November 2011 - Wine-devel - list.winehq.org

Re: [PATCH 1/2] jscript: Added tests of unary '+' on not existent property
by Marvin 22 Nov '11

22 Nov '11

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 http://testbot.winehq.org/JobDetails.pl?Key=15587 Your paranoid android. === WNT4WSSP6 (32 bit) === No test summary line found === W2KPROSP4 (32 bit) === No test summary line found === WXPPROSP3 (32 bit) === No test summary line found === W2K3R2SESP2 (32 bit) === No test summary line found === WVISTAADM (32 bit) === No test summary line found === W2K8SE (32 bit) === No test summary line found === W7PRO (32 bit) === No test summary line found === W7PROX64 (32 bit) === No test summary line found === TEST64_W7SP1 (32 bit) === No test summary line found === W7PROX64 (64 bit) === No test summary line found === TEST64_W7SP1 (64 bit) === No test summary line found

1 0

Re: [4/5] ntdll: Remove an invalid test.
by Alexandre Julliard 22 Nov '11

22 Nov '11

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > The test initializers that call GetProcAddress() do write to the same section > where the static buffer resides, which leads to the image section protection > change. That doesn't make the test invalid. -- Alexandre Julliard julliard(a)winehq.org

2 3

Re: [PATCH] dxgi: 0 / 0 means 0 for a DXGI_RATIONAL
by Henri Verbeet 22 Nov '11

22 Nov '11

On 22 November 2011 00:42, Eduard - Gabriel Munteanu <eduard.munteanu(a)linux360.ro> wrote: > This fixes a division by zero generated in Unigine Tropics in D3D10 > mode. It also adds a helper to handle such conversions, which takes the > form of a macro so we don't need to worry about the result's type. > I'd prefer an actual function. A small test for this probably wouldn't hurt either.

1 0

Re: winefile: Make it possible to translate the disk size units.
by Francois Gouget 22 Nov '11

22 Nov '11

On Sat, 19 Nov 2011, Francois Gouget wrote: [...] > - static const WCHAR sFmtGB[] = {'%', '.', '1', 'f', ' ', 'G', 'B', '\0'}; > - static const WCHAR sFmtMB[] = {'%', '.', '1', 'f', ' ', 'M', 'B', '\0'}; > - static const WCHAR sFmtkB[] = {'%', '.', '1', 'f', ' ', 'k', 'B', '\0'}; To be totally correct these should actually be 'GiB', 'MiB' and 'kiB' (the code divides by powers of 1024). However that may be too geeky for regular users so I'm not sure we want to do that. But now is probably a good time to debate this. Here are two reference points: * On Windows winefile and explorer use 'KB' & co for the disk space although they divide by '1024'. * Nautilus uses 'kiB' & co. Another point is that we always display the exact byte count for file sizes instead of converting it to 'KB', 'MB', etc. That actually matches winfile but seems pretty cubersome. On Windows Explorer does show file sizes as 'KB', 'MB', etc. so that's what users would expect and I think it's a place where we could follow suite. -- Francois Gouget <fgouget(a)free.fr> http://fgouget.free.fr/ If you think the whole world revolves around you, quit staring at the GPS display while driving.

3 2

regression bisected to commit: gdi32: Implement nulldrv_StretchDIBits using the PutImage gdi driver function.
by Wolfgang Walter 21 Nov '11

21 Nov '11

Hello, I just tested version 1.3.33. With this version an application we use shows its icons with wrong colors. The background of the icons is wrong, too. I bisected it to commit c9a7bb715d2db1512db30deb11e4676e76791a07. I then checked that by replacing nulldrv_StretchDIBits in 1.3.33 with its older implementation. c9a7bb715d2db1512db30deb11e4676e76791a07 is the first bad commit commit c9a7bb715d2db1512db30deb11e4676e76791a07 Author: Huw Davies <huw(a)codeweavers.com> Date: Mon Oct 17 15:46:07 2011 +0100 gdi32: Implement nulldrv_StretchDIBits using the PutImage gdi driver function. :040000 040000 e4cfeee3adcb0fe2c0502909c87b2740a8de08e5 00851b246b095bdf4cafd3e6f9965c811d1c87a4 M dlls Regards, -- Wolfgang Walter Studentenwerk München Anstalt des öffentlichen Rechts Abteilungsleiter IT Leopoldstraße 15 80802 München

2 1

Question about messages and the windows these need.
by Joerg-Cyril.Hoehle＠t-systems.com 21 Nov '11

