May 2007 - Wine-devel - list.winehq.org

include: Include local not system sql.h header
by Andrew Talbot 12 May '07

12 May '07

I have just sent in a patch to use Wine's sql.h instead of the system one, because I noticed that the latter includes <sqltypes.h>, which appears to redefine DWORD as unsigned long, based on its reckoning that the size of a long integer is four bytes on default platforms, whereas - with _MSC_VER undefined - DWORD is an unsigned int. However, I now see that Wine's sql.h includes sqltypes.h, too. So my patch may be unnecessary or wrong, but should odbcinst.h preceed its inclusion of sql.h with some sort of definition of SIZEOF_LONG to prevent this redefinition? Please advise. Thanks, -- Andy.

1 1

Re: PATCH - implement LockWorkStation using xdg-screensaver
by Marcus Meissner 12 May '07

12 May '07

On Sat, May 12, 2007 at 06:13:42AM -0400, Steven Edwards wrote: > Hi, > This patch was an attempt to implemented LockWorkStation using > xdg-screensaver. It works for my simple test app that calls the > function but my real world application Process Explorer ends up > hanging now rather than crashing and still does not properly lock the > workstation. I am not quite sure why. Perhaps someone else is > interested in this to help improve wine usability. Also I guess I > should use HeapAlloc rather than malloc. > > -- > Steven Edwards > > "There is one thing stronger than all the armies in the world, and > that is an idea whose time has come." - Victor Hugo > From 8e7001c9e3309492c881e9c1f717f26e3e379dfb Mon Sep 17 00:00:00 2001 > From: Steven Edwards <winehacker(a)gmail.com> > Date: Sat, 12 May 2007 05:49:52 -0400 > Subject: [PATCH] Implemented LockWorkStation using xdg-screensaver > > + * LockWorkStation (USER32.@) > + */ > +BOOL WINAPI LockWorkStation(void) > +{ > + /* Use xdg-screensaver */ > + static const char screensaver[] = "/usr/bin/xdg-screensaver"; > + static const char lock_command[] = "lock"; > + char **argv; > + > + FILE *f = fopen( screensaver, "r" ); > + if (!f) > + { > + FIXME_(graphics)( "no support for xdg-screensaver detected\n"); > + return 0; > + } You should use either stat/lstat or just check for failure from execve. > + argv = xmalloc(sizeof(lock_command)); > + memcpy( argv[1] , lock_command, sizeof(lock_command)); This is not going to work. char *argv[2]; argv[0] = "/usr/bin/xdg-screensaver"; argv[1] = "lock"; argv[2] = NULL; > + execv( screensaver, argv ); And actually you cannot use execve(), since it replaces the current process with the new one. Use system("/usr/bin/xdg-screensaver lock") or CreateProcess(). Ciao, Marcus

