May 2011 - Wine-Devel - list.winehq.org

Re: no line numbers in backtraces on Mac OS X?

by Eric Pouech

we miss dwarf support on mac it's likely the cause here (if you can, you can try to compile with stabs, it should work) (there's an open bug in bugzilla for this) A+ 2011/5/23 <Joerg-Cyril.Hoehle(a)t-systems.com> > Hi, > > Is it normal that there are no source line numbers in backtraces on Mac OS? > Despite many patches to winedbg etc., the backtraces I get are > no better that those from 1.1.24 times. > 18 0x41f5b2a1 _OpenDriverA+0x51() in winmm (0x0032ce78) > 19 0x4041467f _initAudioDlg+0x75f() in winecfg (0x0032dc28) > > I'm using configure + make to build. On Linux, the line numbers help > a lot in locating bogus code. > > Thank for your help, > Jörg Höhle. > > > -- -- Eric Pouech

May 23, 2011

no line numbers in backtraces on Mac OS X?

by Joerg-Cyril.Hoehle＠t-systems.com

Hi, Is it normal that there are no source line numbers in backtraces on Mac OS? Despite many patches to winedbg etc., the backtraces I get are no better that those from 1.1.24 times. 18 0x41f5b2a1 _OpenDriverA+0x51() in winmm (0x0032ce78) 19 0x4041467f _initAudioDlg+0x75f() in winecfg (0x0032dc28) I'm using configure + make to build. On Linux, the line numbers help a lot in locating bogus code. Thank for your help, Jörg Höhle.

May 23, 2011

mmsystem16.h incorrect? SEGPTR vs. LPSTR

by Joerg-Cyril.Hoehle＠t-systems.com

Alexandre, Damjan, Eric, I understand your answers mean that a patch is needed to MCI_*_PARMS16 in mmsystem16.h to turn all occurrences of LPSTR into SEGPTR. E.g. MCI_LOAD_PARMS16, MCI_DGV_INFO_PARMS16, MCI_DGV_OPEN_PARMS16 ... I suppose "DWORD dwCallback;" is ok to copy 1:1 (what the code currently does) because it contains a window handle, not pointer. It would be wrong if the 16bit SDK contained HWND16 wCallback; UINT16 wReserved0; because then Wine might see half garbage and be lucky when wReserved is 0. Regards, Jörg Höhle

May 23, 2011

Re: mmsystem16.h incorrect? SEGPTR vs. LPSTR

by Damjan Jovanovic

On Mon, May 23, 2011 at 2:54 PM, <Joerg-Cyril.Hoehle(a)t-systems.com> wrote: > Executive summary: I believe LPSTR is all wrong in mmsystem16.h's MCI defs. > > Hi, > > I've some doubts about the correctness of some definitions in > mmsystem16.h, esp. the use of SEGPTR vs. LPSTR types. > > Please compare > typedef struct { > DWORD dwCallback; > LPSTR lpstrReturn; > ... > } MCI_DGV_INFO_PARMS16, * LPMCI_DGV_INFO_PARMS16; > http://source.winehq.org/source//include/wine/mmsystem16.h > typedef struct { > DWORD dwCallback; > SEGPTR lpstrReturn; > ... > } MCI_INFO_PARMS16, *LPMCI_INFO_PARMS16; > > These definitions do not match. SEGPTR requires the use of the > MapSL() helper function to correctly access the segmented pointer, > while LPSTR is a linear string pointer. > > For reasons I recently explained in > http://www.winehq.org/pipermail/wine-devel/2011-May/090060.html > I very much doubt that both pointer types were used simultaneously -- unless > the digital video device (mciavi16.dll?) embedded different return > types in its resource file. It's not impossible, just very unlikely. > > Within one winmm, The return type MCI_STRING (value 0001) can *either* denote > linear C strings, or segmented ones. Should 2 types really have been > used as the above structure definitions suggest, then one will > have a distinct return value in its resource definition file > (e.g. mciavi_res.rc or extractable from the .dll). For instance, > that's how we derived that there must exist a distinct value for > 64bit use we then named MCI_INTEGER64 in Wine's mmddk.h > > However, I feel it's more likely that LPSTR above is wrong, that no > distinct return type was used and that segmented pointers were used in > most places within the MCI in 16 bit times, partly because Wine > currently uses the SEGPTR definitions and MapSL() helpers in its > current code base and would have crashed otherwise (but if that code > is never exercised?). > > Likewise for MCI_OPEN_PARMS16 and MCI_DGV_OPEN_PARMS16. Wine's > current MCI code would choke if it had to deal with 2 different types > of pointers here. > > To give another example, I find it illogical that MCI_LOAD_PARMS16 > takes an LPCSTR, while MCI_SOUND_PARMS16 takes a SEGPTR according to > mmsystem16.h. It's not about a different name, it's entirely > different content! > > I'd be pleased if somebody could grab out an old SDK and have a look > at the includes or otherwise shed light on this issue. > > Thank you for your help, > Jörg Höhle > > > AFAICT Wine's 16 bit headers cannot be used to compile Win16 code into a Winelib. The only purpose of Wine's 16 bit headers is to allow 32 bit code to read and write 16 bit structures and thus play nicely with 16 bit binaries. That's why Wine's types and names of 16 bit fields, structures, functions and headers don't always match what Win16 actually used, and why those headers are in a special directory that won't exist on any Windows API (include/wine/). According to Openwatcom's headers, LPSTR is a FAR pointer to char. "FAR" is a 16:16 segmented pointer, which makes sense, because Win16, being 16 bit, had no 32 bit linear pointer type in user-space (only kernel VxDs had it). So the Win16 MCI headers' LP[C]STRs were still 16:16 pointers to char. The SEGPTR type doesn't exist on Win16 at all. It's a Wine invention to conveniently mark a 16:16 pointer and since it's an integer type, it catches attempts to assign directly between 32 bit linear pointers and SEGPTRs with compiler warnings. So it seems Wine's headers should never use LP[C]STR in 16 bit structures, SEGPTR should always be used instead. Damjan Jovanovic

May 23, 2011

Re: mmsystem16.h incorrect? SEGPTR vs. LPSTR

by Alexandre Julliard

<Joerg-Cyril.Hoehle(a)t-systems.com> writes: > I'd be pleased if somebody could grab out an old SDK and have a look > at the includes or otherwise shed light on this issue. You won't find anything in the SDK. SEGPTR is a Wine invention to help distinguish pointers that have already been converted. The SDK uses LPSTR etc. throughout, but these are obviously all segmented pointers in 16-bit code. -- Alexandre Julliard julliard(a)winehq.org

