World Wine News

From Sven Paschukat I was informed about the discussion of removing the WineTools link from the winehq website. I read all remarks made to that hotly discussed topic and I am willing to tell you all a bit about the purpose and history of WineTools as well as answer the questions made by Tom Wickline. So this will be a little longish mail and I beg your pardon for that and hope you have enough patience to read it all.

To start from the beginning, I took over the remainings of a project winetools initially started by Frank Hendriksen <frank at frankscorner.org> which was somehow orphaned. The very reason I did that was that in autumn 2004 I had to write an article in the well known German computer magazine c't. After fiddling with wine for a long time only to install a stable IE6 or Office, I found that it was so absolute complicated to describe all the steps needed only to install the IE6 (which is much better now) and that also following these steps was very erroneous, that I started to rewrite the winetools script for the readers of this article.

I never meant it to become so famous and wide spread, but the download statistics after only two month showed me that there was a massive demand for such a tool. So this is how it started and it shows very clearly that it was always an installer to make certain important applications easily usable under Linux. It was never meant for testing or developing. It was always meant for the end user being able to use his software.

At the time of the first releases it was almost impossible to install or work with many applications without using native DCOM and/or many of the builtin DLLs that come with the IE6. To find a configuration for as many apps as possible was one of the goals. Not to force users (with very low skills) to have many different wine installations *and* many wine user directories was a second goal. So this is the reason why, until now, everything in WineTools is based on this native M$ software (DCOM, IE6). And believe me, no one would be more happy to wipe this DCOM, IE6 and MSI stuff completely from his harddisk than me!

So as so many programs rely on parts of M$ software (MFC, VB and MDAC are other important examples) and because to install and use this software leads to so many restrictions in the wine usage, I had to pin wine to the emulation of Win98 and also had to massively tamper around with DLLOverrides. The goal was to built up an environment, that can install and use as many Windows programs as possible. This did in no way regard to the fact that it is not very helpful for the wine developers. It was just a practical decision.

From time to time, Sven Paschukat and me are testing wine versions to figure out, whether we can skip some of the M$ stuff. We have not been very successful with that until now, so this is why there are so many remainings of the first start of the project in the concept. But again: This tool is for the average windows user. And believe me, I tested this with my brothers, who are just normal users. And they never ever got a piece of windows software running with wine without my heavy intervention. With WineTools they just need to click on some buttons and got their software downloaded, installed and configured.

With an amount of many thousand downloads a month, not calculating the distribution inclusions of WineTools, I get almost the same number of direct mails about problems with the software as are coming into the list. And I get massive positive resonance from users who were for a long time trying to get their stuff running with plain wine and are now happy that it just works with WineTools.

I must admit, that there raise problems out of the fact, that the same users are trying to install not WineTools-tested software with the setup of WineTools and coming then to the list. And your are right that this is an undebuggable situation. The only way around that is to smoothly migrate WineTools to more and more builtin features as long as this does not make the programs uninstallable or unusable.

Sven Paschukat suggested to implement a mode for WineTools to install without native components and without that tampering in the registry. This allows automatic installations *and* insight for developers if something goes wrong. We can also add a debugging mode to enable the logging of wine's debug channels, to make a mailable report of bugs. Also the other suggested corrections will be looked over and changed if applicable.

If you look through the mailing list you will find, that Sven and me are already giving support on the wine-users mailing list. That we will not and cannot maintain a channel should be clear as we are not earning our money by looking at channels. Also the demand by Vitaly to provide rapid fixes to requests on such a channel or the list is something that can be only meant as a joke!!! If Vitaly is a man of independent means it is nice for him but we are a free time project and can only work for WineTools, as the name suggests, in our *free* time. And we have also a family, friends, mistresses and other hobbies...

Well, we also do not like winehq to remove the project's link. As for the support I already wrote a line above about that. Note that there is a massive amount of users using Wine via WineTools. Almost all users coming from Windows are using this software if they like to use their programs. This might lead to more and more end user questions on the list. As a possible solution for that, if you think this gets the upper hand, I would suggest to have a new mailing list wine-winetools and to redirect all users to that one. We will propagate then this list in the readmes and on the WineTools sites. And for sure we will sit on this list like spiders lurking for our prey...

At least I want to thank all wine developers for their absolute fantastic work! In my opinion this is one of the most interesting and most ambitious projects in the community of open source software. And it really works astonishingly good.