21 Nov '11

Hi, Alexandre Julliard wrote in http://www.winehq.org/pipermail/wine-devel/2011-July/090969.html >If you are going to wait for a reply it would be better to create a >HWND_MESSAGE window and do a SendMessage to it, then you get all the >synchronization for free. MMSYSTEM_MidiStream_Player needs a patch to fix bug #28388. Such a patch will likely introduce sending Suspend & Resume messages to the MIDI player (to be able to wait for the resume signal). So I believe this same advice applies. MMSYSTEM_MidiStream_PostMessage already contains a PostThreadMessage + MsgWait pair that looks like it should be replaced, correct? OTOH, I'm unsettled whether MidiStreamSuspend & -Restart should wait for a reply to the message, so PostThreadMessage may be enough. Furthermore, I'm unsure about the implications of adding a dummy window and message queue. The MSDN blogs contain repeated warnings about the obligation to pump the message queue and GUI apps hanging when some random app fails to do so in all situations. Won't that window receive all kinds of broadcast messages? No, says MSDN http://msdn.microsoft.com/en-us/library/windows/desktop/ms632599%28v=vs.85%… The MIDI sequencer uses simplistic timing based on GetTickCount. Won't that allow arbitrary foreign window callbacks to disturb the timing of the MIDI sequencer? Instead of inventing message names like #define WINE_MSM_STOP (WM_USER+1) Why not re-use message names like MCI_STOP even if it's not in MCI format? (sounds like a hack...) Thank you for your help, Jörg Höhle

1 0

Re: [PATCH 1/1] user32: Fix 64-bit 'msg' test hang in test_timers.
by Alexandre Julliard 21 Nov '11

21 Nov '11

"Erich E. Hoover" <ehoover(a)mines.edu> writes: > Real Name: > Erich Hoover > > Description: > The TestBot ran into a hanging issue (noticed with, but unrelated > to, my previous user32 patch) and I took some time today to track the > problem down. Currently, on the 64-bit test system the test_timers() > tests hang waiting for PeekMessage() to clear the message queue. The > attached patch kills the timer just prior to the PeekMessage() call, > thereby keeping the test from hanging indefinitely. > > Changelog: > user32: Fix 64-bit 'msg' test hang in test_timers. That defeats the purpose of the test. -- Alexandre Julliard julliard(a)winehq.org

2 1

Re: [1/4] msxml3: Added common ISupportErrorInfo implementation
by Jacek Caban 21 Nov '11

21 Nov '11

Hi Nikolay, On 11/21/11 08:24, Nikolay Sivov wrote: > Added common ISupportErrorInfo implementation +/* common ISupportErrorInfo implementation */ +typedef struct { + ISupportErrorInfo ISupportErrorInfo_iface; + LONG ref; + + const tid_t* iids; +} SupportErrorInfo; + +static inline SupportErrorInfo *impl_from_ISupportErrorInfo(ISupportErrorInfo *iface) +{ + return CONTAINING_RECORD(iface, SupportErrorInfo, ISupportErrorInfo_iface); +} + +static HRESULT WINAPI SupportErrorInfo_QueryInterface(ISupportErrorInfo *iface, REFIID riid, void **obj) +{ + SupportErrorInfo *This = impl_from_ISupportErrorInfo(iface); + TRACE("(%p)->(%s %p)\n", This, debugstr_guid(riid), obj); + + *obj = NULL; + + if (IsEqualGUID(riid,&IID_IUnknown) || IsEqualGUID(riid,&IID_ISupportErrorInfo)) { + *obj = iface; + ISupportErrorInfo_AddRef(iface); + return S_OK; + } + + return E_NOINTERFACE; +} This doesn't seem to be the right approach. You need to return here interfaces of an object that inherits your ISupportErrorInfo implementation - that's the COM rule. Also you probably don't want to create new SupportErrorInfo instance on each QI call. I'd suggest DispatchEx-like approach in this case (or even abuse DispatchEx object and add ISupportErrorInfo there). Jacek

2 5

Re: [PATCH 1/2] jscript: Use bytecode for binary negation implementation
by Marvin 21 Nov '11

21 Nov '11

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 http://testbot.winehq.org/JobDetails.pl?Key=15559 Your paranoid android. === WNT4WSSP6 (32 bit) === No test summary line found === W2KPROSP4 (32 bit) === No test summary line found === WXPPROSP3 (32 bit) === No test summary line found === W2K3R2SESP2 (32 bit) === No test summary line found === WVISTAADM (32 bit) === No test summary line found === W2K8SE (32 bit) === No test summary line found === W7PRO (32 bit) === No test summary line found === W7PROX64 (32 bit) === No test summary line found === TEST64_W7SP1 (32 bit) === No test summary line found === W7PROX64 (64 bit) === No test summary line found === TEST64_W7SP1 (64 bit) === No test summary line found

1 0

Re: qedit/tests: Add SampleGrabber COM aggregation test.
by Marvin 21 Nov '11

21 Nov '11

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 http://testbot.winehq.org/JobDetails.pl?Key=15558 Your paranoid android. === WNT4WSSP6 (32 bit mediadet) === mediadet.c:147: Test failed: CoCreateInstance failed with 80040154 mediadet.c:148: Test failed: pM is NULL mediadet: unhandled exception c0000005 at 004015A2 === W2KPROSP4 (32 bit mediadet) === mediadet.c:147: Test failed: CoCreateInstance failed with 80040154 mediadet.c:148: Test failed: pM is NULL mediadet: unhandled exception c0000005 at 004015A2 === W2K8SE (32 bit mediadet) === mediadet.c:147: Test failed: CoCreateInstance failed with 80040154 mediadet.c:148: Test failed: pM is NULL mediadet: unhandled exception c0000005 at 004015A2

2 1