September 2011 - Wine-devel

Re: comctl32: Property sheets must be able to redraw themselves even if no page is selected.
by Marvin 05 Sep '11

05 Sep '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=13992 Your paranoid android. === WVISTAADM (32 bit propsheet) === Failure running script in VM: The specified guest user must be logged in interactively to perform this operation

1 0

Undocumented registry settings
by Vitaliy Margolen 05 Sep '11

05 Sep '11

Henri, Would you mind documenting recently added registry entries to wined3d on Wiki page http://wiki.winehq.org/UsefulRegistryKeys ? I can't find description of "AlwaysOffscreen". Vitaliy.

2 2

Re: cmd: Avoid rereading batch file for every call/goto executed (try 2)
by Henri Verbeet 05 Sep '11

05 Sep '11

2011/9/5 Frédéric Delanoy <frederic.delanoy(a)gmail.com>: > + keys = HeapAlloc(GetProcessHeap(), 0, capacity * sizeof(WCHAR*)); > + vals = HeapAlloc(GetProcessHeap(), 0, capacity * sizeof(LARGE_INTEGER)); ... > + keys = HeapReAlloc(GetProcessHeap(), 0, keys, > + capacity * sizeof(WCHAR*)); > + vals = HeapReAlloc(GetProcessHeap(), 0, vals, > + capacity * sizeof(LARGE_INTEGER)); Memory allocations can fail. You also never free these, but I suppose that happens implicitly on process termination.

2 1

Re: msxml3: Correct MSXML Object Safety tests (try 2)
by Alexandre Julliard 05 Sep '11

05 Sep '11

Alistair Leslie-Hughes <leslie_alistair(a)hotmail.com> writes: > Hi, > Add the enabled test as well. > > Changelog: > msxml3: Correct Object Safety tests It doesn't work here: ../../../tools/runtest -q -P wine -M msxml3.dll -T ../../.. -p msxml3_test.exe.so domdoc.c && touch domdoc.ok domdoc.c:4706: Test failed: expected INTERFACESAFE_FOR_UNTRUSTED_CALLER | INTERFACESAFE_FOR_UNTRUSTED_DATA got 2 domdoc.c:7532: Test failed: expected INTERFACESAFE_FOR_UNTRUSTED_CALLER | INTERFACESAFE_FOR_UNTRUSTED_DATA got 2 make[1]: *** [domdoc.ok] Error 2 -- Alexandre Julliard julliard(a)winehq.org

1 0

Re: ole32/tests: Test for correct storage creation.
by Vincent Povirk 05 Sep '11

05 Sep '11

You might just want to try reopening the file and checking its CLSID, instead of directly checking the contents. That should make it more obvious that something is seriously wrong. With the direct checks, it's hard to figure out what the test actually means. Anyway, this should be easy to fix. We just need to keep track of whether the file has ever had a header, instead of assuming the header is valid if the file is large enough to have one.

1 0

Re: [1/2] programs: Add a stub implementation of regasm.exe.
by Vincent Povirk 05 Sep '11

05 Sep '11

On Windows, this is a .NET assembly. We should probably do it the same way.

1 0

Re: [dsound] InstanceId holds a pointer, so it needs to be as wide as a pointer.
by Michael Stefaniuc 05 Sep '11

05 Sep '11

Hello Dan, Dan Kegel wrote: > Not that we do anything with it... still the correct fix would be to rewrite the driver to use a more modern header version than DirectX 5. I guess that's part of Andrew's dsound rewrite. bye michael

1 0

re: configure.ac: add -Werror to default compiler options
by Dan Kegel 05 Sep '11

05 Sep '11

(Reviving http://www.winehq.org/pipermail/wine-devel/2008-September/069266.html ) After http://www.winehq.org/pipermail/wine-patches/2011-August/106150.html and http://www.winehq.org/pipermail/wine-patches/2011-August/106151.html, configuring with -Wall -Werror gives the right results, judging by a quick diff of config.h and config.status. (sys/asoundlib.h is not detected, but that's ok, it's marked as obsolete, and wine uses alsa/asoundlib.h instead.) This takes care of Maarten's concerns from the earlier thread. Then building yields only a tiny handful of errors: dllfunc.c: In function ‘AMovieDllRegisterServer2’: dllfunc.c:180: warning: dereferencing type-punned pointer will break strict-aliasing rules qualitycontrol.c: In function ‘QualityControlImpl_Notify’: qualitycontrol.c:70: warning: dereferencing type-punned pointer will break strict-aliasing rules transform.c: In function ‘TransformFilterImpl_QueryInterface’: transform.c:248: warning: dereferencing type-punned pointer will break strict-aliasing rules ../../include/winternl.h: In function ‘get_vm86_teb_info’: ../../include/winternl.h:2660: warning: dereferencing type-punned pointer will break strict-aliasing rules ... register.c: In function ‘create_registrar’: register.c:66: warning: dereferencing type-punned pointer will break strict-aliasing rules Those are potential, if unlikely, bugs; we ought to either fix them or add -fno-strict-aliasing to CFLAGS for the seven or so affected .c files. At which point it would probably be a fine idea to add -Werror by default; buildbot will help keep everyone in sync, even if they're using a compiler that doesn't catch as many warnings as the one buildbot uses. - Dan

7 9

Re: [1/4] urlmon: Use CoTaskMemFree instead of HeapFree for memory allocated with CoTaskMemAlloc
by Jacek Caban 05 Sep '11

05 Sep '11

Hi Thomas, AFAICS secur_url wasn't allocated with CoTaskMemAlloc until this patch. Why are you changing it? Jacek

2 1

Re: [PATCH] cmd: Avoid rereading batch file for every call/goto executed (try 3)
by Eric Pouech 05 Sep '11

05 Sep '11

this won't work if a .bat file calls another .bat file the cache must be stored inside the context structure... A+ -- Eric Pouech

1 0