Xorg crashes after using Microsoft Word 2003 in Wine.

Hi Anthony,

could you send us your ~/.xsession-errors and your /var/log/Xorg.*.log right after the crash when you reproduced it?
Please attach it to the bugreport.

Thx,

\sh

Anything that crashes Xorg is an Xorg (or driver)
bug. It cannot be a wine bug. Please reassign.

Here is ~/.xsession-errors.

Here is /var/log/Xorg.*.log.

On Jan 18, 2008 10:34 AM, dankegel <email address hidden> wrote:
> Anything that crashes Xorg is an Xorg (or driver)
> bug. It cannot be a wine bug. Please reassign.

@Dankegel
You contribute much to our community, and I for one appreciate your
participation, but I have to disagree with you on this one.

While I do not know the particulars of this bug, I know that the
categorical statement above is not true. Whether it's a bug in Wine or
Xorg depends on the facts. Killing the session may be the right thing
for Xorg to do if another program misbehaves badly enough.

Documents created from Microsoft Office components in particular are
susceptible to extremely bad code execution. The biggest reason is
that the VBasic scripting language that is embedded in Office macros
makes it possible to embed dangerous code in a document, sometimes
without realizing it. This turns out to be a prime suspect in such
cases because macro execution is the only technical reason to run
MSOffice on a Linux box in the first place. OpenOffice can read and
write the MS formats as well or better than MSOffice itself. However,
many companies have developed documents with complex and sometimes
sophisticated macros and scripts, which OpenOffice cannot yet run with
any safety. Consequently, such companies are often compelled to use
MSOffice, even when they have otherwise converted to open source.

Again, I don't know the specifics of this case, but we can't just
dismiss the possibility that the issue should be fixed in Wine. A
program such as Wine that undertakes to run code written for another
operating system should generally be responsible for containing any
malfeasance that the code causes.

Happy Trails

--
Loye Young
Isaac & Young Computer Company
Laredo, Texas
http://www.iycc.biz

Anthony,

thx for your information...but I can't see any errors in your xorg logfile, which should be the normal way when crashing ;).
The problem, I can't try to reproduce the crash, just because I don't own something like MS Office.

Would you like to file a bug at upstream (http://www.winehq.org/) and provide them as much information as possible?
Furthermore attach the bug url to this report, so we can track it.

In this very moment I think wine is misbehaving and xorg takes care of it with a crash, which is not nice but can happen.

Thx for your help shaping the future of opensource :)

\sh

reassigning this bug to wine, because I have the feeling it's not Xorg...

You seem to underestimate the requirements for X.
X is supposed to be bulletproof and
immune to crashes caused by non-root apps.
Any chink in its armor is a bug in X or its drivers.
One should be able to run ANY userspace program
(that does not DOS the system in general)
and have X stay up. If this seems unreasonably
demanding of X, go ask the X maintainers, they'll
tell you the same thing.
- Dan

On Jan 18, 2008 10:50 AM, Loye Young <email address hidden> wrote:
> On Jan 18, 2008 10:34 AM, dankegel <email address hidden> wrote:
> > Anything that crashes Xorg is an Xorg (or driver)
> > bug. It cannot be a wine bug. Please reassign.
>
> @Dankegel
> You contribute much to our community, and I for one appreciate your
> participation, but I have to disagree with you on this one.
>
> While I do not know the particulars of this bug, I know that the
> categorical statement above is not true. Whether it's a bug in Wine or
> Xorg depends on the facts. Killing the session may be the right thing
> for Xorg to do if another program misbehaves badly enough.
>
> Documents created from Microsoft Office components in particular are
> susceptible to extremely bad code execution. The biggest reason is
> that the VBasic scripting language that is embedded in Office macros
> makes it possible to embed dangerous code in a document, sometimes
> without realizing it. This turns out to be a prime suspect in such
> cases because macro execution is the only technical reason to run
> MSOffice on a Linux box in the first place. OpenOffice can read and
> write the MS formats as well or better than MSOffice itself. However,
> many companies have developed documents with complex and sometimes
> sophisticated macros and scripts, which OpenOffice cannot yet run with
> any safety. Consequently, such companies are often compelled to use
> MSOffice, even when they have otherwise converted to open source.
>
> Again, I don't know the specifics of this case, but we can't just
> dismiss the possibility that the issue should be fixed in Wine. A
> program such as Wine that undertakes to run code written for another
> operating system should generally be responsible for containing any
> malfeasance that the code causes.
>
> Happy Trails
>
> --
> Loye Young
> Isaac & Young Computer Company
> Laredo, Texas
> http://www.iycc.biz
>
>
> --
> Xorg crashes after using Microsoft Word 2003 in Wine.
> https://bugs.launchpad.net/bugs/183922
> You received this bug notification because you are a member of Ubuntu
> Wine Team, which is a bug assignee.
>