May 23, 2011

mmsystem16.h incorrect? SEGPTR vs. LPSTR

by Joerg-Cyril.Hoehle＠t-systems.com

Executive summary: I believe LPSTR is all wrong in mmsystem16.h's MCI defs. Hi, I've some doubts about the correctness of some definitions in mmsystem16.h, esp. the use of SEGPTR vs. LPSTR types. Please compare typedef struct { DWORD dwCallback; LPSTR lpstrReturn; ... } MCI_DGV_INFO_PARMS16, * LPMCI_DGV_INFO_PARMS16; http://source.winehq.org/source//include/wine/mmsystem16.h typedef struct { DWORD dwCallback; SEGPTR lpstrReturn; ... } MCI_INFO_PARMS16, *LPMCI_INFO_PARMS16; These definitions do not match. SEGPTR requires the use of the MapSL() helper function to correctly access the segmented pointer, while LPSTR is a linear string pointer. For reasons I recently explained in http://www.winehq.org/pipermail/wine-devel/2011-May/090060.html I very much doubt that both pointer types were used simultaneously -- unless the digital video device (mciavi16.dll?) embedded different return types in its resource file. It's not impossible, just very unlikely. Within one winmm, The return type MCI_STRING (value 0001) can *either* denote linear C strings, or segmented ones. Should 2 types really have been used as the above structure definitions suggest, then one will have a distinct return value in its resource definition file (e.g. mciavi_res.rc or extractable from the .dll). For instance, that's how we derived that there must exist a distinct value for 64bit use we then named MCI_INTEGER64 in Wine's mmddk.h However, I feel it's more likely that LPSTR above is wrong, that no distinct return type was used and that segmented pointers were used in most places within the MCI in 16 bit times, partly because Wine currently uses the SEGPTR definitions and MapSL() helpers in its current code base and would have crashed otherwise (but if that code is never exercised?). Likewise for MCI_OPEN_PARMS16 and MCI_DGV_OPEN_PARMS16. Wine's current MCI code would choke if it had to deal with 2 different types of pointers here. To give another example, I find it illogical that MCI_LOAD_PARMS16 takes an LPCSTR, while MCI_SOUND_PARMS16 takes a SEGPTR according to mmsystem16.h. It's not about a different name, it's entirely different content! I'd be pleased if somebody could grab out an old SDK and have a look at the includes or otherwise shed light on this issue. Thank you for your help, Jörg Höhle

May 23, 2011

Request for comments: d3dX : wined3d implicit enum conversion warnings and resolution

by Frédéric Delanoy

Hi guys, while trying to compile wine with clang, I get a whole bunch of "implicit conversion from enumeration type FOO to different enumeration type WINED3DFOO" (and conversely) warnings when an enum value of type A is assigned to a var of enum type B. In order to tackle those, instead of using (probably ugly) casts everywhere, I put them in a "static inline" function in d3dX_private.h (conv functions put in the header so inlining is possible everywhere, so a potential performance impact is limited) Originally, I intended to issue a FIXME("Unhandled/unrecognized value...") in the default: clause of a switch statement to catch values not in in the original enum's domain, but I couldn't find a way to make FIXMEs work in a .h file 'cos of a macro expansion hell I'm not quite sure however what the best solution is wrt performance, so I'd like your comments on this Some explanations: *D3DPOOL and WINED3DPOOL have a full 1-1 mapping, so the conv functions simply do a cast. *D3DRESOURCETYPE and WINED3DRESOURCETYPE have *not* a 1-1 mapping, and I present here 3 possible solution types: D3DRESOURCETYPE => D3DRTYPE_SURFACE = 1, D3DRTYPE_VOLUME = 2, D3DRTYPE_TEXTURE = 3, D3DRTYPE_VOLUMETEXTURE = 4, D3DRTYPE_CUBETEXTURE = 5, D3DRTYPE_VERTEXBUFFER = 6, D3DRTYPE_INDEXBUFFER = 7, D3DRTYPE_FORCE_DWORD = 0x7fffffff WINED3DRESOURCETYPE => WINED3DRTYPE_SURFACE = 1, WINED3DRTYPE_VOLUME = 2, WINED3DRTYPE_TEXTURE = 3, WINED3DRTYPE_VOLUMETEXTURE = 4, WINED3DRTYPE_CUBETEXTURE = 5, WINED3DRTYPE_BUFFER = 6, WINED3DRTYPE_FORCE_DWORD = 0x7fffffff Type 1 --------- Identical semantics to direct assigment All cases specified, even "obvious" ones (D3D_FOO <-> WINED3D_FOO), using same enum value; non mapped values use the default: clause (e.g. D3DRTYPE_INDEXBUFFER=7; no 7 enum val in WINED3DRESOURCETYPE) Default clause: only used for values not in original enum Type 2 --------- Same as v1 except obvious values aren't specified in a case: Non directly obvious values (same enum val) can still be "highlighted" (e.g. WINED3DRTYPE_BUFFER=6 ==> D3DRTYPE_VERTEXBUFFER=6) Default clause: used for values not in original enum AND obvious values Maybe more optimal than "type 1", but isn't that a micro/premature optimisation? Type 3 --------- Potentially more risky/invasive: conversion does not simply return value "as-is" Semantics *changes* from direct assigment; some additional code changes are necessary; could help centralise, though e.g. D3DRTYPE_VERTEXBUFFER=6 => WINED3DRTYPE_BUFFER=6 D3DRTYPE_INDEXBUFFER=7 => WINED3DRTYPE_BUFFER=6 Adapted code (directx.c): - switch(RType) { - case D3DRTYPE_VERTEXBUFFER: - case D3DRTYPE_INDEXBUFFER: - WineD3DRType = WINED3DRTYPE_BUFFER; - break; - - default: - WineD3DRType = RType; - break; - } + WineD3DRType = wined3dresourcetype_from_d3dresourcetype(RType); What do you think would be best? Frédéric

May 23, 2011

Re: ddraw: Don't interpret end padding as dwCaps2 for x64.

by Alexandre Julliard

Dylan Smith <dylan.ah.smith(a)gmail.com> writes: > @@ -3123,6 +3123,8 @@ static HRESULT CreateSurface(IDirectDrawImpl *ddraw, DDSURFACEDESC2 *DDSD, > memset(&desc2, 0, sizeof(desc2)); > desc2.dwSize = sizeof(desc2); /* For the struct copy */ > DD_STRUCT_COPY_BYSIZE(&desc2, DDSD); > + if (DDSD->dwSize == sizeof(DDSURFACEDESC)) > + CLEAR_STRUCT_END_PADDING((DDSURFACEDESC*)&desc2, ddsCaps); I'd suggest to rethink the whole thing. We don't really need to set the same structure fields 4 times. -- Alexandre Julliard julliard(a)winehq.org

