August 2008 - Wine-Devel - list.winehq.org

Re: [msi/tests 3/3] Create a system restore point when needed
by James Hawkins Aug. 31, 2008

Aug. 31, 2008

On Sat, Aug 30, 2008 at 3:14 PM, Paul Vriens <paul.vriens.wine(a)gmail.com> wrote: > Hi, > > The slowness on some machines for the install tests comes from the system > restore facility. When 'System Restore' is active you will find a great > number > of system restore points after these tests have run. Adding our own system > restore point circumvents this as Windows will now wait till we tell it that > the > system change has ended. Only 1 system restore point will be created now > and > this one is nicely removed afterwards. > > On my testboxes (XP and Vista) this makes sure that the tests run within the > limits of the timeout (2 minutes). Before this patch the tests would take > about > 8 minutes on my Vista box (now 1.something). > > Changelog > Create a system restore point when needed > Please factor all of this out into functions. + /* Create a system restore point. We need to do this on a system where 'System + * Restore' is active as we would otherwise timeout on these tests. + * By doing so we circumvent the multitude of restore points that would have been + * created by all the installation and removal tests. + * + * The restore point is removed afterwards. + * + * On systems where 'System Restore' is not active, this won't have any effect + * as the first call to SRSetRestorePointA will fail with a status of + * ERROR_SERVICE_DISABLED. + */ Does the comment really need to be that long? -- James Hawkins

2 1

Re: ntdll: add some tests for NtDeleteFile
by Vitaliy Margolen Aug. 31, 2008

Aug. 31, 2008

Jeff Zaroyko wrote: > > > + static const WCHAR testdirW[] = {'n','t','d','e','l','e','t','e','f','i','l','e',0}; > + static const WCHAR subdirW[] = {'\\','s','u','b',0}; > + > + ret = CreateDirectoryW(testdirW, NULL); You need to create test files and directories in windows temp dir. > + pRtlDosPathNameToNtPathName_U(testdirW, &nameW, NULL, NULL); You are leaking unicode string nameW. You need to call this to free it: RtlFreeUnicodeString( &nameW ); > + ret = pNtDeleteFile(&attr); > + ok(ret == STATUS_SUCCESS, "expected NtDeleteFile to ret STATUS_SUCCESS on failure\n"); How do you know it's a failure? You should check if dir is still there. Vitaliy

2 1

Solution to msi/install test timeouts
by Paul Vriens Aug. 31, 2008

Aug. 31, 2008

Hi, I've been experimenting on my XP and Vista boxes and the only way to make the tests run fast is by stopping 'System Restore'. It makes sense that this has a huge impact on performance, especially with the big number of installations we are doing in these tests. Could people that are currently seeing timeouts for their boxes on test.winehq.org check if 'System Restore' is indeed activated? -- Cheers, Paul.

2 1

Re: ntdll/file.c: Allow to TRACE NtCreateFile return
by Dmitry Timoshkov Aug. 30, 2008

Aug. 30, 2008

"Markus Hitter" <mah(a)jump-ing.de> wrote: > Providing the file handle allows to map > read/write requests to the corresponding file name. As pointed out by Alexander, you can use an appropriate debug channel for that, +relay or +server. There is no need to pollute the source with additional traces. -- Dmitry.

5 8

Help for my first patch
by Michael Bonfils Aug. 30, 2008

Aug. 30, 2008

Hi, I'm using Wine with xUbuntu and you may be know, the menu and toolbar of xfce are often visible. So I've looked at xfce forum and saw that wine doesn't inform xfce that window is fullscreen... At http://standards.freedesktop.org/wm-spec/wm-spec-latest.html, I've found a _NET_WM_STATE_FULLSCREEN flag. After digging in wine source, I've found in dlls/winex11.drv/window.c the function create_whole_window (I think that this function is called when a game want a resolution which isn't supported by X, wine used this function to fake fullscreen ?, correct me I'm wrong) If in set_wm_hints, I force the window_type to _NET_WM_STATE_FULLSCREEN, toolbar and menu of xfce disappears. (tested with Wizardry 8) My problem is how detect in this function that I came from create_whole_window and not from others functions ? Do I have to test size of window to current resolution and use this flag ? Or call XChangeProperty in create_whole_window ? Thanks in advance, Michael

2 1

Re: [try4] d3d9: Set the initial scissorrect ...
by Rico Schüller Aug. 30, 2008

Aug. 30, 2008

Rico Schüller schrieb: > Hi, > > this patch adds the initial scissorrect for d3d9. The improvement from > the last patch is, that it doesn't try to set it for d3d8 and ddraw > because both don't support scissorrects. > > Cheers > Rico > ------------------------------------------------------------------------ > > This patch is wrong!

1 0

What does patchwatcher do?
by Rico Schüller Aug. 30, 2008

Aug. 30, 2008

Hi, some days ago I've send a patch which broke some other tests, see http://www.kegel.com/wine/patchwatcher/results/274.txt . But this test got a "conformance test ok". I think patchwatcher should also mark patches as bad when they let crash other tests. Cheers Rico

2 3

Re: [try4] d3d9: Set the initial scissorrect ...
by Henri Verbeet Aug. 30, 2008

Aug. 30, 2008

2008/8/30 Rico Schüller <kgbricola(a)web.de>: > Hi, > > this patch adds the initial scissorrect for d3d9. The improvement from the > last patch is, that it doesn't try to set it for d3d8 and ddraw because both > don't support scissorrects. > > Cheers > Rico > I'd rather keep that in wined3d for now.

2 1

Re: [Patch] 0001-Support-for-winscard.dll.patch
by Mounir IDRASSI Aug. 30, 2008

Aug. 30, 2008

Hi All, I have just been informed by one of my contacts that this discussion about Winscard has started. As the original author of Winscard for Wine, I'd like to inform you that I have already posted a fully functional Winscard patch for Wine on April 2007. Here is the link to it : http://www.winehq.org/pipermail/wine-devel/2007-April/056432.html Afterwards, my complete patch was not accepted because I was relying on the fact that alignment of the Winscard structures coming from the WIN32 world is the same as the alignment expected by pcsclite on the system on which Wine is running. My assumption is true for Linux 32Bit but I clearly can't guarantee that on other platforms. Anyway, my implementation is running production systems on two of my customers under Linux 32bit (Ubuntu, CCID readers) and it has been tested by many users who contacted me directly to provide them with the full featured winscard version. For that, I distribute the winscard binary for 32-bit Linux systems under the following link : http://www.idrix.fr/Root/SCard4Wine/winscard.tar.gz I think it would be more efficient if your effort can benefit from my patch and all the experience I have put on it. The source from my side has evolved a little bit but it remains close to the original posting. I hope there will room for collaboration of this to provide a maximum number of Wine users with smart cards functionalities. Cheers, -- Mounir IDRASSI IDRIX http://www.idrix.fr

2 2

Patchwatcher offline
by Dan Kegel Aug. 30, 2008

Aug. 30, 2008

The patchwatcher robot is offline today while I move it to a faster computer and add a timeout to individual tests (since one hangs reliably on the new computer). Here's the timeout patch, if anyone's interested in doing a little prereview. I finally realized the most windows-y way of doing a timeout was to run the test in its own thread. include/wine/test.h | 26 +++++++++++++++++++++++--- 1 file changed, 23 insertions(+), 3 deletions(-)

1 0