--
Wine for Windows ISVs: http://kegel.com/wine/isv

I just sent a note to the x-ubuntu list asking for clarification on
whether X is allowed to crash. Let's see what they say.

Here's my message:
https://lists.ubuntu.com/archives/ubuntu-x/2008-January/000072.html

On Jan 18, 2008 1:19 PM, dankegel <email address hidden> wrote:
> You seem to underestimate the requirements for X.
> X is supposed to be bulletproof and
> immune to crashes caused by non-root apps.
> Any chink in its armor is a bug in X or its drivers.
> One should be able to run ANY userspace program
> (that does not DOS the system in general)
> and have X stay up. If this seems unreasonably
> demanding of X, go ask the X maintainers, they'll
> tell you the same thing.
> - Dan
>
One of the many reasons I appreciate your participation in the
community is that you demand excellence. You are persuasive that
complete and absolute stability should be required of X. That does not
imply, however, that this bug should be removed from Wine's bug list.

"The Right Thing(tm)" is for all of us to expect that that _every_
package be "bulletproof and immune to crashes caused by non-root
apps."

Happy Trails,

Loye

Thanks for bringing this to X's attention Dan.

Yes, in fact what you've stated sums up what I'd consider to be requirement #1 for X: Non-root applications should not cause X to crash.

I generally consider if an app is crashing X, it indicates two bugs: First, a bug in the app for being naughty and causing X to crash, and second a bug in X for allowing the app to cause the crash in the first place.

It's not unlikely that the fix for making X not crash would involve returning an error to the app. Thus the app would still be broken, just less severely, and would still need some sort of fix.

Anyway, so I've added an xorg component to this bug in addition to the wine component. Both will need fixes I suspect.

acuozzo, thanks for attaching the Xorg.0.log. Unfortunately, it does not show error messages or a backtrace relating to the crash; usually this appears at the end of the log. You might look to see if you have an Xorg.0.log.old that contains the crash backtrace.

Make sure you have the dbg symbols for xserver and your video driver installed. Info on this is available at https://wiki.ubuntu.com/DebuggingProgramCrash

If that doesn't have a backtrace (and sometimes even if it does), we may need to have you gather a better backtrace. Instructions on doing this for Xorg are available at https://wiki.ubuntu.com/DebuggingXorg. Since you're experiencing an exit crash rather than a lockup, I'd recommend sshing into the box and invoking X in gdb, run it until it crashes, and then use the command 'backtrace full' to get the trace. Copy that into a file and attach it to this bug report.

Once we know what X call is getting triggered to make X exit, it'll help identify where checks can be inserted to prevent the behavior from occurring.

I would just like to confirm that this happens to me as well under Wine 0.9.54 and Ubuntu 7.10.

Is this still occurring with the latest Wine and X in Hardy?

Doesn't seem to be a problem with latest Wine.

Closing the bug since there hasn't been any confirmation that the crash occurs with 8.04 or 8.10. If it still does, please attach a proper backtrace and reopen the bug. Debugging instructions at wiki.ubuntu.com/X.