October 2008 - Wine-devel - list.winehq.org

Re: mscms: quiet a noisy fixme (1/3)
by Henri Verbeet 17 Oct '08

17 Oct '08

I don't think you want to change the return code based on whether the FIXME is shown or not.

2 1

Re: mscvrt: updated use of pmode flag in _s_open - set permissions of thefile after it's been created.
by Dmitry Timoshkov 17 Oct '08

17 Oct '08

"Ivan Peevski" <cyberhorse(a)gmail.com> wrote: > + /* Ignore _S_IREAD and _S_IWRITE if file exists. */ > + if (stat(path, &buf) != 0) > + pmode = pmode & (MSVCRT__S_IREAD | MSVCRT__S_IWRITE); /*relevant flags*/ You need to use MSVCRT__stat(), not the system stat(), and appropriately typed buf. -- Dmitry.

2 1

Re: dsound: Add typedefs for uint8_t, uint16_t and uint32_t for systems that don't have stdint.h.
by Dmitry Timoshkov 17 Oct '08

17 Oct '08

"Rob Shearman" <robertshearman(a)gmail.com> wrote: > --- > dlls/dsound/dsound_convert.c | 4 ++++ > 1 files changed, 4 insertions(+), 0 deletions(-) I'd argue that dsound_convert.c should use win32 types instead. -- Dmitry.

1 0

Re: dsound: Add typedefs for uint8_t, uint16_t and uint32_t for systems that don't have stdint.h.
by Michael Karcher 17 Oct '08

17 Oct '08

Am Freitag, den 17.10.2008, 11:06 +0100 schrieb Rob Shearman: > +typedef unsigned char uint8_t; > +typedef unsigned short uint16_t; > +typedef unsigned int uint32_t; This probably works on all platforms Wine is designed for, but why don't you use the autoconf provision for these types called AC_CREATE_STDINT_H or AX_CREATE_STDINT_H to create "wine/stdint.h" (which will create a one-liner including <stdint.h> on systems where it is available)? Regards, Michael Karcher

1 0

Re: [PATCH 03/10] jscript: Added NaN value implementation.
by Alexandre Julliard 17 Oct '08

17 Oct '08

Jacek Caban <jacek(a)codeweavers.com> writes: > +#include <math.h> > +#include "basetsd.h" > + > +#ifndef NAN > + > +#ifdef __GNUC__ > + > +# define NAN \ > + (__extension__ \ > + ((union { unsigned __l __attribute__ ((__mode__ (__SI__))); float __d; }) \ > + { __l: 0x7fc00000UL }).__d) > + > +#else /* __GNUC__ */ > + > +#ifdef WORDS_BIGENDIAN > +# define __nan_bytes { 0x7f, 0xc0, 0, 0 } > +#else > +# define __nan_bytes { 0, 0, 0xc0, 0x7f } > +#endif > + > +static union { unsigned char __c[4]; float __d; } __nan_union > + __attribute_used__ = { __nan_bytes }; > +# define NAN (__nan_union.__d) > + > +#endif /* __GNUC__ */ > + > +#endif /* NAN */ I'd suggest to take advantage of the variant union instead of copying all these glibc definitions. Something like #ifdef NAN V_R8(&vt) = NAN; #else V_I8(&vt) = nan_magic_pattern; #endif -- Alexandre Julliard julliard(a)winehq.org

4 4

Re: Kernel32: 3GB UserSpace boot option
by Stefan Reimer 17 Oct '08

17 Oct '08

Hi, compiling binaries with the LARGE_ADDRESS_AWARE flag set is not enough to be able to use more than 2GB user space. To behave like WindowsXP booted with the /3GB option, at least the lpMaximumApplicationAddress in GetSystemInfo has to be changed to 0xBFFEFFFF and the virtual user space has to be reserved in mmap_init in mmap.c up to that address. See Windows memory management: http://ymei.freeshell.org/gopher/Book/Programming%20Applications%20for%20MS… http://codeidol.com/other/inside-windows-2000/Memory-Management/Address-Spa… I am quite sure that my proposed hack is not complete to behave like the /3GB option, but as stated in cpu.c 384 /* FIXME: the two entries below should be computed somehow... */ 385 cachedsi.lpMinimumApplicationAddress = (void *)0x00010000; 386 cachedsi.lpMaximumApplicationAddress = (void *)0x7FFEFFFF; we should look at this in order to be able to use more than 2GB user space for apps that need it. Regards Stefan Reimer

2 3

Re: [1/3] netapi32/tests: Add a test for well known account RIDs
by Nicolas Le Cam 17 Oct '08

