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

Re: user32: fix potential crash when loading unsupported files (with tests)
by Alexandre Julliard 03 Aug '11

03 Aug '11

Vitaly Perov <vitperov(a)etersoft.ru> writes: > @@ -1002,6 +1002,54 @@ static unsigned char gif4pixel[42] = { > 0x02,0x00,0x00,0x02,0x03,0x14,0x16,0x05,0x00,0x3b > }; > > +/* Corrupted images */ > +/* 1x1 pixel corrupted bmp */ > +static unsigned char corruptedbmpimage[38] = { > +0x42,0x4d,0x26,0x00,0x01,0x00,0x00,0x00,0x00,0x00,0x22,0x00,0x00,0x00,0x0c,0x00, > +0x00,0x00,0x01,0x00,0x01,0x00,0x01,0x00,0x01,0x00,0xff,0xff,0xff,0x00,0x55,0x55, > +0x55,0x00,0x00,0x00,0x00,0x00 > +}; You should modify the data in place like other tests already do. There's no reason to duplicate it all. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: oleaut32: Skip tests if interface is not accessable
by Alexandre Julliard 03 Aug '11

03 Aug '11

André Hentschel <nerv(a)dawncrow.de> writes: > The hole test_ function depends on succeeding here (more than i thought first) Actually you'd want to skip even more than that, the real failure is earlier. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: wined3d: Update GLSL spam filter for recent fglrx versions.
by Stefan Dösinger 03 Aug '11

03 Aug '11

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Am 03.08.2011 um 17:31 schrieb Matteo Bruni: > Users tend to get confused by those messages on the terminal. Since we > already have code to filter irrelevant GLSL compiler messages, let's > update it. I think the plan a while ago was to remove the spam filter and not print driver messages as FIXMEs(just as WARNs or TRACEs) when the compile was successful. -----BEGIN PGP SIGNATURE----- Version: GnuPG/MacGPG2 v2.0.17 (Darwin) iQIcBAEBAgAGBQJOOWwJAAoJEN0/YqbEcdMwRF8QAIMOF0XTEsl8+DSmBsuDzIX4 dSjR96gAtHlQGaq11P4Auv2cX1slykcLQrfJWy4zPOCty97psJHQSzSmiEqiRyK2 7Eqa8iKvlqsXJhBQmkOWZM3XBIUIdsfsZrT+jo7jDOjnwF0CNZ7DP6ch3q01gWOw UWlNukHd2qZ29Awq5zuf6T9O45/4W1i85XPrWzLlzhAGeQ7dhTHzddpLfjO1Tb5B GxcmeSmyBP2RMEgjl6eXGk2mgz9OmjSPJrFTgaav0l5mCvhO/HcRczqhvtV08DDd gO3WCvxRywOwL78cOrudM1Aqd8o33etYK0hQGT3zw1kreXbFL7rPS/APs38ZF/53 m+3pRg2JtJiQQmWSjM4ou4H15H72Z8zoXMxAvvguL8YveuFNu2Q18SfLkNc9Qhwv sMtRkxM55ap0D8RAYw60eWORlPYBwMD8zF1lKz3xImIemoMflo7CSrvrWv7KUXJE 0IPOGpJyCtTIrShfh6Hvj53FwWb16FfphETnvwFffY7k42mkbwU+Sa76aZ9Wklnl DODFQ2VLgMtCyO/3he2b4rc2KAlc567HKJI2ND8hYXvQH8M2q6ebps5LrphglPnb 87e3jdBcNBISJ0t2/jamNRFj+uPY2UNoga+5xKwnC1uqKe1m9mmzfvm25SIAMBbE bihjIFyAg+Bv5Z6Lrzag =Qbf6 -----END PGP SIGNATURE-----

2 1

Re: [PATCH 1/3] winmm: Memory related fixes.
by Alexandre Julliard 03 Aug '11

03 Aug '11

Joerg-Cyril.Hoehle(a)t-systems.com writes: > Hi, > > IMHO, this does not need further splitting. We have had that big 100KB patch which introduced > the bugs and don't need N patches of a fraction that size each to fix them one by one. > Instead, we need to find and remove them quickly so users can be happy again with sound in Wine. They would go in more quickly if you would simply follow the rules. Independent changes should be separate patches. There are many good reasons for this, that I'm sure you can figure out. -- Alexandre Julliard julliard(a)winehq.org