The problem with this as I see, is that we as Wine developers creating an environment of one kind. And these "tools" almost totally override that environment.

How often do you guys check if a piece of software can be installed on Wine as-is? Without any additional configuration or native components? The point and click user have to know what they are clicking on, or not be presented with the bad choices. With the way it is _ALL_ new times visiting http://www.winehq.org/download have an impressions that winetools are the absolute requirement and download and install them.

I can't tell how many times I had to tell the user to vipe their ~/.wine dir and start from scratch _not_ using winetools because what they want to run does not require anything that winetools install.

Also I don't see any post from you or Sven about these problems in the bugzilla. How do we know that there are any problems?

As I mentioned before, every single person visiting download page thinks that winetools are something mandatory. I would really like too see what programs run and don't run with and without winetools. Also please redirect all those users to appDB. We need information there available for everyone to see, not in your e-mail box (no offense).

WineTools has setup the environment in a such a way that it is unified! If you're saying that all you did is tinkered with lots of stuff to get few applications to work, this is not good. Wine is not cedega designed to run games and games only. That just proves my point that environment you have setup with winetools is less flexible and more restrictive than Wine itself.

I think it's really silly for Wine trying to be the pure replacement of Windows and yet advertise such a utility that clearly undermines development in a number of critical areas.

Hm. Felt I needed to offer a counterpoint to Vitaliy's rather enthusiastic response.

I support winetools, though I don't use it myself. Not only as a matter of principle (this is open source we're talking about, and one of the beauties is we don't get to constrain how it's used.) Also because it's in line with the aims of the Wine project: run Windows apps. Wine is nice in that it doesn't require you to have any MS license to run Windows apps, but it can use MS software if you have a licensed copy. So winetools offers folks with MS licenses a relatively easy way to get apps to run.

Still, while that's nice for the users, it's not so nice for us Wine developers, because it's hard to get widespread testing of things we're improving if they're not being widely used.

So, as Vitaliy suggested, it would be nice for us if you guys, who know what problems you encounter running with builtin DLLs, could file bug reports in bugzilla for us. That way at least we have a chance to say, hey, that problem's now fixed, consider using builtin OLE (or whatever) in the next release of winetools.

Now for specifics. For instance, you use native advpack in all cases, but that's seen some more work recently. What bugs do you encounter if you use builtin instead? Same with ole32. Ideally, each of the apps that has a native override should have a bug report associated with it, what goes wrong if you use the builtin version instead.

Also, the default override, "*"="native,builtin" seems wrong. If we don't have a DLL at all, we use a native version of course. I don't think it should be the policy to assume the Wine version isn't up to snuff by default. This could even cause problems, e.g. I don't believe any native version of netapi32.dll will work on Wine, yet you don't specify to use the builtin version.

Also, you direct users to http://winehq.org/forums. That's fine since this is community support anyway, and it is often the case that winetools knowledgeable people are around (like you guys.) But it would be nicer if you could be clearer about it. Like, that we developers had nothing to do with it, and may be unfriendly to it if we're in a bad mood, or it's a Tuesday :)

It'd also be nice (for us, even though no one actually reads anything these days) if you could be clearer about what it does. That it installs MS components, that it uses these instead of the builtin versions sometimes, that you can change how it does this using winecfg, etc. That'd make it easier for us to help folks that have only used winetools that want support, because often the first question we ask is, what version of wine, and what DLLs are you using?

This is the first of 5 patches that implement the Windows tape API on top of the Linux SCSI tape driver (st).

I admit this is probably not the last missing piece that kept people from flocking to Linux and Wine, but I do know that I really enjoyed hacking away on Wine to make my dusty old 2GB DDS 1 SCSI tape unit spin again. Hacking Wine is fun!

There's one missing piece after this patch, which is the association of an NT device name with an appropriate Unix device file. Until Wine learns how to autodetect these you need to do it by hand.

For example, to associate NT tape device \\.\TAPE0 with your Linux SCSI tape device /dev/st0 you need to create this symbolic link:

$ ln -s /dev/st0 ~/.wine/dosdevices/tape0
$ ls -l ~/.wine/dosdevices/
lrwxrwxrwx 1 hans hans 10 2006-01-07 22:58 c: -> ../drive_c
lrwxrwxrwx 1 hans hans 8 2006-01-25 22:22 tape0 -> /dev/st0
lrwxrwxrwx 1 hans hans 1 2006-01-07 22:58 z: -> /

Yes. Great, thank you very much :-) !

