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

Re: mciqtz32: Implement driver messages

by Christian Costa

You're right. Not good for git blame. I will ask to Alexandre. Thanks. By, Christian Detlef Riekenberg a écrit : > On So, 2009-03-29 at 18:12 +0200, Christian Costa wrote: > >> That said, I don't understand then why we don't just convert wine's >> code (in >> several steps off course) and avoid troubles particularity switch from >> one >> style to another depending on the dll we are working on. >> > > Changing only the indention is not desired, as the history is falsified > (Original Author and date/time of the the last code-change). > Expect patches with unrelated indention changes to be rejected. > > >> I recently implemented d3dxof using a 2 spaces indentation style. >> I don't know either how this code could have gone into git. :-| >> > > Parts of the old code use 4 SPACE, other parts use 2 SPACE, but the > Styple was not enforced tha hard in the old days. > > >> So next time I will convert d3dxof to 4 spaces and see what happens >> to >> the patch. ;-) >> > > I suggest to ask Julliard before you start with that. > He might exceptional accept that style changes only Patch, > as you did the major part of the code in that dll. > > >

March 31, 2009

Re: GetTempFileName should return 0 when the path isn't valid

by Oskar Eisemuth

Hello The patch "kernel32: GetTempFileName should return 0 when the path isn't valid, including test" has the necessary tests, anything I missed? -- Oskar

March 31, 2009

Re: Implement [Un]RegisterHotKey with test

by Detlef Riekenberg

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

March 31, 2009

Re: Implement [Un]RegisterHotKey with test

by Vincent Povirk

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. > > > > >

March 31, 2009

GSoC Mentors, please start looking at the student proposals

by Kai Blin

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.

March 31, 2009

Re: Tracing weirdness in the kernel32:profile tests

by Paul Vriens

Michael Karcher wrote: > Am Dienstag, den 31.03.2009, 09:42 +0200 schrieb Paul Vriens: > >> 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? > > Expected behaviour. You must not call trace from within the argument > list of an ok() statement. ok()/trace() first sets a global variable to > the current location, and then calls a printing function passing and > evaluating the parameters. If one of them changes the global variable > already set, the location printed is wrong, as you see here. > > Regards, > Michael Karcher > > Thanks for clearing that up. -- Cheers, Paul.

March 31, 2009

Re: Tracing weirdness in the kernel32:profile tests

by Michael Karcher

Am Dienstag, den 31.03.2009, 09:42 +0200 schrieb Paul Vriens: > 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? Expected behaviour. You must not call trace from within the argument list of an ok() statement. ok()/trace() first sets a global variable to the current location, and then calls a printing function passing and evaluating the parameters. If one of them changes the global variable already set, the location printed is wrong, as you see here. Regards, Michael Karcher

March 31, 2009

Re: [1/2](try 2)advapi32: alias some sids in LookupAccountName and make comparison case insensitive

by Paul Vriens

Aric Stewart wrote: > > with tests > --- > dlls/advapi32/security.c | 10 +++++-- > dlls/advapi32/tests/security.c | 56 > ++++++++++++++++++++++++++++++++++++++++ > 2 files changed, 63 insertions(+), 3 deletions(-) > > > > ------------------------------------------------------------------------ > > Hi Aric, Could you have a look at the test results as you are still busy with this stuff. - Some (all?) of the new tests fail on a Non-English locale. - None of the newly checked names are available on NT4. - LocalService and NetworkService are not available on W2K (all?) -- Cheers, Paul.

March 31, 2009

Tracing weirdness in the kernel32:profile tests

by Paul Vriens

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.

March 31, 2009

httpapi compile broken on OpenBSD

by Austin English

Howdy, The compile is broken on OpenBSD in httpapi/httpapi_main.c: gcc -c -I. -I. -I../../include -I../../include -D__WINESRC__ -D_REENTRANT -fPIC -Wall -pipe -fno-strict-aliasing -Wwrite-strings -Wpointer-arith -I/usr/local/include -I/usr/local/include -I/usr/local/include/libpng -o httpapi_main.o httpapi_main.c In file included from ../../include/http.h:25, from httpapi_main.c:27: ../../include/ws2tcpip.h:60: error: conflicting types for `socklen_t' /usr/include/sys/types.h:163: error: previous declaration of `socklen_t' *** Error code 1 Stop in /home/austin/wine-git/dlls/httpapi (line 92 of /usr/share/mk/sys.mk). Talking with Vincent, narrowed it down to the file ignoring Wine's definition of socklen_t. Is there any reason we shouldn't be defining it, like we already do in iphlpapi/iphlpapi_main.c, winhttp/net.c, wininet/netconnection.c, wsock32/protocol.c and server/sock.c? E.g.,: diff --git a/dlls/httpapi/httpapi_main.c b/dlls/httpapi/httpapi_main.c index 7b29a15..b96f730 100644 --- a/dlls/httpapi/httpapi_main.c +++ b/dlls/httpapi/httpapi_main.c @@ -18,6 +18,8 @@ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA */ +#define USE_WS_PREFIX + #include "config.h" #include <stdarg.h> -- -Austin

March 31, 2009