June 2006 - Wine-Devel - list.winehq.org

Re: wine autorun utility

by William Knop

On Jun 30, 2006, at 1:08 AM, Troy Rollo wrote: > On Friday 30 June 2006 14:52, William Knop wrote: > >> Um hold on a second. Clearly many developers have different ideas >> about what's reasonable. > > Actually I don't think that's true. As far as I can see all of the > *developers* participating in this thread agreed that autorun (1) > is a bad > idea and should not be implemented, and (2) is on the other side of > the > demarcation line between Wine and the desktop environment. > > I would be surprised if any significant number of developers > disagreed with > these two points. As for the first, it reflects the difference between > Windows' security model and the Unix/Linux security model (the > difference > being that the latter systems actually have a model deserving of the > label "security"). As for the second, an end user may well not > understand the > distinction between Wine's role and that of the desktop > environment, but a > developer should. Wine is for making Windows applications (and native > applications coded to the Windows API) run - I am not aware of any > developer > who thinks it should be a complete reimplementation of Windows. 1) While I agree maintaining a staunch security policy is important, that has nothing to do with autorun. Making the user browse to find an executable is not security. 2) The line you refer to I believe would put detecting media inserts on the desktop environment side, and the parsing and execution of windows autorun inf files on the wine side. Hell, the user would have to run `wine --media-autorun /mountpoint/autorun.inf`. That has to be as secure as the user running `cd /mountpoint/somedir; wine ./some.exe`. Will

June 30, 2006

Re: wine autorun utility

by Tim Schmidt

On 6/30/06, Chris <chris.kcat(a)gmail.com> wrote: > This sounds like you missed my point. I think you're missing our points Chris... > it's not inherently detramental (since the user would be instructed to manually do > what autorun does automatically Yeah. The Sony rootkit users would have gladly followed the instructions on their _audio CDs_ telling them to install software that prevented fair use and installed illegal software. We're not talking about preventing legitimate use of auto-running CDs. We're talking about a sound, simple, and easy way to prevent illegitimate exploitation. If you can develop a way for Wine to automatically determine whether or not an executable on removable media is something useful or a rootkit, then you may get a little more enthusiasm for this 'feature'. --tim

June 30, 2006

Re: wine autorun utility

by Troy Rollo

On Friday 30 June 2006 14:52, William Knop wrote: > Um hold on a second. Clearly many developers have different ideas > about what's reasonable. Actually I don't think that's true. As far as I can see all of the *developers* participating in this thread agreed that autorun (1) is a bad idea and should not be implemented, and (2) is on the other side of the demarcation line between Wine and the desktop environment. I would be surprised if any significant number of developers disagreed with these two points. As for the first, it reflects the difference between Windows' security model and the Unix/Linux security model (the difference being that the latter systems actually have a model deserving of the label "security"). As for the second, an end user may well not understand the distinction between Wine's role and that of the desktop environment, but a developer should. Wine is for making Windows applications (and native applications coded to the Windows API) run - I am not aware of any developer who thinks it should be a complete reimplementation of Windows. -- Troy Rollo - wine(a)troy.rollo.name

