- Wine-Devel - list.winehq.org

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

Oct. 13, 2003

> If an app installs msvcrt, use it before Wine's builtin one? That could theoreticly create legal problems, but it's OK for me. > I used the default values of documentation/samples/config. If those need > changes, it's best to do so there. I don't really care, but I personally prefer using builtin only (But that's a personal view) > Then those people use rpms or debs. And in that case, it's usually > already configured out of the box if the packager knows what he's doing. I (Following the example of codeweavers, at least until preview 4 that's the last I tried) use winesetuptk for packages. If you give a default config, you impose something that users may not want and may not know how to change. If you use winesetuptk, you give the user a choice of using a real or fake windows setup, and the user can also choose where to have it. The only difference is that codeweavers has a fancy pop up that asks you to run winesetuptk if it doesn't find a config file, and I'm not codeweavers, so I use the "By Alexandre" vanilla source (Otherwise it wouldn't be a official build anymore). Anyway, the config will go soon, so there isn't much point in discussing this.

1 0

RE: winnt.h file
by Steven Edwards Oct. 13, 2003

Oct. 13, 2003

Hello Geoff, Comments are inline --- Geoff Peters <Geoff(a)logic.bm> wrote: > Steven - Thanks for the reply. I have 2 C files I am trying to port > to Solaris, one is done (compiles OK), the other one I am working on > has a dependency on wininet.h, along with it comes a huge raft of > other dependencies. I obtained most of these libraries from > http://sources.redhat.com/cgi-bin/cvsweb.cgi/src/winsup/w32api/include/?cvs… The w32api package is only for developing Windows application using Mingw on Windows or cross-compiling Windows apps from a GNUtool chain. Winelib is a better solution for your problem. > There is one block of code that is dependent on this wininet.h > library - I am not a C developer (java!), so I'm kinda stumped as to > whether I should attempt to re write the offending piece of code > using a similar Solaris C API (I am assuming the wininet.h contains > most of the HTTP functions for windows), or simply try to match up > the correct libraries - again coming from Java I am having a hard > time understanding different libraries for different platforms to > perform the same functions. The winnt.h header file I am using is the > latest release you had posted on the above link - I just recently > tried to incorporate pieces of > http://cvs.winehq.com/cvsweb/~checkout~/wine/include/winnt.h?rev=1.172&cont… > which have a context defined for SPARC - which seems to be whats > missing from the redhat winnt.h Yes the structures and defines in winnt.h are Windows and WINE only. It will be nearly impossible to use them in a 100% native application. > In case you can offer any suggestions, I have attached a copy of the > C code I am attempting to port, the last function listed (int > call_cc_gateway) is the one that contains the references to wininet.h > > Haven't really looked at WINE, do you think it would help me out?? I think it could help alot. The first thing you may want to try and do is run your win32 application under WINE on a x86 linux box. If it works there then it should recompile as a Winelib application. The issue then becomes does Wine and Winelib work on Sparc. ATM I think the port is incompleate although we did have some people at Sun who were working on it and Wine does run under Solaris for x86. You can try and port your Win32 application if you can get a WINE to build for Solaris on Sparc. Another option would be to look at what functions Wines wininet.dll imports and see if you can develop a native implementation of wininet that sits on top of the Unix TCP/IP implementation rather than on Wine Winsock. We will try to help you as much as possible if you want to go the Winelib route on the Sparc as we really want to see the Solaris on Sparc port working. Thanks Steven > Geoff Peters, SCJP Phone : (441) 296-9640 > Applications Developer Fax : (441) 292-1509 > Logic Communications E-mail : geoff(a)logic.bm > 12 Par-La-Ville Road WWW : http://www.logic.bm > Hamilton, Bermuda HM JX > > > -----Original Message----- > From: Steven Edwards [mailto:steven_ed4153(a)yahoo.com] > Sent: Monday, October 13, 2003 11:37 AM > To: Geoff Peters; Steven_Ed4153(a)yahoo.com > Subject: Re: winnt.h file > > > Hello, > Are you trying to use Winelib to port your project to solaris? I can > try and help if you are taking this method. If you are using parts of > the w32api package to try and create a native application without > WINE > it may be a little harder. Can you send me your winnt.h header? > > Thanks > Steven > > > --- Geoff Peters <Geoff(a)logic.bm> wrote: > > Hi - Got your email from the cvs.sourceforge.net version info for > the > > winnt.h file. I have a question - wondering if you could help me or > > point me to someone that could - I am porting a win32 app to > Solaris, > > seem to be doing OK with the API's in the repository, however I am > > getting the following error on compilation using gcc on Solaris 8 > > (SPARC): > > > > In file included from /opt/portal/6.2/include/windef.h:246, > > from /opt/portal/6.2/include/windows.h:49, > > from call_tellan_cc_gateway.c:5: > > /opt/portal/6.2/include/winnt.h:1950: #error "undefined processor > > type" > > > > Obviously doesn't recognize the SPARC processor - any idea how I > > could work around this?? > > > > Thanks for your time, Geoff > > > __________________________________ > Do you Yahoo!? > The New Yahoo! Shopping - with improved product search > http://shopping.yahoo.com > > ATTACHMENT part 2 application/octet-stream name=call_tellan_cc_gateway.c __________________________________ Do you Yahoo!? The New Yahoo! Shopping - with improved product search http://shopping.yahoo.com

1 0

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

Oct. 13, 2003

How can msvcrt possibly be native in a without windows setup? Anyway, if you want winver as win98, we might as well have winelook at win98 too. >Which can lead to the question, >what's the purpose of it if wineinstall can do it? wineinstall is a command line tool for installing wine from source code. Most people would say it's a "For geeks only" thing. You can start winesetuptk without knowing what a shell is, and you can set it up with a few clicks of the mouse, it's usually used by people using binary packages, you know, those people that only use computers as tools and don't know/care about source code and tweaking configurations, and that could never use wineinstall.

3 2

RH7.3 packages
by Vincent Béron Oct. 13, 2003

Oct. 13, 2003

Some time ago somebody asked here about RedHat 7.3 Wine packages. I said I'd do it, pending some hardware upgrade on one of my box. That time has now come, and they are available on the sf.net download page. Only i386 and src for now, no athlon nor i686. Vincent

1 0

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