May 23, 2011

Re: urlmon: Fix incorrect pointer arithmetic (too conversative) in map_url_to_zone.

by Nikolay Sivov

On Sun, May 22, 2011 at 10:53 PM, Gerald Pfeifer <gerald(a)pfeifer.com> wrote: > The difference between two pointers (of the same type) is the number > of elements, not the number of bytes. That's why it was divided by element size to get 'number of elements'. What's a problem? > Thus the code below was way > incorrect, luckily only too conversative. > > Gerald > > --- > dlls/urlmon/sec_mgr.c | 2 +- > 1 files changed, 1 insertions(+), 1 deletions(-) > > diff --git a/dlls/urlmon/sec_mgr.c b/dlls/urlmon/sec_mgr.c > index 7b4bb35..75850ee 100644 > --- a/dlls/urlmon/sec_mgr.c > +++ b/dlls/urlmon/sec_mgr.c > @@ -529,7 +529,7 @@ static HRESULT map_url_to_zone(LPCWSTR url, DWORD *zone, LPWSTR *ret_url) > hres = CoInternetParseUrl(secur_url, PARSE_PATH_FROM_URL, 0, path, > sizeof(path)/sizeof(WCHAR), &size, 0); > > - if(SUCCEEDED(hres) && (ptr = strchrW(path, '\\')) && ptr-path < sizeof(root)/sizeof(WCHAR)) { > + if(SUCCEEDED(hres) && (ptr = strchrW(path, '\\')) && ptr-path < sizeof(root)) { > UINT type; > > memcpy(root, path, (ptr-path)*sizeof(WCHAR)); > -- > 1.7.4.1 > > >

May 22, 2011

Re: comctl32: Shed an unused parameter of TOOLTIPS_HitTestT and rename to TOOLTIPS_HitTest. (RESEND)

by Nikolay Sivov

On Sun, May 22, 2011 at 6:07 PM, Gerald Pfeifer <gerald(a)pfeifer.com> wrote: > Any problems with this patch I submitted back in February? > I believe I mentioned it, or I should have. Message handler is incomplete, cause it should return text as well, that's why it has A/W versions. This needs tests and proper implementation, not just reducing parameters list for no reason. (and if tests show no use of text for these messages it's a reason of course) > Gerald > > --- > dlls/comctl32/tooltips.c | 7 +++---- > 1 files changed, 3 insertions(+), 4 deletions(-) > > diff --git a/dlls/comctl32/tooltips.c b/dlls/comctl32/tooltips.c > index a2e58e0..445845c 100644 > --- a/dlls/comctl32/tooltips.c > +++ b/dlls/comctl32/tooltips.c > @@ -1396,8 +1396,7 @@ TOOLTIPS_GetToolInfoT (const TOOLTIPS_INFO *infoPtr, TTTOOLINFOW *ti, BOOL isW) > > > static LRESULT > -TOOLTIPS_HitTestT (const TOOLTIPS_INFO *infoPtr, LPTTHITTESTINFOW lptthit, > - BOOL isW) > +TOOLTIPS_HitTest (const TOOLTIPS_INFO *infoPtr, LPTTHITTESTINFOW lptthit) > { > TTTOOL_INFO *toolPtr; > INT nTool; > @@ -2209,8 +2208,8 @@ TOOLTIPS_WindowProc (HWND hwnd, UINT uMsg, WPARAM wParam, LPARAM lParam) > > case TTM_HITTESTA: > case TTM_HITTESTW: > - return TOOLTIPS_HitTestT (infoPtr, (LPTTHITTESTINFOW)lParam, > - uMsg == TTM_HITTESTW); > + return TOOLTIPS_HitTest (infoPtr, (LPTTHITTESTINFOW)lParam); > + > case TTM_NEWTOOLRECTA: > case TTM_NEWTOOLRECTW: > return TOOLTIPS_NewToolRectT (infoPtr, (LPTTTOOLINFOW)lParam); > -- > 1.7.4.1 > > >

May 22, 2011

Re: TestBot job 11094 results: Re: comctl32/tests: Add casts to avoid comparison of different int types. [take 2]

by Gerald Pfeifer

On Mon, 23 May 2011, Marvin wrote: > VM Status Number of test failures > WINEBUILD completed > WNT4WSSP6 completed 0 > W2KPROSP4 completed 0 > WXPPROSP3 completed 0 > W2K3R2SESP2 failed > WVISTAADM completed 0 > W2K8SE completed 0 > W7PRO completed 0 > W7PROX64 completed 0 > W7PROX64 completed 0 For the record, I plead not guilty as far as the following goes. === W2K3R2SESP2 (32 bit tab) === Can't copy exe to VM: Failed to resolve host :-) Gerald

May 22, 2011

Re: Big ugly build changes might be needed for Debian/Ubuntu

by Ove Kåven

Den 22. mai 2011 21:48, skrev Scott Ritchie: > On 05/22/2011 03:53 AM, Ove Kåven wrote: >> Den 21. mai 2011 19:38, skrev Scott Ritchie: >>> Well now we can just Depends: wine:i386 or similar, I think. >> >> Last time I asked, they said the first multiarch implementation will not >> support that. The implementation cost of that would be very high (among >> other things, it might violate the current architecture >> self-containment), and the number of packages needing it is too small. >> > > Looking at the spec, it seems like Depends: wine:i386 isn't allowed, > however we can create a unique package name that only exists on i386 (eg > wine32), declare it Multiarch: foreign, and then have wine depend on wine32. Yes, that was what I was thinking of doing if I had to. But it's really not very clean...

May 22, 2011

Re: user32: Partially implement FlashWindowEx with tests

by Detlef Riekenberg

On So, 2011-05-22, André Hentschel wrote > Thanks, will get into try 2 > pfwi->cbSize < sizeof(FLASHWINFO) || pfwi->cbSize > sizeof(FLASHWINFO) That code can be reduced to: pfwi->cbSize != sizeof(FLASHWINFO) -- By by ... Detlef

May 22, 2011

Re: ddraw: Don't interpret end padding as dwCaps2 for x64.

by Dylan Smith

On Sun, May 22, 2011 at 3:10 AM, Stefan Dösinger <stefandoesinger(a)gmx.at> wrote: > On Sunday 22 May 2011 06:57:30 Dylan Smith wrote: >> 4 bytes of padding are at the end of DDSURFACEDESC when compiling for >> x64, because it contains a pointer field followed by an odd number >> of DWORD fields. > Does DDSURFACEDESC have the same padding when using the microsoft headers and > msvc? > I tried and got the same results for x64: offsetof(DDSURFACEDESC, ddsCaps) = 112 offsetof(DDSURFACEDESC, ddsCaps) + sizeof(DDSCAPS) = 116 sizeof(DDSURFACEDESC) = 120

May 22, 2011

Re: Big ugly build changes might be needed for Debian/Ubuntu

by Scott Ritchie

On 05/22/2011 03:53 AM, Ove Kåven wrote: > Den 21. mai 2011 19:38, skrev Scott Ritchie: >> Well now we can just Depends: wine:i386 or similar, I think. > > Last time I asked, they said the first multiarch implementation will not > support that. The implementation cost of that would be very high (among > other things, it might violate the current architecture > self-containment), and the number of packages needing it is too small. > Looking at the spec, it seems like Depends: wine:i386 isn't allowed, however we can create a unique package name that only exists on i386 (eg wine32), declare it Multiarch: foreign, and then have wine depend on wine32.

