July 2009 - Wine-devel - list.winehq.org

Re: winex11: Fix Office2007 XRender regression [resubmit]
by Alexandre Julliard 24 Jul '09

24 Jul '09

Roderick Colenbrander <thunderbird2k(a)gmail.com> writes: > This is a resubmit of the previous fix. No changes have been made to > the patch itself. Alexandre preferred to store source coordinates by > default in a transformation matrix but this proved to be very slow on > most XRender implementations (software xrender can easily be 20-30x > faster while inte/nvidia drivers take a 2x hit). A bug report has been > submitted to Xorg as this issue can be fixed inside X without changes > to drivers. My concerns about complexity still stand. Please hide the performance hack inside a helper function. -- Alexandre Julliard julliard(a)winehq.org

1 0

blocking/non-blocking socket question.
by Rein Klazes 24 Jul '09

24 Jul '09

Hi, I am looking at bug #12048 http://bugs.winehq.org/show_bug.cgi?id=12048 >From a relay trace it is easy enough to find where the app is getting dead-locked and why it should not. It is calling recv() on a socket that was previously: 1) set to blocking with a WS_ioctlsock() call with cmd WS_FIONBIO 2) AsyncSelect()'ed that should make the socket non-blocking. The first call set the unix file descriptor to blocking: fcntl( fd, F_SETFL, 0 ); while the second does not touch the unix fd. This leaves the unix file descriptor in the blocking state and the unix recvmsg() call will not terminate. I thought the unix file descriptor should always by non-blocking, but it is a while since I was looking at socket code. Hopefully someone can tell me if I am barking up the right tree. Is what I think correct, and should the fcntl()'s simply go. Or should AsyncSelect() set the unix fd to non blocking instead? Rein.

3 5

Re: [PATCH] Resize fullscreen window when DirectDraw changes the display mode [resubmit 3]
by Johan Gill 23 Jul '09

23 Jul '09

Was there any problems with this patch? It seems to not have made it to CVS. /Johan Gill

1 0

Re: winhlp32: Ensure Open dialog is modal
by Nikolay Sivov 23 Jul '09

23 Jul '09

Owen Rudge wrote: > When selecting the Open option in winhlp32, the window would not have > an owner, and as such would not be modal. This patch passes the active > window's hWnd to GetOpenFileName, partially fixing bug 19081. > > --- > programs/winhlp32/winhelp.c | 2 +- > 1 files changed, 1 insertions(+), 1 deletions(-) --- + openfilename.hwndOwner = (Globals.active_win ? Globals.active_win->hMainWnd : 0);; --- Hi, Owen. Please change ";;" to ";".

1 0

Re: [patch 1/3] user32: put some repeated code for checking and extending of a MENUITEMINFO structure in a separate function.
by Alexandre Julliard 23 Jul '09

23 Jul '09

Rein Klazes <wijn(a)online.nl> writes: > - if( lpmii->cbSize != sizeof( mii) && > - lpmii->cbSize != sizeof( mii) - sizeof ( mii.hbmpItem)) { > - SetLastError( ERROR_INVALID_PARAMETER); > - return FALSE; > - } > - memcpy( &mii, lpmii, lpmii->cbSize); > - if( lpmii->cbSize != sizeof( mii)) { > - mii.cbSize = sizeof( mii); > - mii.hbmpItem = NULL; > - } > - > item = MENU_InsertItem(hMenu, uItem, bypos ? MF_BYPOSITION : 0 ); > - return SetMenuItemInfo_common(item, (const MENUITEMINFOW *)&mii, FALSE); > + if( MENU_NormalizeMenuItemInfoStruct( (MENUITEMINFOW *)lpmii, &mii)) > + return SetMenuItemInfo_common(item, &mii, FALSE); > + return FALSE; You should do the check before the item insertion, like the existing code does. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: winhttp(1/2): Implement connect timeout (try 2)
by Alexandre Julliard 23 Jul '09

