January 2007 - Wine-devel - list.winehq.org

Re: [2/2] shell32: Add proper support for SHGetFileInfo(SHGFI_ICONLOCATION | SHGFI_USEFILEATTRIBUTES).
by Frank Richter 22 Jan '07

22 Jan '07

On 18.01.2007 16:23, Francois Gouget wrote: > + if (dwFileAttributes & FILE_ATTRIBUTE_DIRECTORY) > + { > + lstrcpyW(psfi->szDisplayName, swShell32Name); > + psfi->iIcon = -IDI_SHELL_FOLDER; > + } At least on Windows, folders have a file type, too - see HKCR\Folder. So maybe use 'HCR_GetDefaultIconW(L"Folder", ...)' when the directory attribute is set, and 'HCR_GetDefaultIconW(sTemp, ...)' otherwise? -f.r.

2 1

Re: server: Reverse return value and status in token_access_check to be consistent.
by Robert Shearman 22 Jan '07

22 Jan '07

Vitaliy Margolen wrote: > Change Log: > - server: Reverse return value and status in token_access_check to be consistent. > - server: Do not set successful status by default. Set it explicitly in each case. > - advapi32: Add few tests. > > It is some-what confusing to have token_access_check return 2 statuses. And > even more so, when it doesn't match with the native. Usually function returns > a status if call to it succeeded (all parameters are valid, caller allowed > to perform requested operation, etc). And extra parameters would indicate > if some check(s) succeeded or not. > I agree with the spirit of this patch, but it is not clear which part of this patch fixes the new tests. So please split this patch up into tests, cleanups and bug fixes. > @@ -866,24 +862,17 @@ static unsigned int token_access_check( > ace = ace_next( ace ); > } > > + done: > The indentation of this label is incorrect. -- Rob Shearman

1 0

Re: user32: Some apps pass a color bitmap as a mask to CreateIconIndirect, convert it to b/w
by Kirill K. Smirnov 22 Jan '07

22 Jan '07

Hi, these patches http://www.winehq.org/pipermail/wine-patches/2007-January/035014.html http://www.winehq.org/pipermail/wine-patches/2007-January/035052.html totally break qip IM icons. (www.qip.ru) I suspect the reason is hidden in GetDIBits(...) - resulting AND mask is not correct. XOR mask is correct. BTW, these patches heal systray taskmgr icon! -- Kirill

2 1

Re: [PATCH 1/2] ntdll: Implement RtlDecodePointer and RtlEncodePointer
by Alexandre Julliard 22 Jan '07

22 Jan '07

Mike McCormack <mike(a)codeweavers.com> writes: > + /* there's no ntdll.RtlInterlockedCompareExchangePointer() ... */ > + RtlEnterCriticalSection( &pointer_obfuscator_cs ); No, but we have interlocked_cmpxchg_ptr() in libwine... -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH 3/8] server: export few functions we will need for security descriptor checks.
by Dmitry Timoshkov 22 Jan '07

22 Jan '07

"Vitaliy Margolen" <wine-patches(a)kievinfo.com> wrote: > +++ b/server/security.h > @@ -41,10 +41,16 @@ extern const LUID SeCreateGlobalPrivileg > > extern const PSID security_interactive_sid; > > +extern void map_generic_mask(unsigned int *mask, const GENERIC_MAPPING *mapping); > +++ b/server/token.c > /* maps from generic rights to specific rights as given by a mapping */ > -static inline void map_generic_mask(unsigned int *mask, const GENERIC_MAPPING *mapping) > +inline void map_generic_mask(unsigned int *mask, const GENERIC_MAPPING *mapping) 'inline' above probably should be removed. -- Dmitry.

1 0

Re: user32: Restore old WM_SIZE handler in mdi child
by Peter Oberndorfer 22 Jan '07

22 Jan '07

On Monday 08 January 2007 11:31, Dmitry Timoshkov wrote: > Hello, Hi, this patch seems to cause http://bugs.winehq.org/show_bug.cgi?id=7190 Each time a MDI child window gets a WM_SIZE the left most menu entry disapperars. The problem is that MDI_RestoreFrameMenu is called for each WM_SIZE message, that does not maximize the window, instead of only beeing called when window goes from maximized to normal And MDI_RestoreFrameMenu removes the left most menu entry, in belief is the document icon of a maximized child. What is the reason for restoring the old code? Does the new code break some apps? The attached patch, which reverts a part of the revert "fixes" the problem for my test app(ollydbg) But i guess it is totally wrong :-) > > this patch restores (almost completely) an old WM_SIZE handler in mdi child, > including the comment /* do not change */. > > Changelog: > user32: Restore old WM_SIZE handler in mdi child. > Greetings Peter