May 22, 2011

Re: comctl32/tests: Add casts to avoid comparison of different int types. [take 2]

by Gerald Pfeifer

On Wed, 16 Feb 2011, Alexandre Julliard wrote: >> This is the first hunk of a patch a few days ago; testbot.winehq.org >> thinks it's fine: https://testbot.winehq.org/JobDetails.pl?Key=9291 > You don't need casts. If the variable has the wrong type you can change > that. At first I didn't realize what you had in mind, but now I believe the patch below does this. :-) Gerald ChangeLog: comctl32/tests: Rewrite a test in test_TCS_OWNERDRAWFIXED to avoid a type mismatch or cast. --- dlls/comctl32/tests/tab.c | 7 ++++--- 1 files changed, 4 insertions(+), 3 deletions(-) diff --git a/dlls/comctl32/tests/tab.c b/dlls/comctl32/tests/tab.c index 8b64764..7f5bbb7 100644 --- a/dlls/comctl32/tests/tab.c +++ b/dlls/comctl32/tests/tab.c @@ -1313,6 +1313,7 @@ static void test_TCM_SETITEMEXTRA(void) static void test_TCS_OWNERDRAWFIXED(void) { LPARAM lparam, lparam2; + ULONG_PTR itemdata; TCITEMA item; HWND hTab; BOOL ret; @@ -1335,9 +1336,9 @@ static void test_TCS_OWNERDRAWFIXED(void) ShowWindow(hTab, SW_SHOW); RedrawWindow(hTab, NULL, 0, RDW_UPDATENOW); - lparam = 0; - memset(&lparam, 0xde, 4); - ok(g_drawitem.itemData == lparam, "got %lx, expected %lx\n", g_drawitem.itemData, lparam); + itemdata = 0; + memset(&itemdata, 0xde, 4); + ok(g_drawitem.itemData == itemdata, "got %lx, expected %lx\n", g_drawitem.itemData, itemdata); DestroyWindow(hTab); -- 1.7.4.1

May 22, 2011

Re: Big ugly build changes might be needed for Debian/Ubuntu

by Marcus Meissner

On Sun, May 22, 2011 at 06:26:45AM -0700, James McKenzie wrote: > On 5/21/11 11:51 PM, Marcus Meissner wrote: > >On Sat, May 21, 2011 at 05:57:40PM -0700, James McKenzie wrote: > >>On 5/19/11 1:51 PM, Marcus Meissner wrote: > >>>It is trivial. You basically build wine twice, once in the 32bit > >>>environment and once inthe 64bit one. > >>> > >>Maybe Wine can look into some sort of merge process to build both > >>and at runtime select one or the other? I don't know if this is > >>possible, I'm just 'throwing the idea out there'. > >Not sure what you mean. > > > >If you install 64 bit wine + 32bit stuff I listed, you will have > >a 64+32 bit capable wine setup. > > > That was the answer I was looking for. Thank you, Marcus. The > ability is already there. So, if I install 64 bit wine on a 64 bit > machine it should work something like what Windows does now, one > directory (wineprefix) for 64 bitness and one directory (wineprefix) > for 32 bitness. wineprefix for the installed .exe and .dlls? as in ~/.wine/ ? A 64bit install and/or a 32bit install will use just 1 wineprefix (~/.wine) If the setup is done by the 64bit binary, it will be set up in away that both 64 and 32bit windows binaries will work. But 64bit and 32bit ~/.wine are not seperate. Ciao, Marcus

May 22, 2011

Re: user32: Partially implement FlashWindowEx with tests

by André Hentschel

Am 22.05.2011 09:58, schrieb Dmitry Timoshkov: > André Hentschel <nerv(a)dawncrow.de> wrote: > >> + if (!pfwi->hwnd || pfwi->cbSize < sizeof(FLASHWINFO)) >> + { >> + SetLastError(ERROR_INVALID_PARAMETER); >> + return FALSE; >> + } > > I don't see in your tests the case for 'cbSize > sizeof(FLASHWINFO)'. > Also using !IsWindow() instead of !hwnd would fix one more test case. > Thanks, will get into try 2 -- Best Regards, André Hentschel

May 22, 2011

Re: Big ugly build changes might be needed for Debian/Ubuntu

by James McKenzie

On 5/21/11 11:51 PM, Marcus Meissner wrote: > On Sat, May 21, 2011 at 05:57:40PM -0700, James McKenzie wrote: >> On 5/19/11 1:51 PM, Marcus Meissner wrote: >>> It is trivial. You basically build wine twice, once in the 32bit >>> environment and once inthe 64bit one. >>> >> Maybe Wine can look into some sort of merge process to build both >> and at runtime select one or the other? I don't know if this is >> possible, I'm just 'throwing the idea out there'. > Not sure what you mean. > > If you install 64 bit wine + 32bit stuff I listed, you will have > a 64+32 bit capable wine setup. > That was the answer I was looking for. Thank you, Marcus. The ability is already there. So, if I install 64 bit wine on a 64 bit machine it should work something like what Windows does now, one directory (wineprefix) for 64 bitness and one directory (wineprefix) for 32 bitness. James McKenzie

May 22, 2011

Re: Big ugly build changes might be needed for Debian/Ubuntu

by Ove Kåven

Den 21. mai 2011 19:38, skrev Scott Ritchie: > Well now we can just Depends: wine:i386 or similar, I think. Last time I asked, they said the first multiarch implementation will not support that. The implementation cost of that would be very high (among other things, it might violate the current architecture self-containment), and the number of packages needing it is too small. > I don't mean the multiarch change, I mean the "build daemons will not > allow have foreign multiarch build dependencies" change, which I guess > isn't a change so much as it is a property of the new system. Well that's not any different from how it works now. There wouldn't be any management implications of that. >> Multiarch has no management implications whatsoever for the build >> daemons. All official Debian architectures are perfectly self-contained >> as they are. (Some of them currently contain cross-compiled pieces of >> other architectures, which would go away, but that does not concern the >> build daemons or the infrastructure.) > > Multiarch with foreign architecture build support would eliminate the > self-containment, wouldn't it? As far as the infrastructure is concerned, it is the runtime dependencies that are important. Until now, nothing in Debian had foreign architecture support, apart from Arch:all packages (which don't really count), which is why we now have the abomination that is ia32-libs. For the first version of multiarch, the buildds and dak are not going to change - only the package managers (and package build tools) would need to be aware of multiarch. Maybe foreign architecture support for the archive tools are added later, maybe not. Either way, infrastructure maintenance is not going to get easier.