1 0

Fast events
by Maarten Lankhorst 12 May '07

12 May '07

Hi all, Since I'm trying to improve on winmm's timing scheduling, the best improvement I can find is getting rid of the wineserver calls, they add a huge delay to the timer and that is bad for something that should sleep all the time to get better scheduling. I tried sending in a patch that used linux sockets but it was rejected. Similar patches already do exist: winealsa and wineoss use them, some other sound drivers might use them too. Instead of duplicating that code everywhere I think it's better to add a generic implementation of it that can very easily be changed back to a win32 implementation. What are your thoughts on this? Maarten fevent patch attached. >From 5e5cb0a326f51dc2edef245fcaab4a071f9e0c42 Mon Sep 17 00:00:00 2001 From: Maarten Lankhorst <m.b.lankhorst(a)gmail.com> Date: Sat, 12 May 2007 15:19:45 +0200 Subject: [PATCH] wine headers: introduce fevent interface --- include/wine/fevent.h | 145 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 145 insertions(+), 0 deletions(-) diff --git a/include/wine/fevent.h b/include/wine/fevent.h new file mode 100644 index 0000000..e80391d --- /dev/null +++ b/include/wine/fevent.h @@ -0,0 +1,145 @@ +/* + * Fast event interface + * Use this for when default wineserver calls are extremely bad for time critical usages + * This is also REALLY *ONLY* useful in case of 1 thread doing a 'WaitForSingleObject' + * else weird effects might occur. If someone wants to use this in any other way, results + * are undetermined. + * + * Copyright (C) 2007 Maarten Lankhorst + * + * This library is free software; you can redistribute it and/or + * modify it under the terms of the GNU Lesser General Public + * License as published by the Free Software Foundation; either + * version 2.1 of the License, or (at your option) any later version. + * + * This library is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU + * Lesser General Public License for more details. + * + * You should have received a copy of the GNU Lesser General Public + * License along with this library; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA + */ + +#ifndef __WINESRC__ +#error This should only be used internally +#endif + +#ifndef __WINE_WINE_FEVENT_H +#define __WINE_WINE_FEVENT_H + +#if (defined(HAVE_POLL_H) || defined(HAVE_SYS_POLL_H)) && !defined(WINE_FEVENT_USE_EVENT) + +#ifdef HAVE_POLL_H +#include <poll.h> +#endif +#ifdef HAVE_SYS_POLL_H +#include <sys/poll.h> +#endif + +#include <errno.h> + +WINE_DECLARE_DEBUG_CHANNEL(fevent); + +struct TAGfevent { + int fd[2]; + DWORD manreset; +}; + +typedef struct TAGfevent *FEVENT; + +static void reset_fd(FEVENT fev) +{ + char readme; + struct pollfd pfd; + pfd.fd = fev->fd[0]; + pfd.events = POLLIN; + while (poll(&pfd, 1, 0) > 0) + read(fev->fd[0], &readme, sizeof(readme)); +} + +static void set_fd(FEVENT fev) +{ + char foo = 'A'; + + write(fev->fd[1], &foo, sizeof(foo)); +} + +static FEVENT create_fd(LPSECURITY_ATTRIBUTES lpEventAttributes, BOOL bManualReset, BOOL bInitialState, LPCWSTR lpName) +{ + FEVENT fev = HeapAlloc(GetProcessHeap(), 0, sizeof(*fev)); + + if (!fev) + goto err; + + if (lpEventAttributes || lpName) + FIXME_(fevent)("FEVENTS are meant to be internal and shouldn't have security attributes or a name\n"); + + if (pipe(fev->fd) < 0) + { + fev->fd[0] = fev->fd[1] = -1; + goto err; + } + fev->manreset = bManualReset; + if (bInitialState) + set_fd(fev); + return fev; + + err: + if (fev) + { + close(fev->fd[0]); + close(fev->fd[1]); + HeapFree(GetProcessHeap(), 0, fev); + } + return NULL; +} + +static void close_fd(FEVENT fev) +{ + close(fev->fd[0]); + close(fev->fd[1]); + fev->fd[0] = fev->fd[1] = -1; + HeapFree(GetProcessHeap(), 0, fev); +} + +static DWORD poll_fd(FEVENT fev, DWORD timeout) +{ + int err; + struct pollfd pfd; + pfd.fd = fev->fd[0]; + pfd.events = POLLIN; + + while ((err = poll(&pfd, 1, (int)timeout)) == -1) + if (errno != EAGAIN && errno != EINTR) + break; + switch (err) + { + case 0: return WAIT_TIMEOUT; + case -1: return WAIT_FAILED; + default: + if (!fev->manreset) + reset_fd(fev); + return WAIT_OBJECT_0; + } +} + +#define RESETEVENT reset_fd +#define SETEVENT set_fd +#define CREATEEVENT create_fd +#define CLOSEEVENT close_fd +#define WAITEVENT poll_fd + +#else + +#define FEVENT HANDLE +#define RESETEVENT ResetEvent +#define SETEVENT SetEvent +#define CREATEEVENT CreateEventW +#define CLOSEEVENT CloseHandle +#define WAITEVENT WaitForSingleObject + +#endif + +#endif /* __WINE_WINE_FEVENT_H */ -- 1.4.4.2

1 0

Re: PATCH - implement LockWorkStation using xdg-screensaver
by Stefan Dösinger 12 May '07

12 May '07

Am Samstag 12 Mai 2007 12:13 schrieb Steven Edwards: > Hi, > This patch was an attempt to implemented LockWorkStation using > xdg-screensaver. It works for my simple test app that calls the > function but my real world application Process Explorer ends up > hanging now rather than crashing and still does not properly lock the > workstation. I am not quite sure why. Perhaps someone else is > interested in this to help improve wine usability. Also I guess I > should use HeapAlloc rather than malloc. Yeah, you should use HeapAlloc, and perhaps remove the xmalloc() wrapper :-) . If you test for the existence of xdg-screensaver that way you should close the opened file afterwards. If I remember correctly execvp would check $PATH for the file, so you don't have to hardcode /usr/bin/xdg-screensaver, but I think some "winescreensaver" program simmilar to winebrowser which is configureable via a registry setting would be even better. Also you should return TRUE and FALSE instead of 1 and 0.

