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

Re: Disabling File attachments on Wiki
by Dmitry Timoshkov 30 Apr '09

30 Apr '09

[please don't forget to cc: wine-devel] "Jeff Latimer" <lats(a)yless4u.com.au> wrote: >>> It would be also helpful to to get all wiki user account list (check it >>> time from time for suspicious names) and have an ability to delete >>> accounts >>> of known spammers for Wiki admins. Is it possible? >> Another thing that would probably help to get rid of spammers' accounts >> is to (regularly) delete all user accounts without a personal page. > It would also get rid of anyone who did not want to create a home page. Then it's time to do that, with a short description what your intentions/ activity is in the Wine project. > I just created one /just in case/ but really don't want to. What's the point in having a Wiki account then? -- Dmitry.

9 28

Re: SoC: DirectShow/Gstreamer
by Maarten Lankhorst 09 Apr '09

09 Apr '09

Hello Trevor, I think this is a great idea. I have no idea how gstreamer works, but a lot of quartz is automatic pulling the pipes etc together. There are 3 categories of filters in directshow: - Source filters (can be a file, can be a webcam, can be any filter that has only an output) - Intermediate filters (can decode or parse a file, multiple can be linked together) - Renderers I have only peeked at gstreamer, but it seems it is very similar. Wine's directshow however still misses a few things that are probably important, like sending information about when new segments start, etc. If you implemented a gstreamer codec, a big part of it could be to fix this to be more similar to what windows does. You will probably find graphedt.exe from the microsoft sdk useful, it used to be part of the directx sdk, but later I think it became part of the platform sdk. It will probably still run, but the save/load function for saving/restoring filtergraphs is missing. However last time I looked most other functionality from it worked, if not patches should be easy to write. Also some older directx sdk's had in the extras a debug version of windows' quartz dll which is very useful if you want to test your stuff against windows' quartz. It would output extra debug information which you could use to debug stuff. I would have to do some digging to find it back again. I have also made a stand-alone codec for schroedinger last year, which falls in the 'intermediate filters' category. It's source is available at: http://www.diracvideo.org/git?p=direct-schro.git;a=tree;h=HEAD This is a very basic filter based on wine's directshow code and using the schroedinger open source codec to do the decoding and encoding (they're separate filters). In encoder.c there is also a beginning of a properties dialog, which might be useful if you need one for the gstreamer filter. Cheers, Maarten.

3 3

Single login for Wine sites?
by David Gerard 08 Apr '09

08 Apr '09

What's standing in the way of unified login for the various Wine sites (appdb, wiki, Bugzilla)? Is there anyone running any of these sites who doesn't consider single unified login a good idea? - d.

5 6

Re: [1/5] resend include: bring in6_addr into line with the MS definition
by Alexandre Julliard 02 Apr '09

02 Apr '09

Jeff Latimer <lats(a)yless4u.com.au> writes: > diff --git a/include/ws2tcpip.h b/include/ws2tcpip.h > index a38ccda..9ddb3d1 100644 > --- a/include/ws2tcpip.h > +++ b/include/ws2tcpip.h > @@ -85,26 +85,41 @@ struct WS(ip_msfilter) { > struct WS(in_addr) imsf_slist[1]; > }; > > -typedef struct WS(in_addr6) > -{ > - WS(u_char) s6_addr[16]; /* IPv6 address */ > +struct WS(in6_addr) { > + union { > + WS(u_char) Byte[16]; > + WS(u_short) Word[8]; > + } u; > } IN6_ADDR, *PIN6_ADDR, *LPIN6_ADDR; > > +#define in_addr6 WS(in6_addr) > + > +#define _S6_un u > +#define _S6_u8 Byte > +#ifndef USE_WS_PREFIX > +#define s6_addr _S6_un._S6_u8 > +#else > +#define WS_s6_addr _S6_un._S6_u8 > +#endif > + > +#define s6_bytes u.Byte > +#define s6_words u.Word This doesn't seem to match the PSDK, I don't see anything like this in ws2tcpip.h. -- Alexandre Julliard julliard(a)winehq.org

