February 2015 - Wine-Devel - list.winehq.org

My Ideas for GSoC 2015

by Zhenbo Li

Hello, everyone I'm Zhenbo Li, a college student who participated in GSoC 2014 for mshtml.dll. Here are some ideas for GSoC 2015. 1. IDropTargetHelper ================== This interface is to display images [1]. I'm sure that it is necessary for VPT[2], and I guess implementing this interface will contribute to programs like 7zip. Although IDropTargetHelper is a useful interface, I haven't found a proper demo(most demos depend on MFC). So I'm writing a simple demo for this interface[3], with no MFC or .NET code. 2. IXMLHttpRequest ================== XMLHttpRequest is used heavily in AJAX programming.[4] Wine has implemented most of it in msxml component, but it's a stub in mshtml. Most js code doesn't treat wine IE as IE6 or earlier, so they will not choost ActiveX[5] to work around it(like Firebug Lite[6]) 3. sessionStorage =============== IHTMLStorage[7] is used widely, including Firebug Lite[6]. A well-implemented IHTMLStorage interface should support both sessionStorage and localStorage. I tried to write a hack, but found that it may be necessary to adjust wine-gecko code to support that(Thank Jacek). Therefore, implementing sessionStorage can be the first step of implementing localStorage, and it can be an example for WebIDL binding for further development. These are my humble ideas for GSoC 2015, and I'm glad to choose one of them as my proposal, if they meet the criteria for GSoC. Also, it is my honor if my ideas make a positive effect on other GSoC students. Thank you for reading this mail. I'm looking forward to your feedback. [1]: https://msdn.microsoft.com/en-us/library/windows/desktop/bb762028%28v=vs.85… [2]: http://virtualplaytable.com/ [3]: https://gitcafe.com/endle/IDrop_Helper_Example [4]: https://developer.mozilla.org/en-US/docs/Web/API/XMLHttpRequest [5]: http://en.wikipedia.org/wiki/XMLHttpRequest#Support_in_Internet_Explorer_ve… [6]: https://getfirebug.com/firebuglite [7]: https://msdn.microsoft.com/en-us/library/cc288313%28v=vs.85%29.aspx

Feb. 28, 2015

Re: advapi32 redefines SECURITY_INFORMATION

by Nikolay Sivov

On 28.02.2015 13:33, Francois Gouget wrote: > --- > > So svcctl.idl needs the SECURITY_INFORMATION definition and it's not > available from any other idl file. This results in this type being > defined twice when compiling dlls/advapi32/service.c: once in winnt.h > and then again in dlls/advapi32/svcctl.h. > > We normally avoid this type of situation. An easy fix is to introduce > the usual _XXX_DEFINED macro check but this seems to be a first in an > idl file. Is there a better way? > I think I screwed up it a little, does it work if you place it inside #if 0, after this: --- cpp_quote("#if 0 /* already defined in winsvc.h */") --- I think we only need it for widl, and later advapi32 or services should include winnt.h in usual way.