June 30, 2006

Re: wine autorun utility

by Vincent Povirk

On 6/29/06, Chris <chris.kcat(a)gmail.com> wrote: > If you notice, Sony got into a lot of trouble over that. And the problem > wasn't autorun. The problem was that the disc installed the rootkit > anyway /even if the user said no/. The same exact thing would've happened if > the user had to browse the CD and double-click setup.exe, or whatever the > file was called. Should Wine disable running .exe files because they may > install rootkits on users' machines? Of course not, because that would be > couter-productive to what Wine is trying to achieve. It's the same thing with > autorun. It may or may not cause problems, but it's the user's responsibility > to take proper care of their machine. It's just as true in Windows as it is > in Linux, or any other OS. [skip to the last paragraph to get to the point; I thought this other stuff was necessary and now I don't, but I also don't feel like deleting it having already written it] I'm not sure we should install programs in that way on Linux. I think one of the most noticeable differences between Linux and Windows for most people is how you install software. In Windows, you get an exe file from somewhere (hopefully from someone you can trust), and you run it. You then follow the instructions, possibly reboot the computer, and then when you're done you expect the program to be installed and working. If it doesn't work, you complain to the authors of the software. In Linux, most of the software you need will be provided by your distribution. If you're on, say, gentoo, you type "emerge software", and when it's done you expect the program will be installed and working (in this case, you don't have to do anything; you just wait). If it doesn't work, you don't complain to the authors, you complain to gentoo. If it turns out to be an upstream bug, it gets passed along to the authors. If it turns out to be a bug in gentoo, then hopefully they fix it. Of course, Linux users can also do something similar to what the Windows users do and handle the whole process themselves. Then they'll report problems directly to the authors. Now, here's the part where I reward you for reading this far and tell you what this has to do with Wine. Wine users get software from the authors, try to install it themselves, and expect it to work. Then when it doesn't work they complain to Wine. That's backwards. Unless they checked the appdb first, probably no one from Wine claimed it would work to begin with. Yes, most people aren't like that, and yes, people should report bugs. But they shouldn't act like someone said it would probably work if no one has. I think this is a good reason to have something like Wine Doors. For applications that are known to work, possibly with some override or hack, someone could write an installer and vouch for the workingness of at least some apps. People who don't know how to do things like overrides can use the installer. If something doesn't work (and you should expect it to work if there's an installer that's supposed to work on that wine version) then those users have somewhere to complain to. Conversely, applications without installers would be hit-and-miss. You'll have to do a bit more work for those, and you may be paving the way for others by submitting testing data, writing howtos, bug reports, an installer, maybe even fixing something in wine. So it's clear that they might not work. Now, when you make someone's environment say "This CD has a program on it, shall I run it?", you're making the claim that this will probably work. That's just not true. It probably will not work as well as it's supposed to. I'd much rather have people go to a package manager, the applications database, the documentation, ANYWHERE where they will not be promised something that isn't true, even if it makes Wine seem harder to use. -- Vincent Povirk

June 30, 2006

Re: wine autorun utility

by William Knop

> Gee sounds like a "great" idea. We all waiting too see some patches... > > It sure would be cool to have: > - Multiuser Wine > - Wine stable enough to run as service (err hmm whatever the hell > that means... > ah you mean daemon ?) > - Run something more complicated then 'printf("hello world\n");' > without X > - Talk to WMs to show icons and ask questions. > > Chris if you think that autostart is such a great idea - you are > very welcome to > start sending patches in. And if they are reasonable enough they > might get in. > But if you want to rant about that Linux doesn't have some > absolutely required > "feature" that windows has - this not the right place. > > Vitaliy Um hold on a second. Clearly many developers have different ideas about what's reasonable. It makes sense to obtain a semblance of unanimity and mutual understanding before taking action. You shouldn't go around quashing discussions like that. Will

June 30, 2006

Re: wine autorun utility

by Chris

On Thursday 29 June 2006 21:13, Vitaliy Margolen wrote: > Chris if you think that autostart is such a great idea - you are very > welcome to start sending patches in. And if they are reasonable enough they > might get in. But if you want to rant about that Linux doesn't have some > absolutely required "feature" that windows has - this not the right place. This sounds like you missed my point. My argument isn't that autorun is cool and Linux should have it (when in fact KDE *does* have it, to some degree), my argument is that Windows has it and it's not inherently detramental (since the user would be instructed to manually do what autorun does automatically anyway, and hence have the same effects), so Wine, if it is trying to match Windows feature-for-feature and bug-for-bug, should have it too. IMO, of course. If I knew how dbus and hal worked, I would probbly try to make the necesarry patches. But as I'm not familiar with Win32 or Wine's internals, I'm not really qualified to work on it, at this time. I'm just a user voicing my opinion.

