateOffscreenPlainSurface(device, 4, 4, D3DFMT_A8R8G8B8, D3DPOOL_SCRATCH, &surface, NULL); > if (FAILED(hr)) { > @@ -1248,6 +1252,45 @@ static void test_D3DXSaveSurfaceToFileInMemory(IDirect3DDevice9 *device) > ID3DXBuffer_Release(buffer); > } > > + SetRectEmpty(&rect); > + hr = D3DXSaveSurfaceToFileInMemory(&buffer, D3DXIFF_DDS, surface, NULL, &rect); > + todo_wine ok(hr == D3D_OK, "Got unexpected hr %#x\n", hr); Missing period. > + if (SUCCEEDED(hr)) > + { > + header = ID3DXBuffer_GetBufferPointer(buffer); > + > + ok(header->magic == MAKEFOURCC('D','D','S',' '), "Invalid DDS signature.\n"); > + ok(header->size == 124, "Invalid DDS size %u.\n", header->size); > + ok(!header->height, "Got unexpected height %u.\n", header->height); > + ok(!header->width, "Got unexpected width %u.\n", header->width); > + ok(!header->depth, "Got unexpected depth %u.\n", header->depth); > + ok(!header->miplevels, "Got unexpected miplevels %u.\n", header->miplevels); > + ok(!header->pitch_or_linear_size, "Got unexpected pitch_or_linear_size %u.\n", header->pitch_or_linear_size); > + ok(header->caps == DDS_CAPS_TEXTURE || broken(header->caps == BROKEN_CAPS), > + "Got unexpected caps %x.\n", header->caps); Indentation. > + ok(header->flags == (DDS_CAPS | DDS_HEIGHT | DDS_WIDTH | DDS_PIXELFORMAT), > + "Got unexpected flags %x.\n", header->flags); > + ID3DXBuffer_Release(buffer); > + } > + > + hr = D3DXSaveSurfaceToFileInMemory(&buffer, D3DXIFF_DDS, surface, NULL, NULL); > + ok(hr == D3D_OK, "Got unexpected hr %#x\n", hr); Missing period. > + header = ID3DXBuffer_GetBufferPointer(buffer); > + > + ok(header->magic == MAKEFOURCC('D','D','S',' '), "Invalid DDS signature.\n"); > + todo_wine ok(header->size == 124, "Invalid DDS size %u.\n", header->size); > + ok(header->height == 4, "Got unexpected height %u.\n", header->height); > + ok(header->width == 4, "Got unexpected width %u.\n", header->width); > + todo_wine ok(!header->depth, "Got unexpected depth %u.\n", header->depth); > + todo_wine ok(!header->miplevels, "Got unexpected miplevels %u.\n", header->miplevels); > + todo_wine ok(!header->pitch_or_linear_size, "Got unexpected pitch_or_linear_size %u.\n", header->pitch_or_linear_size); > + ok(header->caps == DDS_CAPS_TEXTURE || broken(header->caps == BROKEN_CAPS), > + "Got unexpected caps %x.\n", header->caps); > + todo_wine ok(header->flags == (DDS_CAPS | DDS_HEIGHT | DDS_WIDTH | DDS_PIXELFORMAT), > + "Got unexpected flags %x.\n", header->flags); > + ID3DXBuffer_Release(buffer); > + > +#undef BROKEN_CAPS > IDirect3DSurface9_Release(surface); > } Looks good to me otherwise.

1 0

Release status
by Alexandre Julliard 10 Dec '15

10 Dec '15

Folks, I'm pleased with the progress we've made since code freeze started, with about 40 regressions and many other bugs fixed. Thanks to everybody for taking the time to fix them! At this point it looks like we are running out of easily fixable bugs, so my plan is that rc4 is going to be the last release candidate, and that 1.8 will be released next week. Note also that rc4 is going to be a little delayed since I'm travelling on Friday; it will be out during the weekend. Thank you all for your help! -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH] msvcrt: Add tmpnam_s() and _wtmpnam_s() with tests.
by Piotr Caban 10 Dec '15

10 Dec '15

Hi, it this patch fixing any application or is it only letting it execute a little further? It doesn't compile: ../../../wine_src/dlls/msvcr71/msvcr71.spec:667: function 'MSVCRT_wtmpnam' not defined ../../../wine_src/dlls/msvcr70/msvcr70.spec:671: function 'MSVCRT_wtmpnam' not defined ... On 12/10/15 00:08, lejcik wrote: > /********************************************************************* > + * tmpnam_s (MSVCRT.@) > + */ > +int CDECL MSVCRT_tmpnam_s(char *s, MSVCRT_size_t size) > +{ > + char tmpstr[16]; > + char *p; > + int count, tmpsize; > + > + if (!MSVCRT_CHECK_PMT(s != NULL)) > + return MSVCRT_EINVAL; > + > + msvcrt_int_to_base32(GetCurrentProcessId(), tmpstr); > + tmpsize = MSVCRT__snprintf(s, size, "\\s%s.", tmpstr); > + if (tmpsize == -1) > + return MSVCRT_ERANGE; > + > + p = s + tmpsize; You can reuse s variable here (s += tmpsize). > + size -= tmpsize; > + for (count = 0; count < MSVCRT_TMP_MAX; count++) According to MSDN TMP_MAX_S should be used here. It's different then TMP_MAX. > + { > + tmpsize = msvcrt_int_to_base32(tmpnam_unique++, tmpstr); > + if ((MSVCRT_size_t) tmpsize > size) > + return MSVCRT_ERANGE; Is invalid parameter handler called in this case? > char * CDECL MSVCRT_tmpnam(char *s) > { ... > + return MSVCRT_tmpnam_s(s, MAX_PATH) == 0 ? s : NULL; You can't use tmpnam_s here because of TMP_MAX and TMP_MAX_S inequality (unless MSDN is wrong here). Thanks, Piotr