Feb. 28, 2015

advapi32 redefines SECURITY_INFORMATION

by Francois Gouget

--- So svcctl.idl needs the SECURITY_INFORMATION definition and it's not available from any other idl file. This results in this type being defined twice when compiling dlls/advapi32/service.c: once in winnt.h and then again in dlls/advapi32/svcctl.h. We normally avoid this type of situation. An easy fix is to introduce the usual _XXX_DEFINED macro check but this seems to be a first in an idl file. Is there a better way? include/wine/svcctl.idl | 3 +++ include/winnt.h | 3 +++ 2 files changed, 6 insertions(+) diff --git a/include/wine/svcctl.idl b/include/wine/svcctl.idl index e8463b5..1c3d22f 100644 --- a/include/wine/svcctl.idl +++ b/include/wine/svcctl.idl @@ -65,7 +65,10 @@ interface svcctl typedef [context_handle] void *SC_RPC_LOCK; typedef [context_handle] void *SC_NOTIFY_RPC_HANDLE; + cpp_quote("#ifndef _SECURITY_INFORMATION_DEFINED") + cpp_quote("#define _SECURITY_INFORMATION_DEFINED") typedef DWORD SECURITY_INFORMATION; + cpp_quote("#endif") /* undocumented access rights */ cpp_quote("#define SERVICE_SET_STATUS 0x8000") diff --git a/include/winnt.h b/include/winnt.h index 4b06b2c..0a4185d 100644 --- a/include/winnt.h +++ b/include/winnt.h @@ -3971,7 +3971,10 @@ static const WCHAR SE_CREATE_GLOBAL_NAME[] = { 'S','e','C','r','e','a','t','e',' #define SE_RM_CONTROL_VALID 0x00004000 #define SE_SELF_RELATIVE 0x00008000 +#ifndef _SECURITY_INFORMATION_DEFINED +#define _SECURITY_INFORMATION_DEFINED typedef DWORD SECURITY_INFORMATION, *PSECURITY_INFORMATION; +#endif typedef WORD SECURITY_DESCRIPTOR_CONTROL, *PSECURITY_DESCRIPTOR_CONTROL; /* The security descriptor structure */ -- 2.1.4

Feb. 28, 2015

Re: [PATCH 1/2] d3d9/tests: Extend the FPU setup test.

by Alexandre Julliard

Henri Verbeet <hverbeet(a)gmail.com> writes: > On 27 February 2015 at 12:43, Stefan Dösinger <stefandoesinger(a)gmail.com> wrote: >> Am 2015-02-27 um 12:34 schrieb Henri Verbeet: >>> Not that I would terribly mind if this got comitted, but I have to >>> wonder if this really has a lot of value as a regression test? >> As a regression test not really, except for cases where someone wants >> to add FPU CW set / restore in the way we're doing it in d3d7. I think >> the code has value as documentation though. >> >> I don't have a strong opinion on this either, but my preference is to >> have it in the tree rather than in the archives. >> > I guess it's mostly a matter of how Alexandre feels about this kind of > thing then. I don't mind having them in. At least they don't prevent me from using the computer while running, unlike some of the other d3d tests ;-) -- Alexandre Julliard julliard(a)winehq.org

Feb. 27, 2015

Re: [PATCH 1/2] d3d9/tests: Extend the FPU setup test.

by Henri Verbeet

On 27 February 2015 at 12:43, Stefan Dösinger <stefandoesinger(a)gmail.com> wrote: > Am 2015-02-27 um 12:34 schrieb Henri Verbeet: >> Not that I would terribly mind if this got comitted, but I have to >> wonder if this really has a lot of value as a regression test? > As a regression test not really, except for cases where someone wants > to add FPU CW set / restore in the way we're doing it in d3d7. I think > the code has value as documentation though. > > I don't have a strong opinion on this either, but my preference is to > have it in the tree rather than in the archives. > I guess it's mostly a matter of how Alexandre feels about this kind of thing then.

Feb. 27, 2015

Re: [PATCH 1/2] d3d9/tests: Extend the FPU setup test.

