February 2011 - Wine-devel - list.winehq.org

Killing WineAPI project on SourceForge.
by Max TenEyck Woodbury 07 Feb '11

07 Feb '11

No feedback and apparently no interest in the WineAPI project on SourceForge. Understandable since there is now a WineAPI on the regular Wine Wiki. Unless someone objects, I'm going to ask that it be shutdown on SourceForge. Max

5 6

Including winegcc-built shared library in Debian binary package
by Robert McDonald 07 Feb '11

07 Feb '11

I've got a Debian binary package configured. A wrapper script is installed in /usr/bin that invokes wine on the winegcc-built shared library that implements the program. The shared library is installed in /usr/lib and then ldconfig is run during the installation process. The shared library build product seems to be a special case because it is built by winegcc and referenced on the wine command line when run. I've got questions about this even though the product works OK as packaged. Is there a more appropriate installation directory than /usr/lib for the shared library (although lintian complains if I don't put it in /usr/lib)? Does ldconfig need to be run after installing/uninstalling the shared library (other than to satisfy lintian complaints if ldconfig isn't run from postinst/postrm)? Is there any reason to bother with version numbering the shared library (other than using one version number for all library releases just to keep lintian happy by having the version number) since the library is prepared by winegcc and will only be loaded by wine? To look for answers, I've done the following: * searched the web * searched the WineHQ web site and read the FAQ there * scanned the subject lines of a couple of months of this mailing list * read http://wiki.winehq.org/Packaging * used https://wiki.ubuntu.com/PackagingGuide/Complete for some packaging instructions * read http://www.bytehold.com/index.php/projects/176-debian-binary-package-and-re… for general instructions in setting up a Debian binary package * read http://www.pathname.com/fhs/ for recommendations on locating files URLs to relevant resources would be welcome. -- Robert McDonald BitTorrent, Inc.

2 1

Re: iphlpapi: Changed return value of NotifyAddrChange() stub and added test cases.
by André Hentschel 07 Feb '11

07 Feb '11

Hi, in todo_wine blocks only ok() calls should be made. So consider splitting them down to "todo_wine ok(..." for each. > -CancelIPChangeNotify > CreateProxyArpEntry > DeleteIPAddress > DeleteProxyArpEntry > EnableRouter > FlushIpNetTable > GetAdapterIndex > -NotifyAddrChange > -NotifyRouteChange > +NotifyRouteChange + CancelIPChangeNotify this change regarding CancelIPChangeNotify is also not needed. -- Best Regards, André Hentschel

2 1

Re: oleaut32: Fix copying an array of 0 elements
by Marvin 07 Feb '11

07 Feb '11

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 http://testbot.winehq.org/JobDetails.pl?Key=8987 Your paranoid android. === WNT4WSSP6 (32 bit safearray) === safearray: unhandled exception c0000005 at 0040E5A5 === W2KPROSP4 (32 bit safearray) === safearray: unhandled exception c0000005 at 0040E5A5 === WXPPROSP3 (32 bit safearray) === safearray: unhandled exception c0000005 at 0040E5A5 === W2K3R2SESP2 (32 bit safearray) === safearray: unhandled exception c0000005 at 0040E5A5 === WVISTAADM (32 bit safearray) === safearray: unhandled exception c0000005 at 0040E5A5 === W2K8SE (32 bit safearray) === safearray: unhandled exception c0000005 at 0040E5A5 === W7PRO (32 bit safearray) === safearray: unhandled exception c0000005 at 0040E5A5 === W7PROX64 (32 bit safearray) === safearray: unhandled exception c0000005 at 0040E5A5 === W7PROX64 (64 bit safearray) === Timeout

1 0

Patch to fix ntdll:exception tests on Mac OS
by Charles Davis 07 Feb '11

07 Feb '11

Hi, This is a small patch I wrote to fix the exception tests on Mac OS X. Right now, if you look at the Mac results I submitted, you'll notice that there are two failures. The first failure happens because Mach doesn't allow the 'global' flag to be set when setting the debug registers. Changing the test not to set the global flag causes the test to pass. I get the feeling, though, that there is a good reason the global flag was set, so I'm reluctant to send that change in. The other failure happens because 32-bit Mac OS only protects stacks from execution. If you clear PROT_EXEC on any other data, you can still execute code from it. (64-bit Mac OS, on the other hand, always respects the PROT_EXEC flag.) I changed ntdll on 32-bit Mac OS to report DEP as permanently disabled, but I wonder if this was the right change to make. I just want to know: am I doing the right thing here? If not, what should I do instead to fix the tests? Chip

1 0

Re: user32: Don't try to move mouse to (-1,0).
by Vitaliy Margolen 06 Feb '11

06 Feb '11

On 02/05/2011 07:10 AM, Jacek Caban wrote: > > This trigers screen edge action on KDE, which is 'Present Windows' in > default configuration. mouse_event() generates relative mouse movement events unless MOUSEEVENTF_ABSOLUTE is specified. If you don't want a pointer to get near the edge then you should use SetCursorPos() instead. Vitaliy.

2 1

re: Wine FAQ edits
by Dan Kegel 06 Feb '11

06 Feb '11

Hi Albert, http://wiki.winehq.org/FAQ?action=diff&rev2=384&rev1=383 looks like it was a pretty big edit. Try smaller ones, and try to do just one kind of change per edit.

7 20

re: Pulling Patch
by Dan Kegel 06 Feb '11

06 Feb '11

James McKenzie wrote: > Since my Mac is dying I have decided to return to the Windows world. > > Please pull any and all patches. I have envoked the right to copyright > and none of my code can or will be used in Wine. Sorry to see you go, but... why would you want to prohibit people from using your patches? (I only see two typo changes committed: http://source.winehq.org/git/wine.git/?a=search&h=HEAD&st=author&s=James+Mc… Presumably you don't mean those, since they were trivial comment changes.) I guess you mean patches like http://www.winehq.org/pipermail/wine-patches/2010-June/089537.html When you sent those to wine-patches, weren't you placing them under the LGPL? - Dan

4 5

RE: shell32/tests: Clean make new folder test (Valgrind)
by Michael Mc Donnell 05 Feb '11

05 Feb '11

Hi! What is the reason this patch by Nikolay Sivov was not included in the latest version of wine? It fixes a clean-up bug on some Windows XP machines, and tidies up the code nicely. Wine-patches archive link: http://www.winehq.org/pipermail/wine-patches/2011-January/098342.html Regards, Michael Mc Donnell P.S. Thanks Nikolay for fixing my first win32 test.

1 0

Re: user32: Don't try to move mouse to (-1,0).
by Marvin 05 Feb '11

05 Feb '11

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 http://testbot.winehq.org/JobDetails.pl?Key=8925 Your paranoid android. === W7PROX64 (64 bit msg) === Timeout

1 0