May 22, 2011

Re: user32: Partially implement FlashWindowEx with tests

by Dmitry Timoshkov

André Hentschel <nerv(a)dawncrow.de> wrote: > + if (!pfwi->hwnd || pfwi->cbSize < sizeof(FLASHWINFO)) > + { > + SetLastError(ERROR_INVALID_PARAMETER); > + return FALSE; > + } I don't see in your tests the case for 'cbSize > sizeof(FLASHWINFO)'. Also using !IsWindow() instead of !hwnd would fix one more test case. -- Dmitry.

May 22, 2011

Re: ddraw: Don't interpret end padding as dwCaps2 for x64.

by Stefan Dösinger

On Sunday 22 May 2011 06:57:30 Dylan Smith wrote: > 4 bytes of padding are at the end of DDSURFACEDESC when compiling for > x64, because it contains a pointer field followed by an odd number > of DWORD fields. Does DDSURFACEDESC have the same padding when using the microsoft headers and msvc?

May 22, 2011

Re: Big ugly build changes might be needed for Debian/Ubuntu

by Marcus Meissner

On Sat, May 21, 2011 at 05:57:40PM -0700, James McKenzie wrote: > On 5/19/11 1:51 PM, Marcus Meissner wrote: > >It is trivial. You basically build wine twice, once in the 32bit > >environment and once inthe 64bit one. > > > Maybe Wine can look into some sort of merge process to build both > and at runtime select one or the other? I don't know if this is > possible, I'm just 'throwing the idea out there'. Not sure what you mean. If you install 64 bit wine + 32bit stuff I listed, you will have a 64+32 bit capable wine setup. Ciao, Marcus

May 22, 2011

Re: Big ugly build changes might be needed for Debian/Ubuntu

by James McKenzie

On 5/19/11 1:51 PM, Marcus Meissner wrote: > It is trivial. You basically build wine twice, once in the 32bit > environment and once inthe 64bit one. > Maybe Wine can look into some sort of merge process to build both and at runtime select one or the other? I don't know if this is possible, I'm just 'throwing the idea out there'. James McKenzie

May 21, 2011

Re: Big ugly build changes might be needed for Debian/Ubuntu

by James McKenzie

On 5/19/11 1:51 PM, Marcus Meissner wrote: > It is trivial. You basically build wine twice, once in the 32bit > environment and once inthe 64bit one. > Maybe Wine can look into some sort of merge process to build both and at runtime select one or the other? I don't know if this is possible, I'm just 'throwing the idea out there'. James McKenzie

May 21, 2011

Re: Big ugly build changes might be needed for Debian/Ubuntu

by Scott Ritchie

On 05/21/2011 08:20 AM, Ove Kåven wrote: > Den 19. mai 2011 20:58, skrev Scott Ritchie: >> A side effect of this change, however, is the current build daemons ONLY >> have packages for one architecture. This means that, at build time, you >> won't be able to pull in the 32-bit packages on a 64-bit build daemon. > > Well, I don't see the problem. I've been more concerned that there's not > yet any clean way to get the 64-bit package to depend on the 32-bit > package, but for the build itself, it's quite trivial. (At least it was > a year ago or so, and I can't imagine it has gotten worse since then.) > Well now we can just Depends: wine:i386 or similar, I think. >> From what I understand, this change is originating in Debian (and thus >> propagating to Ubuntu). I believe the motivations are mostly ease of >> management of the build daemons -- only by doing this, for instance, can >> an entire architecture be properly isolated and self-contained. > > No. Very, very, far from it. > > The multiarch specs are created and maintained by a Ubuntu (and Debian) > developer, and the motivations are purely user-oriented - making it > possible, easy, and reliable for the user to do things that's currently > difficult, fragile, and error-prone. But for package maintainers, life > will certainly be no easier. > I don't mean the multiarch change, I mean the "build daemons will not allow have foreign multiarch build dependencies" change, which I guess isn't a change so much as it is a property of the new system. You're quite right that multiarch proper is for users; it's something I've been demanding for a while now. > Multiarch has no management implications whatsoever for the build > daemons. All official Debian architectures are perfectly self-contained > as they are. (Some of them currently contain cross-compiled pieces of > other architectures, which would go away, but that does not concern the > build daemons or the infrastructure.) > > Multiarch with foreign architecture build support would eliminate the self-containment, wouldn't it?

May 21, 2011

Re: Big ugly build changes might be needed for Debian/Ubuntu

by Ove Kåven

Den 19. mai 2011 20:58, skrev Scott Ritchie: > A side effect of this change, however, is the current build daemons ONLY > have packages for one architecture. This means that, at build time, you > won't be able to pull in the 32-bit packages on a 64-bit build daemon. Well, I don't see the problem. I've been more concerned that there's not yet any clean way to get the 64-bit package to depend on the 32-bit package, but for the build itself, it's quite trivial. (At least it was a year ago or so, and I can't imagine it has gotten worse since then.) > From what I understand, this change is originating in Debian (and thus > propagating to Ubuntu). I believe the motivations are mostly ease of > management of the build daemons -- only by doing this, for instance, can > an entire architecture be properly isolated and self-contained. No. Very, very, far from it. The multiarch specs are created and maintained by a Ubuntu (and Debian) developer, and the motivations are purely user-oriented - making it possible, easy, and reliable for the user to do things that's currently difficult, fragile, and error-prone. But for package maintainers, life will certainly be no easier. Multiarch has no management implications whatsoever for the build daemons. All official Debian architectures are perfectly self-contained as they are. (Some of them currently contain cross-compiled pieces of other architectures, which would go away, but that does not concern the build daemons or the infrastructure.)

