October 2004 - Wine-devel - list.winehq.org

Re: Fix various function prototypes
by Rolf Kalbermatter 17 Oct '04

17 Oct '04

Vincent Béron wrote: >MSDN says it's void >http://msdn.microsoft.com/library/default.asp?url=/library/en- >us/shellcc/platform/shell/reference/functions/shcocreateinstance.asp > >I don't have latest PSDK. > >This will need to be investigated some more (although I'm leaning >towards HRESULT). My own tests seem to indicate that the function really returns some codes which are consistent with HRESULTs. Also my SDK (from February 2003) declares the functions as SHSTDAPI which is an abbreviation for "extern HRESULT __stdcall". So MSDN seems to be wrong here and besides returning them anyhow in the implementation eventhough some clients may ignore them based on MSDN documentation seems the correct way to me! Rolf Kalbermatter

1 0

Re: Enable _WIN64
by Robert Shearman 17 Oct '04

17 Oct '04

Vincent Béron wrote: >Index: wine/dlls/oleaut32/oaidl_p.c >=================================================================== >RCS file: /home/wine/wine/dlls/oleaut32/oaidl_p.c,v >retrieving revision 1.4 >diff -u -r1.4 oaidl_p.c >--- wine/dlls/oleaut32/oaidl_p.c 7 Oct 2004 03:06:49 -0000 1.4 >+++ wine/dlls/oleaut32/oaidl_p.c 16 Oct 2004 04:50:00 -0000 >@@ -11119,7 +11119,7 @@ > }; > > >-#if !defined(__RPC_WIN32__) >+#if !defined(__RPC_WIN32__) && !defined(__RPC_WIN64__) > #error Invalid build platform for this stub. > #endif > > > oaidl.idl needs to be regenerated from a newer MIDL that knows about Win64 instead of this hack. Rob

1 0

Re: Fix various function prototypes
by Robert Shearman 17 Oct '04

17 Oct '04

Vincent Béron wrote: >SHCoCreateInstance returns void, not HRESULT. >Some functions had a different prototype than their declaration. > >Changelog: >Fix various function prototypes. > > Some of these are wrong if we want binary compatibility with Windows 64-bit because e.g.: sizeof(unsigned long) on linux != sizeof(unsigned long) on Windows If a parameter to a function is "unsigned long" then we will have to change it to "ULONG" instead. Rob

1 0

Re: Freetype portability fix for older versions - improved
by Rein Klazes 17 Oct '04

17 Oct '04

On 15 Sep 2004 11:17:10 -0400, you wrote: > Tested on both freetype 2.1.2 and freetype 2.1.4. > > Changelog: > Use the preferred FT_ENCODING values rather than ft_encoding. > > +#ifndef FT_ENCODING_NONE > +#define FT_ENCODING_NONE ft_encoding_none > +#endif > +#ifndef FT_ENCODING_MS_SYMBOL > +#define FT_ENCODING_MS_SYMBOL ft_encoding_symbol > +#endif > +#ifndef FT_ENCODING_UNICODE > +#define FT_ENCODING_UNICODE ft_encoding_unicode > +#endif > +#ifndef FT_ENCODING_APPLE_ROMAN > +#define FT_ENCODING_APPLE_ROMAN ft_encoding_apple_roman > +#endif The patch works, but is not quite correct and doing the opposite of what you claim: the FT_ENCODING's are never defined but are enums. Therefore "ifndef FT_ENCODING_XYZ" is always true and you always use the depreciated ft_encodings. On old freetype2: enum FT_Encoding { ft_encoding_xyz, ... On Newer freetype2: enum FT_Encoding { FT_ENCODING_XYZ, ... #define ft_encoding_xyz FT_ENCODING_XYZ So to use FT_ENCODING_XYZ in the code and still compile with older freetype2's, we need to test: #ifndef ft_encoding_xyz Patch attached, tested to compile with freetype 2.0.9 and 2.1.7. Changelog: dlls/gdi : freetype.c Test whether the ft_encoding's are defined, not the FT_ENCODING's. Rein. -- Rein Klazes rklazes(a)xs4all.nl

1 0

Re: winesetuptk0.72/tcltk-winesetuptk-0.72
by Ivan Leo Puoti 16 Oct '04

16 Oct '04

> /tcltk-winesetuptk-0.72/build.sh Winesetuptk isn't maintained any more, and it isn't needed to configure wine. Why do you need it? Ivan. ____________________________________________________________ Libero ADSL: navighi gratis a 1.2 Mega, senza canone e costi di attivazione. Abbonati subito su http://www.libero.it

1 0

Re: Honor --disable-debug or --disable-trace completely
by Vincent Béron 16 Oct '04

16 Oct '04

Le sam 16/10/2004 à 22:29, Vincent Béron a écrit : > (Yes, the patch is gzip as it's over 200k) > > wine/debug.h has some tests based on NO_{DEBUG,TRACE}_MSGS, which is set > (or not) in config.h. Quite a few files didn't #include config.h before > wine/debug.h, so some debugmsgs were always compiled in, even with > ./configure --disable-debug. > > Now maybe we want to install config.h as well, as debug.h is already > installed. I'll leave that one to Alexandre. > > Changelog: > Force inclusion of config.h before wine/debug.h. Don't commit yet, make depend fails (config.h has been added to some includes in some dirs). I'll resend. Vincent

1 0

Re: Use LLP64 compatible sized-types for LONG and DWORD
by Dmitry Timoshkov 16 Oct '04

16 Oct '04

"Vincent Béron" <vberon(a)mecano.gme.usherb.ca> wrote: > Changelog: > Change the type of LONG and DWORD to be int based instead of long based > for Win64 compatibility. It would be cleaner IMO to use that only when compiling with _WIN64 defined, that would avoid affecting the Win32 target. -- Dmitry.

2 2

Re: Fix various function prototypes
by Dmitry Timoshkov 16 Oct '04

16 Oct '04

"Vincent Béron" <vberon(a)mecano.gme.usherb.ca> wrote: > SHCoCreateInstance returns void, not HRESULT. According to Platform SDK headers SHCoCreateInstance returns HRESULT. P.S. Do you compile Wine with _WIN64 defined on Alpha? That would be really cool to see that going, clean our headers and fix all warnings caused by 32 bit pointer assumptions. -- Dmitry.

2 3

RFH: winuser.h
by Dimitrie O. Paun 16 Oct '04

16 Oct '04

Can someone please help me with the value of LBS_COMBOBOX? TIA, Dimi.

2 1

programs/winecfg/libraries.c warnings regressions
by Gerald Pfeifer 16 Oct '04

16 Oct '04

I believe the following patch of programs/winecfg/libraries.c revision 1.3 date: 2004/09/28 03:55:16; author: julliard; state: Exp; lines: +288 -501 Mike Hearn <mike(a)navi.cx> - rewrite the transaction system to be based on a settings overlay, to have a nicer API, and to actually work (always a bonus) - change the libraries page to be based on a listbox rather than a treeview, clean up and shrink the code - add accelerator keys to the libraries page, focus management - make the window title reflect what the user is currently editing - remove bogus root warning - remove some unused control IDs in resource.h - start converting the x11drv dialog to kernel_style from javaStyle - bugfixing introduced the following warnings libraries.c:261: warning: `mode' might be used uninitialized in this function libraries.c:264: warning: `str' might be used uninitialized in this function (which are the only ones of this kind for a full Wine build). Gerald

3 2