3 4

Re: mciqtz32: Implement driver messages
by Detlef Riekenberg 01 Apr '09

01 Apr '09

On Sa, 2009-03-28 at 11:48 +0100, Christian Costa wrote: Thanks for working on this part of Wine. > +/************************************************************************** > + * MCIAVI_drvClose [internal] > + */ > +static DWORD MCIQTZ_drvClose(DWORD dwDevID) > +{ With greetings from copy & paste: There is no need to fix the MCIAVI != MCIQTZ mismatch in a second Patch. Fix it here, when you add the code. You use TAB in the comments and also in the Code. Don't do that. -- By by ... Detlef

2 5

Re: GetTempFileName should return 0 when the path isn't valid
by Austin English 31 Mar '09

31 Mar '09

On Sun, Mar 29, 2009 at 11:37 AM, Oskar Eisemuth <patchnow(a)gmail.com> wrote: > GetTempFileName should check the path and should return zero if it isn't a directory. > See Bug 17875 > > --- > dlls/kernel32/path.c | 10 ++++++++++ > 1 files changed, 10 insertions(+), 0 deletions(-) > > > > > Can you add a testcase for this? -- -Austin

3 3

Re: Implement [Un]RegisterHotKey with test
by Detlef Riekenberg 31 Mar '09

31 Mar '09

On Mo, 2009-03-30 at 18:17 +0500, Vladimir Pankratov wrote: > + return DefWindowProcW (hWnd, msg, wParam, lParam); > + WCHAR className[] = {'T','e','s','t','H','o','t','K','e','y',0}; > + WCHAR winName[] = {'H','o','t','K','e','y','T','e','s','t',0}; > + ATOM classatom; > + WNDCLASSW cls; > + cls.hIcon = LoadIconW (0, (LPWSTR)IDI_APPLICATION); > + cls.hCursor = LoadCursorW (0, (LPWSTR)IDC_ARROW); .... Next to the other comments, beware that UNICODE Functions are missing or not implemented in win9x. Please use ANSI Functions in the tests. -- By by ... Detlef

1 0

Re: Implement [Un]RegisterHotKey with test
by Vincent Povirk 31 Mar '09

31 Mar '09

I've been told the functionality needs to go in the server. You should be using git format-patch to create your patches (see http://winehq.org/sending_patches) You're mixing tabs and spaces. Please don't do that. Vincent Povirk 2009/3/30 Vladimir Pankratov <scriptkid(a)mail.ru>: > Hello. > > Based on Vincent Povirk's idea and patch at > http://www.winehq.com/hypermail/wine-devel/2003/02/0636.html. > > This implement RegisterHotKey and UnregisterHotKey API functions, and test. > > Thank you. > > > > >

1 0

GSoC Mentors, please start looking at the student proposals
by Kai Blin 31 Mar '09

31 Mar '09

Hi folks, the student application period is nearing it's end and the proposals are starting to come in. From what I heard the new web-app doesn't support changing the application after the deadline on April 3rd, so I'd like mentors to give as much feedback as possible before the deadline. Also, please indicate if you're willing to mentor specific proposals. Thanks, Kai -- Kai Blin WorldForge developer http://www.worldforge.org/ Wine developer http://wiki.winehq.org/KaiBlin Samba team member http://www.samba.org/samba/team/ -- Will code for cotton.

1 0

Tracing weirdness in the kernel32:profile tests
by Paul Vriens 31 Mar '09

31 Mar '09

Hi, I'm trying to fix the remaining issues in the kernel32 profile tests. They have to do with the fact that we have extra '\r\n' on Win9x and that we need a timeout. The timeout is already in (1000 msec) but apparently that is not even enough on some boxes. If I check the current results on one of my boxes I see: profile.c:977: Test failed: File doesn't match profile.c:992: Test failed: File doesn't match profile.c:1007: Test failed: File doesn't match profile.c:1019: Test failed: File doesn't match This is exactly at the place where the ok() message is. For debugging purposes I added if (lstrcmpA(buf, data)) { trace("buf : (%s)\n", buf); trace("data : (%s)\n", data); } to check_file_data(). When I now run the test I get (and I omitted the buf/data output here): profile.c:835: Test failed: File doesn't match profile.c:835: Test failed: File doesn't match profile.c:835: Test failed: File doesn't match profile.c:835: Test failed: File doesn't match So the linenumbers are not the ones from the original ok() message but instead the ones from the last trace() message. Any idea? -- Cheers, Paul.

2 2