May 21, 2011

hello

by chris.teague＠gmail.com

, i paid off all my debt http://tulsamarketplace.com/redirect.php?to=aHR0cDovL3d3dy5jbmJjNy5jb20vP3B…

May 20, 2011

Re: gdi32: Fetch CURVECAPS, LINECAPS and POLYGONALCAPS for an EMF from parent device.

by Alexandre Julliard

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > Alexandre Julliard <julliard(a)winehq.org> wrote: > >> > This follows the way how it's done for other device capabilities. >> >> It's not clear that this makes sense, a metafile is capable of doing all >> these things even if the device cannot. Have you verified this on a >> device that doesn't have all the caps? > > I'd expect you asked that question to the person adding these caps in > the first place. I just did what's already done with for textcaps and > rastercaps which is quite straightforward. Returning capabilities matching what metafiles can do makes sense to me, returning caps from an unrelated device doesn't. Of course Windows isn't known for doing things that make sense, but I'd still like to see a test case. -- Alexandre Julliard julliard(a)winehq.org

May 20, 2011

Re: gdi32: Fetch CURVECAPS, LINECAPS and POLYGONALCAPS for an EMF from parent device.

by Dmitry Timoshkov

Alexandre Julliard <julliard(a)winehq.org> wrote: > > This follows the way how it's done for other device capabilities. > > It's not clear that this makes sense, a metafile is capable of doing all > these things even if the device cannot. Have you verified this on a > device that doesn't have all the caps? I'd expect you asked that question to the person adding these caps in the first place. I just did what's already done with for textcaps and rastercaps which is quite straightforward. -- Dmitry.

May 20, 2011

Re: gdi32: Fetch CURVECAPS, LINECAPS and POLYGONALCAPS for an EMF from parent device.

by Alexandre Julliard

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > This follows the way how it's done for other device capabilities. It's not clear that this makes sense, a metafile is capable of doing all these things even if the device cannot. Have you verified this on a device that doesn't have all the caps? -- Alexandre Julliard julliard(a)winehq.org

May 20, 2011

Re: Big ugly build changes might be needed for Debian/Ubuntu

by Marcus Meissner

On Thu, May 19, 2011 at 11:58:47AM -0700, Scott Ritchie wrote: > So, multiarch is going along well in Debian/Ubuntu, and by next distro > release Wine should be building as a proper multiarch package. Previous > hacks like ia32-libs will be dead for good. > > > A side effect of this change, however, is the current build daemons ONLY > have packages for one architecture. This means that, at build time, you > won't be able to pull in the 32-bit packages on a 64-bit build daemon. > > > From what I understand, this change is originating in Debian (and thus > propagating to Ubuntu). I believe the motivations are mostly ease of > management of the build daemons -- only by doing this, for instance, can > an entire architecture be properly isolated and self-contained. > > > This means for bi-arch Wine, we will need to either: > - Make Wine support building in pure 64-bit mode and then calling the > 32-bit subsystem from a separate binary > - Make a very persuasive case that Wine isn't the only package that > needs to have multiple architectures at build time > > > Hopefully I'm overestimating the amount of work involved, as this sort > of architecture modularity might be a "nice to have" in Wine anyway. > > If none of this is doable by next release, we'll just have a 32-bit only > Wine, which isn't really any worse than the situation now. It is trivial. You basically build wine twice, once in the 32bit environment and once inthe 64bit one. You just need to take care you can install both. For openSUSE I create the binary packages: wine-32bit (i586 build) /usr/lib/libwine* /usr/lib/wine/* /usr/bin/wine /usr/bin/wine-preloader wine (both a i586 and x86_64 package) both requires the wine-32bit package Contain all other files not in wine-32bit. for 64bit also the runtime things /usr/lib64/libwine* /usr/lib64/wine/* /usr/bin/wine64 /usr/bin/wine64-preloader So basically depending on which architecture it is installed on, you get wine and wine-32bit on the final system ... Just on 64bit the wine package is 64bit. There is no need for 32bit and 64bit cross usage during build. Ciao, Marcus

May 19, 2011

Re: [3/3] d3d9: Try to get the INTZ test working on nvidia cards

by Henri Verbeet

I think I'd rather just change the way the test works.

May 19, 2011

Re: Big ugly build changes might be needed for Debian/Ubuntu

by Michael Stefaniuc

Scott Ritchie wrote: > So, multiarch is going along well in Debian/Ubuntu, and by next distro > release Wine should be building as a proper multiarch package. Previous > hacks like ia32-libs will be dead for good. > > > A side effect of this change, however, is the current build daemons ONLY > have packages for one architecture. This means that, at build time, you > won't be able to pull in the 32-bit packages on a 64-bit build daemon. > > > From what I understand, this change is originating in Debian (and thus > propagating to Ubuntu). I believe the motivations are mostly ease of > management of the build daemons -- only by doing this, for instance, can > an entire architecture be properly isolated and self-contained. > > > This means for bi-arch Wine, we will need to either: > - Make Wine support building in pure 64-bit mode and then calling the > 32-bit subsystem from a separate binary > - Make a very persuasive case that Wine isn't the only package that > needs to have multiple architectures at build time > > > Hopefully I'm overestimating the amount of work involved, as this sort > of architecture modularity might be a "nice to have" in Wine anyway. > > If none of this is doable by next release, we'll just have a 32-bit only > Wine, which isn't really any worse than the situation now. http://wiki.winehq.org/Wine64ForPackagers bye michael

May 19, 2011

Big ugly build changes might be needed for Debian/Ubuntu

by Scott Ritchie

So, multiarch is going along well in Debian/Ubuntu, and by next distro release Wine should be building as a proper multiarch package. Previous hacks like ia32-libs will be dead for good. A side effect of this change, however, is the current build daemons ONLY have packages for one architecture. This means that, at build time, you won't be able to pull in the 32-bit packages on a 64-bit build daemon. >From what I understand, this change is originating in Debian (and thus propagating to Ubuntu). I believe the motivations are mostly ease of management of the build daemons -- only by doing this, for instance, can an entire architecture be properly isolated and self-contained. This means for bi-arch Wine, we will need to either: - Make Wine support building in pure 64-bit mode and then calling the 32-bit subsystem from a separate binary - Make a very persuasive case that Wine isn't the only package that needs to have multiple architectures at build time Hopefully I'm overestimating the amount of work involved, as this sort of architecture modularity might be a "nice to have" in Wine anyway. If none of this is doable by next release, we'll just have a 32-bit only Wine, which isn't really any worse than the situation now. -Scott Ritchie

May 19, 2011

Re: [PATCH 7/7] msvcrt: Make functions working on FILE thread safe

by Alexandre Julliard

Piotr Caban <piotr(a)codeweavers.com> writes: > --- > dlls/msvcrt/file.c | 448 > ++++++++++++++++++++++++++++++++++++---------------- > 1 files changed, 314 insertions(+), 134 deletions(-) This one should probably be split. It also breaks badly on 64-bit: ../../../../wine/tools/runtest -q -P wine -M mshtml.dll -T ../../.. -p mshtml_test.exe.so ../../../../wine/dlls/mshtml/tests/activex.c && touch activex.ok err:seh:setup_exception stack overflow 2336 bytes in thread 0021 eip 00002b829f113a5d esp 0000000000140ce0 stack 0x140000-0x141000-0x240000 -- Alexandre Julliard julliard(a)winehq.org