BTW: I had to set "Hardware Acceleration" to "Emulation" (instead of "Full", but did not try other levels) to make mplayer2.exe play all my test wave files _right_.

I used the c:/Windows/Media/*.wav files. And the strange thing was, that mplayer2.exe (with "Full") played some of them right and others wrong (= sounds like sampling freq. (transformation) issue).

I have posted a comment at

http://bugs.winehq.org/show_bug.cgi?id=4351

The problem seems to be with the name of the executable: "user.exe" (sounds crazy, but it was I could realize after some tests).

That's quite possible. user.exe is a 16 bit part of user32.dll, and Wine treats it in a special way.

You might be able to add a dll override entry in winecfg like this:

*user.exe = native, builtin

In fact there are a little more with same problem: start.exe and sound.dll (as reported in bugs 3950, 4351 and 4311)

However "winecfg" does not allow me to input free form strings (at least I tried and could not find a way). It does only allow me to navigate through the file system and choose the application. And the "Libraries" tab, in winecfg, it is only for DLLs - I guess! (Am I wrong?) The only solution I tried is using "vi" to edit ~/.wine/user.reg but I think I made a mistake because wine still segfaults.

Here is the KEY I have put:

[Software\\Wine\\AppDefaults\\*user.exe\\DllOverrides] 1138094570
"*"="native,builtin"

I have also tried a variation of it .... but it still segfaults when calling "wine user.exe":

[Software\\Wine\\AppDefaults] 1138094570
"*user.exe"="native,builtin"

No, it can be used for .exes too, you only have to specify the full name (e.g. "user.exe" for user.exe compared to "oleaut" for oleaut.dll)

I often use it for dplaysvr.exe, so it works

Trying to track down the problem with our DInput implementation I found some interesting stuff - our global hooks don't work correctly because hook callbacks are never called if the event is generated in a different thread. I don't think this is a revelation to a number of people who knew that before. But how do we fix it?

Here is what I dug up so far:

1. Installing global hook we are setting it as WINEVENT_INCONTEXT while msdn and tests indicate that it qualifies as WINEVENT_OUTOFCONTEXT instead. Unless of course we don't want to call post_win_event in step 5 and go to step 6 instead.
2. All HW mouse messages start here:
   http://source.winehq.org/source/dlls/x11drv/mouse.c#L191
   From looking at it I assume that all the HW mouse hooks should be handled by HOOK_CallHooks function. This is because there is no code related to hooks in send_message server call.
3. First thing that HOOK_CallHooks does is calling HOOK_IsHooked with checks only current thread for installed hooks. So this is the first place that is wrong. Let's comment that check out for now. I think we need to move that check farther down, after the call to server.
4. We call start_hook_chain http://source.winehq.org/source/server/hook.c#L468 to start hook callback chain. There we call get_first_valid_hook on local and then global if no local hooks exist. But there again we check for http://source.winehq.org/source/server/hook.c#L211 the hook callback thread being local (which is again not the case for global hooks). This is problem #2.
5. Now we have post_win_event of but right before it we are hitting an assert so this is not the right way?
6. Say we returned with success and back to HOOK_CallHooks that now tries to call MSG_SendInternalMessageTimeout. Which for some reason fails to deliver message to our global hook.

So my question is to anyone who knows. What is the proper way to deliver these messages to global hooks that are in a different thread/process?

Just an observation I did in a Windows program running under WinXP. Not sure if this is related somehow but it seems quite like the opposite of what you describe here although with windows message hooks.

I had a program that was hooking window messages for a particular window. When using SendMessage from the same thread as in which the Windows message loop was running (doesn't make much sense of course but it was just for testing purposes) I could not receive any events above WM_USER, lower events did seem to get received properly though. When placing SendMessage in a different thread the events did come through.

Not exactly sure yet what this is all about but message handling and especially hooking in Windows for sure can be a tricky business.

That's because the new win event code changed the meaning of the thread field, and the low-level hooks haven't been updated for it. Try something like this .

It does work thank you. The only question I have : does that cover WINEVENT_OUTOFCONTEXT hooks too?

I'm working on a test to test hooks. Hopefully there will be a way to test it.

And it looks like we will have to rethink our DInput - the game still doesn't work. It seems that the thread which acquired the mouse doesn't have a message loop.