1 0

mmdevapi tests
by Joerg-Cyril.Hoehle＠t-systems.com 03 Aug '11

03 Aug '11

Hi, I've a couple of mmdevapi tests that I could put into submittable form soon, however there's one big problem: There are three different backends (ALSA, OSS, MacOS) and each one needs a different set of todo_wine. How to deal with that? In the past, when I worked alone (e.g. on the MCI), I simply submitted problematic tests only after I had patches that prevented Wine from crashing or possibly even implemented the functionality. Now of course, I could open an issue (one per backend?), append the patch and say "Wine must pass this". However that won't get as much test coverage (by neither native machines, nor Wine boxes). What do you suggest? Jörg Höhle

2 1

Re: [PATCH 1/2] comctl32: Fix to ComboBoxEx processing of WM_ENABLE message
by Nikolay Sivov 03 Aug '11

03 Aug '11

On 8/3/2011 02:49, Richard Bradbrook wrote: > comctl32: Fix to ComboBoxEx processing of WM_ENABLE message > This fixes some of the issues mentioned in bug 27920 Hi, Richard. Few comments: > +static LRESULT COMBOEX_Enable (COMBOEX_INFO *infoPtr, BOOL enable) > +{ > + TRACE("hwnd=%p, enable=%s\n", infoPtr->hwndSelf, enable ? "TRUE":"FALSE"); No need for a trace here, all messages are already traced at procedure level. Also please add tests that show disabled edit box and disabled combo after switching ComboEx state.

1 0

Re: advapi32: Add stub TraceMessage, TraceMessageVa
by Hans Leidekker 03 Aug '11

03 Aug '11

On Wed, 2011-08-03 at 13:01 +1000, Alistair Leslie-Hughes wrote: -@ stub TraceMessage -@ stub TraceMessageVa +@ varargs TraceMessage(long long ptr long) +@ stdcall TraceMessageVa(long long ptr long ptr) These functions take a TRACEHANDLE as the first argument which is an int64.

1 0

re: GSoC: dinput8 Action Mapping
by Dan Kegel 02 Aug '11

02 Aug '11

Lucas asked: > Should I correct the tabs and align the > parameters in the function prototype when I implement them? If you do a significant rewrite of a function, or are implementing it for the first time, I think it's ok to use more standard code style in your new code, or to update the style of adjacent old code (like a prototype) related to your change.

7 28

Re: [PATCH 1/3] shell32/tests: Add verification of known folders parsing names.
by Alexandre Julliard 02 Aug '11

02 Aug '11

Mariusz Pluciński <vshader(a)gmail.com> writes: > + KNOWN_FOLDER(FOLDERID_AddNewPrograms, NO_CSIDL, "AddNewProgramsFolder", KF_CATEGORY_VIRTUAL, GUID_NULL, NULL, "::{21EC2020-3AEA-1069-A2DD-08002B30309D}\\::{15eae92e-f17a-4431-9f28-805e482dafd4}\0" > + "shell:::{26EE0668-A00A-44D7-9371-BEB064C98683}\\0\\::{15eae92e-f17a-4431-9f28-805e482dafd4}\0\0" /* 6.1 */), This is now way past a reasonable line length. Please format this differently. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH 4/6] shell32: Implement SHCreateDefaultContextMenu and CDefFolderMenu_Create2 (try 2)
by Alexandre Julliard 02 Aug '11

02 Aug '11

Jay Yang <jkelleyy(a)gmail.com> writes: > Added TRACE statements and removed some stray comments > --- > dlls/shell32/Makefile.in | 1 + > dlls/shell32/composite_cmenu.c | 275 ++++++++++++++++++++++++++++++++++++++++ Please put that in some existing files, there are already way too many files in shell32. -- Alexandre Julliard julliard(a)winehq.org

1 0