by Stefan Dösinger

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Am 2015-02-27 um 12:34 schrieb Henri Verbeet: > Not that I would terribly mind if this got comitted, but I have to > wonder if this really has a lot of value as a regression test? As a regression test not really, except for cases where someone wants to add FPU CW set / restore in the way we're doing it in d3d7. I think the code has value as documentation though. I don't have a strong opinion on this either, but my preference is to have it in the tree rather than in the archives. -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQIcBAEBAgAGBQJU8FhHAAoJEN0/YqbEcdMwtMMQAJWSekh9Yc/6mrj/3JsbhkDy UwlwExx11ldu3h5GwCph7ozjdVuzVTiU1+MqDd5opsSA3rCOEidVByNBCziCh6NH gsDt5VCYWD+Ri3ynGZsGrfI8xNGdffBbyke5irWnZnT49wE5BetprM6+k0q1ti/P bBEiv8J6jkCayBCQ1Ehjgvb6fL1GAcF9HMtC7nVKOuNKWD+OYawaUPdkKhEtWD3+ Qau8tEqrIhwswloBy6UDqxLkazzTQDoTi8/LXmDhnzuDk/hDqarlqpYMgFz0C5vM SF0Y1ADxxTEtDdODnmrwu21OkNq8kWqgv8eEdDWowDIGpsV07x4AI/9BpSaA3NcM FA2563EsBDIb+uF3xZ+PCtxBYfXgvF2gl3hiaH9tSgLVrBdGLIi2ofbAJ7tzCuTO 7VF3Otjhldmi84UEbg/Xc3vHAb1uJnzweNEy1p03nrgUHBxfKFaASilp1PNOmkWz kHgYdzKSm34I3+Jw1hbIeOx2pFPS3kZU4xuQYhg3RtKWUU5OWiOlqNiF987I9Whk +srjOwDaNnkG8E/dWJhqgDXxIbafU7q7UX/j9plRnp1vVysVNbYI5N6z7J5S4WKE roYGg0QYHaBE9ibLh9TOUpu30sV6EHtk25nAOBhVADgy7QamssfB9+tqMY9/jRDX 0nVuTYn9jiO6Z/cRLZfr =2n3Z -----END PGP SIGNATURE-----

Feb. 27, 2015

Help with testing new Gecko downloader

by Jacek Caban

Hi all, My today patch [1] together with changes in Wine code from yesterday changed the way we interact with SourceForge for downloading Gecko. It's a risky change that could use wider testing from different locations. If you'd like to help, you need to: - Remove wine_gecko-2.36*.msi if you already have it. If you used downloader to get it, it's most likely in ~/.cache/wine. If you downloaded it manually, you know where to find it. - Create new wineprefix The most important thing is to ensure that downloading and installing Gecko still works. The expected difference is shorter connection time and faster downloads (or better said, no slow downloads). I appreciate any feedback on this. Thanks, Jacek [1] http://source.winehq.org/patches/data/109462

Feb. 27, 2015

Re: [PATCH 1/2] d3d9/tests: Extend the FPU setup test.

by Henri Verbeet

Not that I would terribly mind if this got comitted, but I have to wonder if this really has a lot of value as a regression test?

Feb. 27, 2015

Re: [PATCH] [shell32] Implement SHRunControlPanel using ShellExecute

by Alex Henrie

+ if((int)ShellExecute(parent,NULL,commandLine,NULL,NULL,SW_NORMAL)>32) + return TRUE; + else + return FALSE; I think this would be more elegantly expressed as return (int)ShellExecute(parent,NULL,commandLine,NULL,NULL,SW_NORMAL)>32; But more importantly, are you sure that SHRunControlPanel will execute anything? What if the filename doesn't end in .cpl? You'll probably have to write some tests before this patch is accepted. -Alex

Feb. 27, 2015

Re: [PATCH] [shell32] Implement SHRunControlPanel using ShellExecute

by Alex Henrie

2015-02-26 7:36 GMT-07:00 Nikolay Sivov <bunglehead(a)gmail.com>: > On 26.02.2015 17:01, Robert Naumann wrote: >> - FIXME("(%s, %p): stub\n", debugstr_w(commandLine), parent); >> - return FALSE; >> + if((int)ShellExecute(parent,NULL,commandLine,NULL,NULL,SW_NORMAL)>32) >> + return TRUE; >> + else >> + return FALSE; > > Why 32? Also please don't remove tracing line. MSDN says that ShellExecute returns a value of 32 or less if it fails, and a value greater than 32 if it succeeds: https://msdn.microsoft.com/en-us/library/windows/desktop/bb762153%28v=vs.85… -Alex

Feb. 27, 2015