17 Oct '08

Its obvious but don't apply, third patch is missing comment and unused variables cleaning. I will submit a new patch later. regards, Nicolas Le Cam 2008/10/17 Nicolas Le Cam <niko.lecam(a)gmail.com>: > Hi, > > This path series intend to fix a failing test on localized windows > version. > The first two patches are needed to reveal and fix a different behavior > between windows and wine in NetQueryDisplayInformation Level 1 that I'm > using in the third patch to fix a failing test on my localized windows > without having to rework completely the test logic. > > regards, > Nicolas Le Cam > > > >

1 0

Re: [PATCH] shlwapi: Fix pointer and size in AssocQueryStringA
by Juan Lang 16 Oct '08

16 Oct '08

Hi Detlef, - if (dwLenOut >= MAX_PATH) + if (dwLenOut >= MAX_PATH) { lpszReturnW = HeapAlloc(GetProcessHeap(), 0, (dwLenOut + 1) * sizeof(WCHAR)); + } This first set of parentheses seems like an unnecessary addition.. + else + { + dwLenOut = sizeof(szReturnW) / sizeof(szReturnW[0]); + } and this one's style doesn't match the above. The rest of the file doesn't appear to use braces with only a single-line expression, so would you mind sticking with the existing style? Same with the next hunk, it introduces braces where none are needed. Thanks, --Juan

1 0

Re: [PATCH] wintrust: Implement CryptCATAdminCalcHashFromFileHandle
by Juan Lang 16 Oct '08

16 Oct '08

Hi Maarten, overall this patch looks pretty good. A few minor comments: + tempbuffer = HeapAlloc(GetProcessHeap(), 0, TEMP_BLOCK_SIZE); + if (!tempbuffer) + goto out; You return TRUE in this case. You probably ought to set last error and return FALSE instead. + while (ReadFile(hFile, tempbuffer, TEMP_BLOCK_SIZE, &readbytes, NULL) && readbytes) + CryptHashData(hash, tempbuffer, readbytes, 0); Why not while ((ret = ReadFile...) && readbytes) instead? This way you can at least return failure if reading the file fails for whatever reason. + SetLastError(ERROR_INSUFFICIENT_BUFFER); return TRUE; Why are you setting ERROR_INSUFFICIENT_BUFFER in the success case? If it's just to pass that test, I think you might remove the test instead: in general we don't like to check last error in case of success, unless we know an app does so. --Juan

2 3

Re: setupapi: Implement installing catalog files
by Vincent Povirk 16 Oct '08

16 Oct '08

The patch you sent doesn't match the subject line. Vincent Povirk On Wed, Oct 15, 2008 at 11:05 AM, Maarten Lankhorst <maarten(a)codeweavers.com> wrote: > --- > The functions called here are stubs, but don't cause this code to bomb out. > > From 04917f75269b6f4f51924ba734c20dc05617d882 Mon Sep 17 00:00:00 2001 > From: Maarten Lankhorst <maarten(a)codeweavers.com> > Date: Tue, 14 Oct 2008 20:16:14 +0200 > Subject: [PATCH] advapi32: Make CheckTokenMembership return FALSE for asking > whether we are connected through the network > > --- > dlls/advapi32/security.c | 21 +++++++++++++++++---- > 1 files changed, 17 insertions(+), 4 deletions(-) > > diff --git a/dlls/advapi32/security.c b/dlls/advapi32/security.c > index ca26217..5a2b3e4 100644 > --- a/dlls/advapi32/security.c > +++ b/dlls/advapi32/security.c > @@ -615,10 +615,23 @@ BOOL WINAPI > CheckTokenMembership( HANDLE TokenHandle, PSID SidToCheck, > PBOOL IsMember ) > { > - FIXME("(%p %p %p) stub!\n", TokenHandle, SidToCheck, IsMember); > - > - *IsMember = TRUE; > - return(TRUE); > + WCHAR *sid = NULL; > + WCHAR network[] = { 'S','-','1','-','5','-','2',0}; > + BOOL ret = TRUE; > + FIXME("(%p %p %p) stub!\n", TokenHandle, SidToCheck, IsMember); > + > + ConvertSidToStringSidW(SidToCheck, &sid); > + FIXME("Sid: %s\n", debugstr_w(sid)); > + /* We are not connected through the network in any way, but locally > + * Internet Explorer 7 installer checks for it and bombs out if > + * we return TRUE on asking whether we are > + */ > + if (sid && !lstrcmpW(sid, network)) > + ret = FALSE; > + LocalFree(sid); > + > + *IsMember = ret; > + return TRUE; > } > > /****************************************************************************** > -- > 1.5.6.5 > > > > >

2 1