June 30, 2006

Re: wine autorun utility

by Tim Schmidt

On 6/29/06, Chris <chris.kcat(a)gmail.com> wrote: > If you notice, Sony got into a lot of trouble over that. And the problem > wasn't autorun. The problem was that the disc installed the rootkit > anyway /even if the user said no/. The same exact thing would've happened if > the user had to browse the CD and double-click setup.exe, or whatever the > file was called. Should Wine disable running .exe files because they may > install rootkits on users' machines? Of course not, because that would be > couter-productive to what Wine is trying to achieve. It's the same thing with > autorun. It may or may not cause problems, but it's the user's responsibility > to take proper care of their machine. It's just as true in Windows as it is > in Linux, or any other OS. Of course. You're right. Everyone's computers _should_ run arbitrary code from any un-authorized source automatically without the user's knowledge or permission. I was wrong. The fact that Windows ran _anything_ upon inserting a CD meant to contain audio only is crap. I understand that Sony exploited a 'feature' of Windows. It's all Sony's fault. Blame Sony. Problem is, that philosophy pushes the trust all the way out to the people who want to install rootkits on your computer. Bad idea. Better to trust Wine not to do anything to endanger your computer without your explicit attention. --tim

June 30, 2006

Re: wine autorun utility

by Vitaliy Margolen

Thursday, June 29, 2006, 8:33:24 PM, Chris wrote: > On Thursday 29 June 2006 07:49, you wrote: >> what is the gain of autorun, while i dont have wine running? would not >> the common user expect, once she knows, that wine support autorun.inf, >> that she can insert a cd and away it goes? while the explorer.exe is >> not running permanently - e.g. via xinitrc - then i still would have to >> deal with starting my setup.exe by hand. i have my doubts, that wine is >> the correct place to implement this - but something like hotplugd or >> whatevernowadaysdoesthetrickwithcdroms should handle this? > One could think of Wine as a service, and start it with X if they wanted, so > they could then run Windows programs whenever they wanted with potentially > faster startup times, and with Windows "features" like autorun, and with a > persistant Windows state (eg. instead of being shut down when all Windows > programs close, it would stay around until Wine itself is explicitly > stopped). Or, like me currently, they could run a Windows desktop/taskbar > replacement in a Wine virtual desktop with managed windows off on a seperate > X desktop and pretend to have Windows running along side X. > I'd think if Wine were trying to "emulate" Windows as much as possible, it'd > be backwards to not have autorun capabilities. As I said before, autorun.inf > needs to be parsed anyway since it can define an icon to use for the drive, > so if you're going to automatically parse it to load an icon, why not also > ask (one time only) if the user wants to turn autorun on or off to run the > defined program? Gee sounds like a "great" idea. We all waiting too see some patches... It sure would be cool to have: - Multiuser Wine - Wine stable enough to run as service (err hmm whatever the hell that means... ah you mean daemon ?) - Run something more complicated then 'printf("hello world\n");' without X - Talk to WMs to show icons and ask questions. Chris if you think that autostart is such a great idea - you are very welcome to start sending patches in. And if they are reasonable enough they might get in. But if you want to rant about that Linux doesn't have some absolutely required "feature" that windows has - this not the right place. Vitaliy

June 30, 2006

Re: wine autorun utility

by Chris

On Thursday 29 June 2006 20:31, Tim Schmidt wrote: > The Sony rootkit fiasco alone should be enough to end this conversation. If you notice, Sony got into a lot of trouble over that. And the problem wasn't autorun. The problem was that the disc installed the rootkit anyway /even if the user said no/. The same exact thing would've happened if the user had to browse the CD and double-click setup.exe, or whatever the file was called. Should Wine disable running .exe files because they may install rootkits on users' machines? Of course not, because that would be couter-productive to what Wine is trying to achieve. It's the same thing with autorun. It may or may not cause problems, but it's the user's responsibility to take proper care of their machine. It's just as true in Windows as it is in Linux, or any other OS.

June 30, 2006

Re: wine autorun utility

by Tim Schmidt

The Sony rootkit fiasco alone should be enough to end this conversation. Period. Say what you want about the theoretical integrity of the media, and the user's security habits. The fact is that hundreds (possibly thousands or millions) of _real_ people were infected by rootkits because of autorun and an unscrupulous corporation. As you've so aptly demonstrated, average users are all too willing to trust people who don't deserve to be trusted. If Wine can protect some of those people sometimes, great. If Wine can protect people without any developers having to do any work, that's simply amazing. Well done Wine devs. --tim

June 30, 2006