1 0

Re: [PATCH] initialize pSurface, since we process it on failure
by Stefan Dösinger 12 May '07

12 May '07

Am Samstag 12 Mai 2007 11:39 schrieb Marcus Meissner: > Hi, > > Not sure if we should perhaps not do the ICOM CAST magic on a failing > CreateSurface call here. > > But initializing pSurface to NULL also helps. The ICOM_CAST thing should set the output to NULL of the surface returned by ddraw7_CreateSurface is NULL, but ddraw7 doesn't always set it to NULL. I wanted to write a test for this because some calls leave the output pointer untouched if they fail and some set it to NULL. AJ also says we should replace transgamings com casting macros with inline functions.

1 0

Current behavior of netapi32-access tests on W2K3 and Vista
by Paul Vriens 12 May '07

12 May '07

Hi, Not so long ago some tests where added for NetUserAdd and friends. While browsing I found that I have about 6 testuser home directories on W2K3 and 4 on Vista (which is the amount of tests I ran on those platforms). Bottom-line is that NetUserDel doesn't delete the home-directory on W2K3 and Vista (which is a good thing) but has no option to delete it explicitly (which is a bad thing IMHO). Any idea how to cleanup after these tests? Just remove the home directory? Cheers, Paul.

2 1

mandriva rpm wine 0.9.37
by marco 11 May '07

11 May '07

As I have still no way to upload the mandriva wine rpm to the official winehq download server. I put it on my own website. So mandriva users you can download the latest wine binaries here: http://www.xs4all.nl/~meijer2/wine/ kind regard, Marco

1 0

Re: [PATCH] Uninstaller: Use a define'd length instead of a magic number
by James Hawkins 11 May '07

11 May '07

On 5/10/07, Tom Spear <speeddymon(a)gmail.com> wrote: > MAX_STRING_LEN is defined in resource.h at 255. According to > http://msdn2.microsoft.com/en-us/library/ms724872.aspx the limit for > the length of a subkey name is 255. > > >> Please let me know if I should make a new define instead of using > MAX_STRING_LEN. > Yes, you should make a new define. MAX_STRING_LEN is the maximum length for a string from the resource file, which is totally unrelated to the limit for the length of a subkey name. -- James Hawkins

8 8

re: kernel32: Implement most of CopyFileEx
by Dan Kegel 11 May '07

11 May '07

Nice to see somebody filling in the gaps like this. A few nits based on a superficial reading: CopyFileA leaks a string. (I know, it did before your change, too.) copy_file_open_dest's interface comment has the wrong name. + for(i = 0; i < sizeof(flags) / sizeof(flags[0]); i++) + { + if(for_write) { Please use the same whitespace conventions as in the rest of the file (in particular, leave a space after keywords like for and if). + if ((h1 = copy_file_open_source(source, copyFlags & COPY_FILE_OPEN_SOURCE_FOR_WRITE)) == INVALID_HANDLE_VALUE) That line's too long... try to keep them under 80 chars. + // FIXME: total_file_size should include the sum of all streams Use C comments, not C++ comments. + DWORD result = progressRoutine(total_file_size, total_bytes_transferred, stream_size, stream_bytes_transferred, + stream_number, CALLBACK_STREAM_SWITCH, h1, h2, appData); Another instance of line longer than 80 chars. (You have several.) You don't have a conformance test, nor was there one for CopyFile. Maybe you should consider writing one, and making sure it passes both on Wine and Windows. Say, what app is this for? - Dan

4 5

Re: wined3d: Add hardware cursor support.
by Stefan Dösinger 11 May '07

11 May '07

Am Freitag 11 Mai 2007 02:34 schrieb Andrew Riedi: > I made a test to figure out the exact Windows behavior, and Stefand helped > me out in telling me what to test. This code behaves how it should. I > will send the test in later once I have cleaned it up. > > changelog: wined3d: Add hardware cursor support. > + if (This->hardwareCursor) DestroyCursor(This->hardwareCursor); > + if (This->haveHardwareCursor) This->haveHardwareCursor = FALSE; Is there a specific reason to test if This->haveHardwareCursor is true before setting it to false? Or would a plain This->haveHardwareCursor = FALSE work too? Otherwise looked fine to me

1 0