- Wine-Devel - list.winehq.org

Re: added implementation of some more options - try3
by Dmitry Timoshkov Nov. 9, 2005

Nov. 9, 2005

"Vijay Kiran Kamuju" <infyquest(a)gmail.com> wrote: > Added implementation for INTERNET_OPTION_VERSION > Added implementation for INTERNET_OPTION_USER_AGENT > Fixed behavior for INTERNET_OPTION_HTTP_VERSION > Added implementation for unsupported options > Added some minimalist testcases In this version there are still the same problems mentioned before. Notably: in the tests last error is not set to a bogus value before an API call while it's tested afterwards, buffer length is not initialized before the API calls. > + FIXME("This request %ld is not supported by native as of now\n",dwOption); This message is misleading, I'd suggest something like the following: FIXME("Request %ld is not supported\n", dwOption); -- Dmitry.

1 0

Re: crypt32: plug memory leaks
by Robert Shearman Nov. 9, 2005

Nov. 9, 2005

Juan Lang wrote: >Hope you don't mind the memory function changes, they helped me debug the >leaks. > >ChangeLog: >- use macros for memory allocation for easier debugging >- plug memory leaks > I'm not sure why you need to use these extra macros for memory allocation. Can't you simply replace the CryptMem* functions during debugging with a macro? -- Rob Shearman

3 3

Re: KERNEL: GetTempFileNameW should not require prefix
by Dimi Paun Nov. 8, 2005

Nov. 8, 2005

> Index: path.c > =================================================================== > RCS file: /home/wine/wine/dlls/kernel/path.c,v > retrieving revision 1.26 > diff -u -r1.26 path.c > --- path.c 8 Nov 2005 11:01:03 -0000 1.26 > +++ path.c 8 Nov 2005 19:54:35 -0000 Please create the diff from the root of th project: $ cvs diff -u dlls/kernel/path.c -- Dimi Paun <dimi(a)lattica.com> Lattica, Inc.

1 0

kernel32/tests thread.ok fails
by Paul Vriens Nov. 8, 2005

Nov. 8, 2005

Hi, the traces are my doing, but: thread.c:660:test_CreateThread_basic() thread.c:662:test_CreateThread_suspended() thread.c:664:test_SuspendThread() thread.c:666:test_TerminateThread() thread.c:668:test_CreateThread_stack() thread.c:670:test_thread_priority() thread.c:672:test_GetThreadTimes() fixme:thread:NtQueryInformationThread Cannot get kerneltime or usertime of other threads thread.c:674:test_thread_processor() fixme:thread:SetThreadIdealProcessor (0xfffffffe): stub thread.c:676:test_GetThreadExitCode() thread.c:679:test_SetThreadContext() err:seh:EXC_DefaultHandling Unhandled exception code c00000fd flags 0 addr 0xa make: *** [thread.ok] Error 1 this doesn't seem right. Anyone? Cheers, Paul.

1 1

Re: KERNEL: GetTempFileNameW should not require prefix
by James Hawkins Nov. 8, 2005

Nov. 8, 2005

On 11/8/05, Phil Lodwick <Phil.Lodwick(a)efi.com> wrote: > Hi, > > _tempnam(".", NULL) fails when using the builtin msvcrt.dll. > This is because GetTempFileNameW was requiring a non-NULL prefix. > > This is my first patch submission. Please let me know if I am doing it wrong > as I have more to come. > A test case showing this behavior would be nice. For a small change like this, it's best to send the test case and the fix in one patch. -- James Hawkins

1 0

cvswine can't find files
by Coleman Jeremy D SrA 50 OSS/OSOT1 Nov. 8, 2005

Nov. 8, 2005

I have been trying to use cvswine because it has the d3d9 patches merged. The problem I run into is that for some reason it is unable to find required files (particularly dlls) even when run from within the same directory. I have seen this happen with the following programs so far. Warcraft 3: Can't find Game.dll NFSU Demo: Can't find d3d9.dll Warhammer 40000 Dawn of War Demo: Can't find spd3d9.dll All of these files are present... Warcraft works fine in cedega and wine 0.9. NFSU Demo works in Cedega. Warhammer DoW runs with Cedega. Any help would be greatly appreciated. If I can get cvswine with d3d9 to work, then my $5/mo for Cedega is likely to become a WineHQ paypal donation.