May 19, 2011

Re: hostname: Add an implementation of the hostname display utility. (try 2)

by Alexandre Julliard

Andrew Nguyen <anguyen(a)codeweavers.com> writes: > diff --git a/programs/hostname/Makefile.in b/programs/hostname/Makefile.in > new file mode 100644 > index 0000000..300d727 > --- /dev/null > +++ b/programs/hostname/Makefile.in > @@ -0,0 +1,12 @@ > +EXTRADEFS = -DWINE_NO_UNICODE_MACROS > +MODULE = hostname.exe > +APPMODE = -mconsole -municode > +IMPORTS = ws2_32 You probably don't want to use winsock, at least not for Unix builds. -- Alexandre Julliard julliard(a)winehq.org

May 19, 2011

Re: [2/2] d3dx9: Implement D3DXLoadMeshHierarchy for basic mesh hierarchy.

by Joris Huizer

On 05/19/2011 06:34 PM, Dylan Smith wrote: > On Thu, May 19, 2011 at 6:16 AM, Joris Huizer<joris_huizer(a)yahoo.com> wrote: >> Hello, >> >> I found a possibly memory leak in the patch. >> In this function: >> >>> +static HRESULT parse_mesh(IDirectXFileData *filedata, struct mesh_data *mesh_data, DWORD provide_flags) >> These may leak the allocated blocks: >>> + >>> + mesh_data->vertices = HeapAlloc(GetProcessHeap(), 0, >>> + mesh_data->num_vertices * sizeof(*mesh_data->vertices)); >>> + mesh_data->num_tri_per_face = HeapAlloc(GetProcessHeap(), 0, >>> + mesh_data->num_poly_faces * sizeof(*mesh_data->num_tri_per_face)); >>> + mesh_data->indices = HeapAlloc(GetProcessHeap(), 0, >>> + (mesh_data->num_tri_faces + mesh_data->num_poly_faces * 2) * sizeof(*mesh_data->indices)); >>> + if (!mesh_data->vertices || !mesh_data->num_tri_per_face || !mesh_data->indices) >>> + return E_OUTOFMEMORY; >>> + > All the pointers are stored in mesh_data, which isn't lost. They are > freed at the end of load_skin_mesh_from_xof. > > +cleanup: > + if (FAILED(hr)) { > + if (d3dxmesh) IUnknown_Release(d3dxmesh); > + if (adjacency) ID3DXBuffer_Release(adjacency); > + } > + HeapFree(GetProcessHeap(), 0, mesh_data.vertices); > + HeapFree(GetProcessHeap(), 0, mesh_data.num_tri_per_face); > + HeapFree(GetProcessHeap(), 0, mesh_data.indices); > + return hr; > +} Thanks for clarification (Sorry for the noise)

May 19, 2011

Re: [2/2] d3dx9: Implement D3DXLoadMeshHierarchy for basic mesh hierarchy.

by Dylan Smith

On Thu, May 19, 2011 at 6:16 AM, Joris Huizer <joris_huizer(a)yahoo.com> wrote: > > Hello, > > I found a possibly memory leak in the patch. > In this function: > >> +static HRESULT parse_mesh(IDirectXFileData *filedata, struct mesh_data *mesh_data, DWORD provide_flags) > > These may leak the allocated blocks: >> >> + >> + mesh_data->vertices = HeapAlloc(GetProcessHeap(), 0, >> + mesh_data->num_vertices * sizeof(*mesh_data->vertices)); >> + mesh_data->num_tri_per_face = HeapAlloc(GetProcessHeap(), 0, >> + mesh_data->num_poly_faces * sizeof(*mesh_data->num_tri_per_face)); >> + mesh_data->indices = HeapAlloc(GetProcessHeap(), 0, >> + (mesh_data->num_tri_faces + mesh_data->num_poly_faces * 2) * sizeof(*mesh_data->indices)); >> + if (!mesh_data->vertices || !mesh_data->num_tri_per_face || !mesh_data->indices) >> + return E_OUTOFMEMORY; >> + > All the pointers are stored in mesh_data, which isn't lost. They are freed at the end of load_skin_mesh_from_xof. +cleanup: + if (FAILED(hr)) { + if (d3dxmesh) IUnknown_Release(d3dxmesh); + if (adjacency) ID3DXBuffer_Release(adjacency); + } + HeapFree(GetProcessHeap(), 0, mesh_data.vertices); + HeapFree(GetProcessHeap(), 0, mesh_data.num_tri_per_face); + HeapFree(GetProcessHeap(), 0, mesh_data.indices); + return hr; +}

May 19, 2011

Re: oleaut32/tests: remove macros from test_VarCyCmpR8

by Nikolay Sivov

On Thu, May 19, 2011 at 11:30 AM, Austin English <austinenglish(a)gmail.com> wrote: > -- > -Austin > > > > This test should use a table+loop I think.

May 19, 2011

Re: [bugzilla] Fix Bug 16822 - Mass Effect: Mouse input does not work (almost completely)

by Vitaliy Margolen

On 05/18/2011 12:03 PM, siro wrote: > This patch don't clear the dinput queue in IDirectInputDevice2WImpl_Acquire Clearly this is wrong. Of course you can prove otherwise with some tests. You can also test with native dinput to see if it has the same issue or not. Vitaliy.

May 19, 2011

Re: gdi32: Add support in DeviceCaps CURVECAPS, LINECAPS, POLYGONALCAPS for EnhMetafiles

by Dmitry Timoshkov

Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> wrote: > gdi32: Add support in DeviceCaps CURVECAPS, LINECAPS, POLYGONALCAPS > for EnhMetafiles I'd guess they should be cached and returned for the parent device instead. -- Dmitry.