2 1

Re: comctl32/tests/updown failures on xp?
by Dan Kegel 21 Jan '07

21 Jan '07

On 1/19/07, James Hawkins <truiken(a)gmail.com> wrote: > It sounds like your build environment isn't set up correctly. The > easiest way to work on and compile the tests is to run > wine/tools/winapi/msvcmaker under cygwin in Windows. This will create > several project files, one for each dll etc, and a winetest project > that can be opened with Visual C++ (and converted to a solution). Nah, I prefer bare metal commandline. I like knowing what's going on under the hood. > No matter which method you use to compile the tests, you should have the > latest Windows SDK installed (not the Platform SDK, which the Windows > SDK is replacing). OK, I'll try that... > If you run 'make test', you shouldn't see the errors > (because they're wrapped by todo_wine). Whoops, forgot about that! When running without 'make test', you have to set the environment variable WINETEST_PLATFORM=wine on wine for the todo_wine to work. Thanks! - Dan

2 1

WINSPOOL_EnumPrinters Found 0 printers on system with 1 parallel CUPS printer
by Michael [Plouj] Ploujnikov 21 Jan '07

21 Jan '07

Hi, I'm sending this message to both wine-users and wine-devel in hopes of getting help from someone with knowledge of Wine's internals. I've had this problem for a while, but only recently have I realized that I better fix it soon! I compile my own Wine out of git. Currently, I'm running wine-0.9.29-g5441536 but this issue appeared long before that. Moreover, I stopped being able to print between Wine updates, so I don't think it is anything to do with changes in Wine's code. The simplest way to reproduce my problem is try printing in notepad. First, I open notepad: $ WINEDEBUG=+winspool wine notepad trace:winspool:WINSPOOL_EnumPrinters Found 0 printers trace:winspool:WINSPOOL_EnumPrinters Found 0 printers Next, I type some text. Then I go to "File -> Printer Setup". Wine outputs this to the console: trace:winspool:WINSPOOL_EnumPrinters Found 0 printers fixme:winspool:WINSPOOL_EnumPrinters We don't handle PRINTER_ENUM_CONNECTIONS and I see a message box saying: "Before you can perform printer-related tasks such as page setup or printing a document, you need to install a printer. Please install one and retry." I've tried deleting and re-creating my ~/.wine multiple times. I know that wine is compiled with CUPS support $ ~/apps/wine-git/configure --x-libraries=/emul/linux/x86/usr/lib/ LDFLAGS="-L/emul/linux/x86/usr/lib -L/emul/linux/x86/lib" --without-libcapi20 --without-libhal --without-libicu --without-libldap --without-libsane > /tmp/wine-config $ grep -i cups /tmp/wine-config checking cups/cups.h usability... yes checking cups/cups.h presence... yes checking for cups/cups.h... yes checking for -lcups soname... libcups.so I have a local CUPS printer installed on my system: $ lpstat -t scheduler is running system default destination: HPDeskjet710c device for HPDeskjet710c: parallel:/dev/lp0 HPDeskjet710c accepting requests since Thu Jan 18 20:26:33 2007 printer HPDeskjet710c is idle. enabled since Thu Jan 18 20:26:33 2007 and I am able to print to it from other Linux applications. I considered tracing the winespool code ot see where it fails to find my printer, but it looks way too cryptic for my uninitiated eyes. Any help would be appreciated. -- Libre Software: http://www.gnu.org/philosophy/free-sw.html

3 2

Re: dbghelp: fix debuglink crash, accessing memory after munmap
by Eric Pouech 21 Jan '07

21 Jan '07

Peter Oberndorfer a écrit : > Winedbg would crash on my system when trying to set a breakpoint (loading the debug info from a .debug file) > The problem is that hash_table_elt adds the symbols of the .debug file to the hashtable of the parent file. > But at the time elf_new_public_symbols is run, the sections of the .debug file are already unmapped. > In addition to that the symbols in .debug would be added 2 times (.so and .debug file) > > Changelog: > dbghelp: fix debuglink crash, accessing memory after munmap > I don't think it's the right fix IMO, we shouldn't call elf_new_public_symbols while processing the .gnu_link target, but in the parent call to elf_load_debug_from_map can you provide more information on the crash itself (which reference is lost to an unmapped section) A+

2 1

3DMark2000 regression
by Stefan Dösinger 21 Jan '07

21 Jan '07

Hi Dmitry, Your "Do not play with WS_VISIBLE" patch(5ddac44dc2a6637613407fa900d3fafac37e70d8) in wined3d causes a regression with 3DMark2000. The rendering in this game is not visible. Readding the SWP_SHOWWINDOW flag to the SetWindowPos calls fixes the problem. But I have the impression that the main point of your patch was removing this flag, so I don't want to just add it again. What were the reasons for removing the SWP_SHOWWINDOW flag? Stefan

2 3