1 0

Regression with icons being black?
by Paul Vriens Nov. 8, 2005

Nov. 8, 2005

Hi, just started Process Explorer with latest CVS and the icons for the processes are black instead of colorful. I will do some regression testing. Anyone experiencing the same? Cheers, Paul.

3 3

[PATCH] variable number of arguments in macro __ONLY_IN_WINELIB()
by Berend Reitsma Nov. 8, 2005

Nov. 8, 2005

Hi, With gcc 2.95.4 I am getting a compile error when compiling dlls/mapi32/tests/prop.c because of the use of the FAR keyword: gcc -c -I. -I. -I../../../include -I../../../include -D_REENTRANT -fPIC -Wall -pipe -mpreferred-stack-boundary=2 -fno-strict-aliasing -gstabs+ -Wpointer-arith -g -O2 -o prop.o prop.c prop.c:54: macro `__ONLY_IN_WINELIB' used without args prop.c:55: macro `__ONLY_IN_WINELIB' used without args prop.c:1158: macro `__ONLY_IN_WINELIB' used without args prop.c:1158: macro `__ONLY_IN_WINELIB' used without args prop.c:1159: macro `__ONLY_IN_WINELIB' used without args make: *** [prop.o] Error 1 In include/windef.h the definition of FAR is a call to the macro __ONLY_IN_WINELIB with no arguments. The attached patch changes the definition of the macro to use variable arguments According to some documentation I found on the web this version with __VA_ARGS__ should be compatible with other compilers as well... Regards, Berend Reitsma.

2 1

Re: RESEND [wininet]add implementation of few more options
by Dmitry Timoshkov Nov. 8, 2005

Nov. 8, 2005

"Vijay Kiran Kamuju" infyquest(a)gmail.com wrote: > as previous implementation had some bugs use this version > + case INTERNET_OPTION_VERSION: > { > - if (*lpdwBufferLength < sizeof(HTTP_VERSION_INFO)) > - INTERNET_SetLastError(ERROR_INSUFFICIENT_BUFFER); > + /* > + * Presently hardcoded to 1.2 > + */ > + if(lpBuffer != NULL) > + { > + ((INTERNET_VERSION_INFO*)lpBuffer)->dwMajorVersion = 1; > + ((INTERNET_VERSION_INFO*)lpBuffer)->dwMinorVersion = 2; You still need to check buffer size before filling the buffer. Also using a new variable instead of casting lpBuffer would be much cleaner. Same applies to all cases. > + case INTERNET_OPTION_OFFLINE_MODE: > + case INTERNET_OPTION_OFFLINE_SEMANTICS: > + case INTERNET_OPTION_POLICY: > + case INTERNET_OPTION_RECEIVE_THROUGHPUT: > + case INTERNET_OPTION_SEND_THROUGHPUT: > + WARN("This request is not supported by native either\n"); > + SetLastError(ERROR_NOT_SUPPORTED); > + break; I'd suggest to use FIXME instead of WARN here. It doesn't really matter that current version of miscrosoft's implementation doesn't support them, they may add support for them later, and the apps will start to use them. > + char useragent[] = {'W','i','n','i','n','e','t',' ','T','e','s','t',0 }; char useragent[] = "Wininet Test"; works just fine. > + retval=InternetQueryOptionA(hinet,INTERNET_OPTION_USER_AGENT,NULL,&len); > + err=GetLastError(); Again, checking last error without setting it first to a bogus value. + SetLastError(0); + buffer=HeapAlloc(GetProcessHeap(),0,len+1); + retval=InternetQueryOptionA(hinet,INTERNET_OPTION_USER_AGENT,buffer,&len); + err=GetLastError(); Do not set last error to 0, you then miss the case when API sets it to 0. You don't initialize length variable before calling InternetQueryOption. -- Dmitry.

2 1

Re: [wininet] Add InternetReadFileEx*
by Dmitry Timoshkov Nov. 8, 2005

Nov. 8, 2005

"Brian Gerst" <bgerst(a)didntduck.org> wrote: > This patch implements the InternetReadFileExA and InternetReadFileExW > functions. I don't know if this completely implements it to MSDN specs > (which are fairly vague), but it is enough to allow World of Warcraft to > read and display the server status message. It would be much better instead of duplicating code implement InternetReadFile via InternetReadFileEx. -- Dmitry.

1 0