23 Jul '09

Juan Lang <juan.lang(a)gmail.com> writes: > -BOOL netconn_connect( netconn_t *conn, const struct sockaddr *sockaddr, unsigned int addr_len ) > +BOOL netconn_connect( netconn_t *conn, const struct sockaddr *sockaddr, unsigned int addr_len, int timeout ) > { > - if (connect( conn->socket, sockaddr, addr_len ) == -1) > + int res, state; > + > + if (timeout > 0) > + { > + state = 1; > + res = ioctlsocket( conn->socket, FIONBIO, &state ); > + } > + res = connect( conn->socket, sockaddr, addr_len ); > + if (res == -1 && (errno == EINPROGRESS || errno == EAGAIN)) > + { > + struct pollfd pfd; > + > + pfd.fd = conn->socket; > + pfd.events = POLLOUT; > + res = poll( &pfd, 1, timeout ); > + } You also need to replace the errno check with the equivalent winsock errors, and you need to correctly check the result of the poll. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: user32: Draw the whole button only if requested. Take 2.
by Alexandre Julliard 23 Jul '09

23 Jul '09

Dmitry Timoshkov <dmitry(a)codeweavers.com> writes: > The application I'm working on relies on this, it doesn't expect that > button completely redraws itself on WM_SETFOCUS/WM_KILLFOCUS messages. > > This version of the patch skips the button drawing in the case > of action == ODA_FOCUS. You then need to draw the focus rect unconditionally for ODA_FOCUS. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: patch: fixed Dutch translation
by Paul Vriens 22 Jul '09

22 Jul '09

Vincent Beers wrote: > GMail decided to throw away my patch for some reason. Reattached. > > -------- Original Message -------- > > Vincent Beers wrote: >> This is my first patch, ever, so please bear with me and tell me if I >> made any mistakes in generating and submitting the patch. :) >> >> This is a simple patch that improves the Dutch translation in winecfg. >> >> Vincent Beers >> Hi, This change looks a bit odd: - CONTROL "Listview",IDC_LIST_SFPATHS,"SysListView32",LVS_REPORT | LVS_AUTOARRANGE | LVS_ALIGNLEFT | + CONTROL "Lijstweergave",IDC_LIST_SFPATHS,"SysListView32",LVS_REPORT | LVS_AUTOARRANGE | LVS_ALIGNLEFT | It won't do any harm I guess as the text is basically ignored. As a matter of fact I even think we are better off setting all these SysListView32 to "" instead of giving it some text (we have both variations in the code currently). -- Cheers, Paul.

2 1

Re: winhttp(6/6): Implement connect timeout
by Alexandre Julliard 22 Jul '09

22 Jul '09

Juan Lang <juan.lang(a)gmail.com> writes: > +BOOL netconn_connect( netconn_t *conn, const struct sockaddr *sockaddr, unsigned int addr_len, int timeout ) > { > - if (connect( conn->socket, sockaddr, addr_len ) == -1) > + int flags = -1, res; > + > + if (timeout > 0) > + flags = fcntl( conn->socket, F_GETFL, 0 ); > + if (flags != -1) > + fcntl( conn->socket, F_SETFL, flags | O_NONBLOCK ); > + res = connect( conn->socket, sockaddr, addr_len ); > + if (res == -1 && (errno == EINPROGRESS || errno == EAGAIN)) This won't compile on Windows. You should use ioctlsocket and friends. -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: [PATCH 1/6] wined3d: Keep track of a thread's wined3d context.
by Alexandre Julliard 22 Jul '09

22 Jul '09

Henri Verbeet <hverbeet(a)codeweavers.com> writes: > +BOOL context_init_current(void) > +{ > + return TlsSetValue(wined3d_context_tls_idx, NULL); > +} Note that this is not necessary, TLS slots are initialized to 0. -- Alexandre Julliard julliard(a)winehq.org

1 0