September 2013 - Wine-Devel

Re: ntdll: Try application manifest path when searching for dependent assemblies
by Alexandre Julliard Sept. 10, 2013

Sept. 10, 2013

Nikolay Sivov <nsivov(a)codeweavers.com> writes: > @@ -2776,8 +2776,7 @@ static NTSTATUS lookup_assembly(struct actctx_loader* acl, > /* FIXME: add support for language specific lookup */ > > nameW.Buffer = NULL; > - if (!(buffer = RtlAllocateHeap( GetProcessHeap(), 0, > - (strlenW(acl->actctx->appdir.info) + 2 * strlenW(ai->name) + 2) * sizeof(WCHAR) + sizeof(dotManifestW) ))) > + if (!(buffer = RtlAllocateHeap( GetProcessHeap(), 0, MAX_PATH * sizeof(WCHAR) ))) Please try to allocate the correct size. If you want to use MAX_PATH you then have to check for overflow when building the string, which is probably worse. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [2/5] ntdll: Make NtWriteFile handle special offset -2 same way as -1. Resend.
by Alexandre Julliard Sept. 10, 2013

Sept. 10, 2013

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > --- > dlls/ntdll/file.c | 31 +++++++++++++++---------------- > 1 file changed, 15 insertions(+), 16 deletions(-) Why do you need that? It doesn't look to me like an improvement. -- Alexandre Julliard julliard(a)winehq.org

2 2

Re: [2/2] oledb32: Implement IDataSourceLocator get/put hWnd (resend)
by Nikolay Sivov Sept. 10, 2013

Sept. 10, 2013

On 9/10/2013 13:30, Alistair Leslie-Hughes wrote: > Hi, > > > Changelog: > oledb32: Implement IDataSourceLocator get/put hWnd > > > Best Regards > Alistair Leslie-Hughes Hi, Alistair. This is wrong, interface definition should be fixed as well, HWND parameter can't be LONG: --- [id(0x60020000), propget] HRESULT hWnd([out, retval] long* phwndParent); [id(0x60020000), propput] HRESULT hWnd([in] long phwndParent); --- PSDK header has COMPATIBLE_LONG thing to make it work on win64.

1 0

Re: xcopy: Add an explicit check to avoid an out of array access. access. (RESEND)
by Alexandre Julliard Sept. 10, 2013

Sept. 10, 2013

Gerald Pfeifer <gerald(a)pfeifer.com> writes: > On Sun, 24 Mar 2013, Gerald Pfeifer wrote: >> This _may_ not be necessary, strictly speaking, but new versions of >> GCC (GCC 4.8) complain about this. (And there might be a fringe case?) > > How about this patch? I see it was not applied, but also did not > see a response. It can't happen (and removing the last char unconditionally is broken anyway). -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [1/5] ntdll: Add support to NtWriteFile for special offset -1. Take 3.
by Alexandre Julliard Sept. 10, 2013

Sept. 10, 2013

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > @@ -961,16 +961,23 @@ NTSTATUS WINAPI NtWriteFile(HANDLE hFile, HANDLE hEvent, > > if (type == FD_TYPE_FILE) > { > + off_t off; > + > if (!(options & (FILE_SYNCHRONOUS_IO_ALERT | FILE_SYNCHRONOUS_IO_NONALERT)) && !offset) > { > status = STATUS_INVALID_PARAMETER; > goto done; > } > > - if (offset && offset->QuadPart != (LONGLONG)-2 /* FILE_USE_FILE_POINTER_POSITION */) > + off = offset ? offset->QuadPart : (off_t)-2 /* FILE_USE_FILE_POINTER_POSITION */; > + > + if (off == (off_t)-1 /* FILE_WRITE_TO_END_OF_FILE */) > + off = lseek( unix_handle, 0, SEEK_END ); You are not supposed to change the position for the asynchronous case. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: ws2_32: Implement get socket option SO_PROTOCOL_INFO (try 5)
by Marvin Sept. 10, 2013

Sept. 10, 2013

Hi, While running your changed tests on Windows, I think I found new failures. Being a bot and all I'm not very good at pattern recognition, so I might be wrong, but could you please double-check? Full results can be found at https://newtestbot.winehq.org/JobDetails.pl?Key=2084 Your paranoid android. === wxppro (32 bit sock) === sock.c:509: Test failed: oob_server (b8): unexpectedly at the OOB mark: 0 sock.c:519: Test failed: oob_server (b8): unexpectedly at the OOB mark: 0 sock: unhandled exception c0000005 at 71AB8D62 === w7pro64 (32 bit sock) === sock.c:528: Test failed: oob_server (138): not at the OOB mark: 1 === w7pro64 (64 bit sock) === sock.c:509: Test failed: oob_server (934): unexpectedly at the OOB mark: 0 sock.c:519: Test failed: oob_server (934): unexpectedly at the OOB mark: 0 sock.c:288: Test failed: do_synchronous_recv (934): error 10054: sock.c:534: Test failed: oob_server (934): not at the OOB mark: 1

1 0

Re: [1/5] ntdll: Add support to NtWriteFile for special offset -1. Take 2.
by Alexandre Julliard Sept. 9, 2013

Sept. 9, 2013

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > --- > dlls/ntdll/file.c | 13 ++++++++++--- > dlls/ntdll/tests/file.c | 15 --------------- > 2 files changed, 10 insertions(+), 18 deletions(-) This is causing test failures all over the place. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH 2/5] wait until all data from earlier test has been written in test_waittxempty
by Dmitry Timoshkov Sept. 9, 2013

Sept. 9, 2013

Wolfgang Walter <wine(a)stwm.de> wrote: > + if (res || (!res && GetLastError() == ERROR_IO_PENDING)) > + /* if data has been sent: wait for termination */ > + Sleep(timeout); I don't see such a problem with real COM-port and serial-USB cable under Windows or Linux here and under testbot VMs. -- Dmitry.

3 12

Re: server: Add a serial event mask change counter.
by Alexandre Julliard Sept. 9, 2013

Sept. 9, 2013

Dmitry Timoshkov <dmitry(a)baikal.ru> writes: > @@ -209,7 +211,7 @@ DECL_HANDLER(get_serial_info) > reply->writemult = serial->writemult; > > /* event mask */ > - reply->eventmask = serial->eventmask; > + reply->eventmask = serial->eventmask | (serial->generation << 16); It would be better to return it as a separate field, particularly since eventmask is already an int. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH 3/6] ws2_32: Fix parameters checking in WSADuplicateSocket
by Alexandre Julliard Sept. 9, 2013

Sept. 9, 2013

Bruno Jesus <00cpxxx(a)gmail.com> writes: > + fd = get_sock_fd( s, FILE_READ_DATA, NULL ); > + if (fd == -1) > + { > + SetLastError(WSAENOTSOCK); > + return SOCKET_ERROR; > + } > + release_sock_fd( s, fd ); You should avoid retrieving the fd if you are not using it. The handle should be checked at the point where you are doing something with it. -- Alexandre Julliard julliard(a)winehq.org

2 2