May 19, 2011

Re: [2/2] d3dx9: Implement D3DXLoadMeshHierarchy for basic mesh hierarchy.

by Joris Huizer

Hello, I found a possibly memory leak in the patch. In this function: > +static HRESULT parse_mesh(IDirectXFileData *filedata, struct mesh_data *mesh_data, DWORD provide_flags) These may leak the allocated blocks: > + > + mesh_data->vertices = HeapAlloc(GetProcessHeap(), 0, > + mesh_data->num_vertices * sizeof(*mesh_data->vertices)); > + mesh_data->num_tri_per_face = HeapAlloc(GetProcessHeap(), 0, > + mesh_data->num_poly_faces * sizeof(*mesh_data->num_tri_per_face)); > + mesh_data->indices = HeapAlloc(GetProcessHeap(), 0, > + (mesh_data->num_tri_faces + mesh_data->num_poly_faces * 2) * sizeof(*mesh_data->indices)); > + if (!mesh_data->vertices || !mesh_data->num_tri_per_face || !mesh_data->indices) > + return E_OUTOFMEMORY; > + HTH, Joris

May 19, 2011

Re: [PATCH] winex11: Make GetCursorPos respect clipping and virtual screen size.

by Lauri Kenttä

Alexandre Julliard writes: > Lauri Kenttä <lauri.kentta(a)gmail.com> writes: >> With my touchpad the results are much more interesting. > > Most likely because XInput raw events don't take acceleration into > account. I'm not sure there's a whole lot we can do about that. I've disabled acceleration, and the problem persists. Unless there's some very well hidden option that doesn't react to xset or xinput. Alexandre Julliard writes: > I think enabling clipping for these apps is a better solution. But clearly that's not an option if the user has a tablet (or other absolute pointing device) and a touchpad (at least my touchpad). Even if the touchpad problem could be avoided by disabling acceleration, that's a pain for the user and makes the touchpad behave differently from what the user is used to. What about adding this feature along with a registry key? The implementation is only a few lines and works great in my case. Of course, this is not an essential feature, as the app works fine in real full-screen mode. So I won't press this any more. -- Lauri Kenttä

May 19, 2011

Re: [PATCH] dinput: Move the c_df* data format constants to the dxguid static library, mark as hidden

by Marcus Meissner

On Wed, May 18, 2011 at 10:57:59AM +0200, Alexandre Julliard wrote: > Marcus Meissner <meissner(a)suse.de> writes: > > > --- > > dlls/dinput/Makefile.in | 1 - > > dlls/dinput/data_formats.c | 566 -------------------------------------------- > > dlls/dxguid/dxguid.c | 545 ++++++++++++++++++++++++++++++++++++++++++ > > 3 files changed, 545 insertions(+), 567 deletions(-) > > delete mode 100644 dlls/dinput/data_formats.c > > What makes you think they should be in dxguid? Wild guessing and web page reading. However, strings on some .lib files in DX SDK June 2010 finds it in dinput8.lib. So please disregard. Ciao, Marcus

May 19, 2011

Re: tests/msvcrt: add test for _splitpath function

by Eric Pouech

adding c:test.txt (and friends) as input should be also interesting UNC paths should also be tested A+ + } Paths[] = { {"C:/directory/test.txt", "C:", "/directory/", > "test", ".txt"}, > + {"C:\\directory\\test.txt", "C:", "\\directory\\", > "test", ".txt"}, > + {"C:\\directory\\test", "C:", "\\directory\\", "test", > ""}, > + {"C:\\directory\\.txt", "C:", "\\directory\\", "", > ".txt"}, > + {"C:\\directory\\", "C:", "\\directory\\", "", ""}, > + {"C:\\test.txt", "C:", "\\", "test", ".txt"}, > + {"C:\\test", "C:", "\\", "test", ""}, > + {"C:\\.txt", "C:", "\\", "", ".txt"}, > + {"C:\\", "C:", "\\", "", ""}, > + {"\\directory\\test.txt", "", "\\directory\\", "test", > ".txt"}, > + {"\\directory\\test", "", "\\directory\\", "test", ""}, > + {"\\directory\\.txt", "", "\\directory\\", "", ".txt"}, > + {"\\directory\\", "", "\\directory\\", "", ""}, > + {"\\test.txt", "", "\\", "test", ".txt"}, > + {"\\test", "", "\\", "test", ""}, > + {"\\.txt", "", "\\", "", ".txt"}, > + {"\\", "", "\\", "", ""}, > + {"test.txt", "", "", "test", ".txt"}, > + {"test", "", "", "test", ""}, > + {".txt", "", "", "", ".txt"}, > + {"", "", "", "", ""}}; > -- > Eric Pouech >

May 19, 2011

Re: d3d: avoid implicit conversion warnings from identical enumeration types D3DPOOL and WINED3DPOOL (Clang) - please discard patch

by Frédéric Delanoy

Patch can be discarded. Spreading casts all over is probably a bad thing.

May 19, 2011

Re: d3d8: Silence implicit enum conversion warnings for D3DFormat (clang) - please discard

by Frédéric Delanoy

Please discard this patch. An updated version follows.

May 18, 2011

Re: [PATCH] winex11: Make GetCursorPos respect clipping and virtual screen size.

by Alexandre Julliard

Lauri Kenttä <lauri.kentta(a)gmail.com> writes: > So is it then possible to clip the cursor to the virtual desktop in > the wineserver, or do you see some apparent problem there? I would be > surprised if this was the only application to choke on strange mouse > coordinates. It's possible, but it's not clear that it's desirable. Pretending that the mouse is still inside the desktop when it has left can lead to strange behaviors too. I think enabling clipping for these apps is a better solution. > With my touchpad the results are much more interesting. The cursor > moves slowly (around 1/3 of normal speed) but warps once per second to > where it really should be. When the cursor stops, the game starts > polling with GetCursorPos and the cursor warps again to the correct > position. The TRACE in GetCursorPos shows that "server pos" follows > the slower movement and is different from what the function returns. > > Maybe I should file a bug about that? Most likely because XInput raw events don't take acceleration into account. I'm not sure there's a whole lot we can do about that. -- Alexandre Julliard julliard(a)winehq.org

May 18, 2011

Re: [PATCH] winex11: Make GetCursorPos respect clipping and virtual screen size.

by Lauri Kenttä

On Wed, 18 May 2011 12:22:57 +0200, Alexandre Julliard wrote: > > The clip rectangle is only valid while the pointer is grabbed, in > which > case the cursor pos will be clipped already. That's true. I only tested the other case. > So for the case where the pointer is not grabbed (for instance if you > don't have XInput) it would have to go through the wineserver. So is it then possible to clip the cursor to the virtual desktop in the wineserver, or do you see some apparent problem there? I would be surprised if this was the only application to choke on strange mouse coordinates. > Have you tried the new desktop clipping mode? Sorry, I missed that because of missing XI2. That indeed fixes the clipping (but only when the window is active) and works fine with an external mouse. With my touchpad the results are much more interesting. The cursor moves slowly (around 1/3 of normal speed) but warps once per second to where it really should be. When the cursor stops, the game starts polling with GetCursorPos and the cursor warps again to the correct position. The TRACE in GetCursorPos shows that "server pos" follows the slower movement and is different from what the function returns. Maybe I should file a bug about that? -- Lauri Kenttä

May 18, 2011