1 0

Re: winhttp/tests: Add an authentication test using IWinHttpRequest.
by Hans Leidekker 10 Dec '15

10 Dec '15

It doesn't build cleanly for me: winhttp.c: In function ‘test_IWinHttpRequest’: winhttp.c:3631:5: warning: implicit declaration of function ‘sprintf’ [-Wimplicit-function-declaration] sprintf( buf, "http://localhost:%d/auth", port ); ^ winhttp.c:3631:5: warning: incompatible implicit declaration of built-in function ‘sprintf’

1 0

Re: programs/iexplore: sync registry and program resource values
by Alexandre Julliard 10 Dec '15

10 Dec '15

Austin English <austinenglish(a)gmail.com> writes: > From: Austin English <austinenglish(a)gmail.com> > Date: Tue, 8 Dec 2015 17:05:28 -0600 > Subject: [PATCH] programs/iexplore: sync registry and program resource values > > Signed-off-by: Austin English <austinenglish(a)gmail.com> > --- > programs/iexplore/iexplore.rc | 8 ++++---- > 1 file changed, 4 insertions(+), 4 deletions(-) Which bug is this for? -- Alexandre Julliard julliard(a)winehq.org

2 2

Re: [PATCH] wined3d: Restore WINEDDBLT_ALPHATEST support.
by Henri Verbeet 09 Dec '15

09 Dec '15

On 7 December 2015 at 23:36, Stefan Dösinger <stefandoesinger(a)gmx.at> wrote: > -static void arbfp_blit_surface(struct wined3d_device *device, DWORD filter, > +static void arbfp_blit_surface(enum wined3d_blit_op op, struct wined3d_device *device, DWORD filter, I think I'd prefer "device" to stay the first parameter for consistency with the other blitter ops. It's probably not a big deal though. > @@ -7881,6 +7885,9 @@ static void arbfp_blit_surface(struct wined3d_device *device, DWORD filter, > /* Leave the opengl state valid for blitting */ > arbfp_blit_unset(context->gl_info); > > + if (op == WINED3D_BLIT_OP_COLOR_BLIT_ALPHATEST) > + glDisable(GL_ALPHA_TEST); > + You shouldn't need this. context_apply_blit_state() invalidates WINED3D_RS_ALPHATESTENABLE, but even if it didn't, disabling the alpha test isn't necessarily the right thing to do. Not that arbfp_blit_surface() necessarily needs all the invalidation context_apply_blit_state() does though. > @@ -4335,8 +4336,16 @@ static void ffp_blit_blit_surface(struct wined3d_device *device, DWORD filter, > wined3d_texture_set_color_key(src_surface->container, WINED3D_CKEY_SRC_BLT, color_key); > > context = context_acquire(device, dst_surface); > + > + if (op == WINED3D_BLIT_OP_COLOR_BLIT_ALPHATEST) > + glEnable(GL_ALPHA_TEST); > + > surface_blt_to_drawable(device, context, filter, > !!color_key, src_surface, src_rect, dst_surface, dst_rect); > + > + if (op == WINED3D_BLIT_OP_COLOR_BLIT_ALPHATEST) > + glDisable(GL_ALPHA_TEST); > + Like above. Some basic tests probably wouldn't hurt. If we had had any the regression might not have happened in the first place.

2 3

Re: [PATCH 3/3] bthprops.cpl: add BluetoothFindFirstRadio/BluetoothFindRadioClose stubs
by Bruno Jesus 09 Dec '15

09 Dec '15

On Sat, Dec 5, 2015 at 7:21 AM, Austin English <austinenglish(a)gmail.com> wrote: > > +HBLUETOOTH_RADIO_FIND WINAPI BluetoothFindFirstRadio(BLUETOOTH_FIND_RADIO_PARAMS *params, HANDLE *radio) > +{ > + FIXME("(%p %p): stub!\n", params, radio); > + *radio = NULL; > + SetLastError(ERROR_CALL_NOT_IMPLEMENTED); > + return NULL; > +} Please set the last error to ERROR_NO_MORE_ITEMS. This is the common return when no devices are found and will make other bluetooth appications more happy.

1 0

Re: [1/2] riched32/tests: Add selection tests.
by Jactry Zeng 09 Dec '15

09 Dec '15

2015-12-08 21:55 GMT+08:00 Jactry Zeng <jzeng(a)codeweavers.com>: > > This patch try to reproduce reason of bug 32112: https://bugs.winehq.org/show_bug.cgi?id=32112 > riched32's selection is different to riched20's. It'll count some multibyte characters as two when riched20 count them as two. Sorry, a typo here: riched32's selection is different to riched20's. It'll count some multibyte characters as two when riched20 counts them as *one*. -- Regards, Jactry Zeng

1 0

Re: [PATCH] Programs: fix extrac32 heap leak
by Alex Henrie 08 Dec '15

08 Dec '15

> diff --git a/programs/extrac32/extrac32.c b/programs/extrac32/extrac32.c > index 63205e8..a6e7438 100644 > --- a/programs/extrac32/extrac32.c > +++ b/programs/extrac32/extrac32.c > @@ -207,6 +207,8 @@ static LPWSTR *get_extrac_args(LPWSTR cmdline, int *pargc) > str++; > } > > + HeapFree(GetProcessHeap(), 0, str); > + > argv[argc] = NULL; > *pargc = argc; I don't think this is right. Pointers to str are returned in the argv array, and it's the caller's responsibility to free them. And in this case, there's not much to be gained by freeing these pointers before the end of the program anyway. By the way, when writing patches for Wine programs in the future, please use the name of the program as the component name. For example, "extrac32: Fix heap leak." -Alex

1 0