October 2003 - Wine-Devel - list.winehq.org

SetShellWindow() and friends
by Martin Fuchs Oct. 13, 2003

Oct. 13, 2003

Hello Alexandre, > Yes, if it works under Windows then obviously we need a fix in Wine. > I think you should investigate exactly how Windows does it, there are > many places where we could add such checks so we need to find out > exactly where Windows has them. OK - wrote a little test program to see, how SetShellWindow() exactly works on Windows. It creates a few windows and tests, what happens when trying to register them as shell window. To run it you have to terminate Windows Explorer at first. Then the output is this the following: first call to SetShellWindow(hwnd): SUCCESS second call to SetShellWindow(hwnd): FAILURE reset shell window by SetShellWindow(0): FAILURE third call to SetShellWindow(hwnd): FAILURE SetWindowExStyle(hwnd, WS_EX_TOPMOST): FAILURE DestroyWindow(hwnd): SUCCESS SetShellWindow(hwnd2) with WS_EX_TOPMOST: FAILURE SetShellWindow(hwnd3): SUCCESS SetWindowPos(hwnd3, HWND_TOPMOST): SUCCESS SetWindowPos(hwnd3, hwnd4: SUCCESS SetShellWindow(hwnd4): FAILURE SetWindowPos(hwnd3, hwnd5): SUCCESS So we have to taking following rules into account for the SetShellWindow/Ex() implementation: - SetShellWindow() only succeedes if there is not yet a registered shell window. - A successfull call moves the window into the background. It becomes the window with the lowest Z-order. This can't be changed by a call to SetWindowPos() or similar functions. The shell window remains at the bottom for ever. - Shell windows can't be reset by calling SetShellWindow(0). To change the shell window, the current window has to be destroyed. - Shell windows must not have the extended style bit WS_EX_TOPMOST set. It it is set, SetShellWindow() failes. If you try to set this bit after registering a window as shell window, the bit is ignored and automatically set to zero. - If you try to move the window into the foreground by calling SetWindowPos() with HWND_TOPMOST this call succeedes, but it does not change anything. The window remains in the background. -- Martin Fuchs martin-fuchs(a)gmx.net

2 2

Re: winesetuptk
by Ivan Leo Murray-Smith Oct. 13, 2003

Oct. 13, 2003

I don't think winesetup needs big changes; I'm still not sure about what is broken. The attached file updates the registry shipped with winesetuptk to the current winedefault.reg, and also updates the dlls that are created when making a fake windows installation. As I've said, I'm not sure of what is broken, but if what has to go isn't determined by winesetuptk while running (Like the AppDefaults) then it's just a question of removing whatever has to go from the cfg_wine_without_windows.ini and/or cfg_wine_with_windows.ini. If anything else has to be changed, then Vincent should tell Alex what exactly has to change, so Alex can get on with it. This is presumably the quickest way of getting a working winesetuptk again, as he know Tcl/Tk very well. It's not worth wasting too much time on it, as were going to trash it soon, it doesn't matter if it can't tweak all aspects of the config file, it just has to make a working configuration for the user, that will often have to tweak the file by hand anyway until winecfg is ready.

3 2

Re: winesetuptk
by Ivan Leo Murray-Smith Oct. 13, 2003

Oct. 13, 2003

> But the problem is that it is currently broken. So someone either fixes > it now, or we hide it until it's fixed or not needed anymore. I can't > see how is it any friendlier to users to download something that does > not work. In what way doesn't it work?

4 3

CVS issues
by Kevin DeKorte Oct. 13, 2003

Oct. 13, 2003

I updated CVS on 10/8 and Lotus notes worked fine. I updated on 10/12 and now Notes gives me a Date not initialized error and all the date and times in the program are all mangled. I am rolling back to 10/8 Kevin

1 0

Re: winesetuptk
by Ivan Leo Murray-Smith Oct. 13, 2003

Oct. 13, 2003

> "The problem is that the tool hasn't really followed Wine development > (for example, removal of $VAR-style variables from config file), and > some users experience problems because of that. Such as?

1 0

Re: winesetuptk
by Ivan Leo Murray-Smith Oct. 13, 2003

Oct. 13, 2003

Most normal people (Users) don't want to edit config files or use a shell. That's why most people don't use linux. Winecfg isn't complete, winesetuptk will go soon, but in the mean time it's the best way to configure wine. I get emails all the time from people thanking me for putting winesetuptk on sourceforge, it's had a lot of downloads, and I don't think we can remove it yet. As soon as winecfg is ready, I'll hide winesetuptk (That won't work anyway because wine won't read the config file). Until then, winesetuptk is the only easy way of configuring wine, and I have no intention of telling users "Go and edit the config file and add this and remove that and make a system directory inside the windows directory that's in your fake c:\ drive that the config file has to point to..." when I can just say "Install winesetuptk, start it, and follow the on-screen instructions". I don't see how such a good level of usability can hurt.

2 1

Re: question about copyright and code, is this legal?
by Ivan Leo Murray-Smith Oct. 13, 2003

Oct. 13, 2003

>also goes over the UK As the UK law must be compatible with the EUCD, I find it hard to believe.

1 0

RE: question about copyright and code, is this legal?
by Patrik Stridvall Oct. 13, 2003

Oct. 13, 2003

> > Protecting trivially collectable collections of fact would serve > > no purpose that I can see... > > Here, if the collection as a whole requires trivial effort to > collect, it's > not protected. It's the aggregate effort that counts. The > problem with the MS > tables is that from what has been said the aggregate has > involved a lot of > effort over many years. Perhaps. However: 1. The data in question is of a more scientific in nature than the facts about the world that law is presumably ment to protect so it might not apply at all. 2. Since Microsoft have intentionally published it in the US where no such protection exists, it could be argued that it have exhausted its right to protection because of this. After all everybody could import it itself from the US if they wanted. In the case of Wine and other open source software this is especially true because of the Internet.

1 0

Sorry for the spam
by Duane Clark Oct. 13, 2003

Oct. 13, 2003

So sorry about the spam that got through. It was a boneheaded moderator error. Reminder to self. Drink coffee before moderation. Sorry again. -Duane

2 1

Winealsa: now fixed, interactive tests still ko.
by Sylvain Petreolle Oct. 12, 2003

Oct. 12, 2003

Applying this patch fixes alsa : [repaired] dlls/winmm/winealsa/audio.c thanks to Daniel (Tom?) and Jerry. Now "Common" tests are ok. But some interactive tests still fail with winealsa. I attached the log. (+wave,+dsound,+dosund3d can be provided if necessary) ===== Sylvain Petreolle (spetreolle_at_users_dot_sourceforge_dot_net) ICQ #170597259 Say NO to software patents Dites NON aux brevets logiciels "What if tomorrow the War could be over ?" Morpheus, in "Reloaded". ___________________________________________________________ Do You Yahoo!? -- Une adresse @yahoo.fr gratuite et en français ! Yahoo! Mail : http://fr.mail.yahoo.com

2 1