[Bug 14863] New: Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

Summary: Crash in application / works with version 0.9.36
           Product: Wine
           Version: 1.1.2
          Platform: PC
        OS/Version: Linux
            Status: UNCONFIRMED
          Severity: critical
          Priority: P2
         Component: -unknown
        AssignedTo: [e-mail]
        ReportedBy: [e-mail]

Created an attachment (id=15406)
 --> (http://bugs.winehq.org/attachment.cgi?id=15406)
Output of Wine

There is an application that crashes with the recent versions of Wine (I tested
1.2.0 and 1.0.0). The same application worked well with version 0.9.36.

I also have seen a crash with that application and the newer version that was
caused somewhere in an installed dbgrid32.dll but currently I'm unable to
reproduce that. I'll submit it as soon as I have it again.

The .wine directory was completely clean before installation, I set up
everything new for that Wine version.

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

Dmitry Timoshkov  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|critical                    |normal
           Keywords|                            |regression

--- Comment #1 from Dmitry Timoshkov   2008-08-14 08:36:07 ---
Not critical. If the application worked before, please run the regression
test: http://wiki.winehq.org/RegressionTesting

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #2 from Vijay Kamuju   2008-08-14 09:23:15 ---
In your bug reports please mention the application name

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #3 from Hadron   2008-08-18 00:08:37 ---
I afraid regression testing is not a solution: I know it worked with 0.9.36 and
I know it doesn't works with 1.0.0. So I can't simply tell GIT to test between
0.9.36 and 0.9.37, there are several more versions to try.

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #4 from James Hawkins   2008-08-18 00:13:22 ---
How does that make regression testing impossible?  You set the good to 0.9.36
and the bad to 1.0.0 then run the test.  The regression test doesn't have to be
run between two consecutive versions.  The document linked to in comment #1
explains the whole process.

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #5 from Hadron   2008-08-18 02:20:35 ---
OK, I did not realize that I can use versions that do not follow directly on
each other.

But here is the next problem: when compiling 0.9.37 I fail with this error:

parse.o: In function `ldap_parse_sort_controlW':
/root/wine-0.9.37/dlls/wldap32/parse.c:339: undefined reference to
`ldap_parse_sort_control'
parse.o: In function `ldap_parse_vlv_controlW':
/root/wine-0.9.37/dlls/wldap32/parse.c:420: undefined reference to
`ldap_parse_vlv_control'
/root/wine-0.9.37/dlls/wldap32/parse.c:420: undefined reference to
`ldap_parse_vlv_control'

openldap-devel is installed and ./configure did not gave me any error. So what
could be the reason for that?

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

Louis Lenders  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #6 from Louis Lenders   2008-08-18 02:33:27 ---
(In reply to comment #5)
> OK, I did not realize that I can use versions that do not follow directly on
> each other.
>
> But here is the next problem: when compiling 0.9.37 I fail with this error:
>
> parse.o: In function `ldap_parse_sort_controlW':
> /root/wine-0.9.37/dlls/wldap32/parse.c:339: undefined reference to
> `ldap_parse_sort_control'
> parse.o: In function `ldap_parse_vlv_controlW':
> /root/wine-0.9.37/dlls/wldap32/parse.c:420: undefined reference to
> `ldap_parse_vlv_control'
> /root/wine-0.9.37/dlls/wldap32/parse.c:420: undefined reference to
> `ldap_parse_vlv_control'
>
> openldap-devel is installed and ./configure did not gave me any error. So what
> could be the reason for that?

I too ran into this a few days ago.The easiest way ( but  a bit tricky i admit)
is just do 'make -k', instead of 'make'. The regression itsself is most likely
not in wldap32
> So what could be the reason for that?

If you compile very old wine-versions on a new system you'll run into these
things unfortunately as headerfiles etc from your system get updated too, and
old wine-versions don't know about that of course. I had some MB's free on an
old laptop, and installed Fedora Core 3. Such an old system will happily
compile all old wine-versions back to 2005.

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #7 from Hadron   2008-08-18 04:00:31 ---
OK, the regression testing fails during the very first run of it:

{standard input}: Assembler messages:
{standard input}:5636: Warning: end of file not at end of a line; newline
inserted
{standard input}:7022: Error: suffix or operands invalid for `mov'
gcc: Internal error: Segmentation fault (program cc1)
Please submit a full bug report.
See http://bugzilla.redhat.com/bugzilla for instructions.

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #8 from Dmitry Timoshkov   2008-08-18 04:05:39 ---
(In reply to comment #7)
> OK, the regression testing fails during the very first run of it:
> {standard input}: Assembler messages:
> {standard input}:5636: Warning: end of file not at end of a line; newline
> inserted
> {standard input}:7022: Error: suffix or operands invalid for `mov'
> gcc: Internal error: Segmentation fault (program cc1)
> Please submit a full bug report.
> See http://bugzilla.redhat.com/bugzilla for instructions.

Have you looked there for instructions as you were asked?

(That's the bug 13532)

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #9 from Louis Lenders   2008-08-18 04:09:57 ---
to summarize quickly:
do:

ulimit -s unlimited

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #10 from Hadron   2008-08-18 06:43:21 ---
Thanks for all the help but I afraid there is an other problem now where also
Google and the Bugzilla search function did not help me.

With the regression testing I now reached version wine-0.9.43-50-g7f901f4

Here I get the LDAP-compileerrors and override the with "make -k". The
resulting version doesn't works, here I get following error:

wine: creating configuration directory '/home/testuser/.wine'...
wine: could not load L"c:\windows\system32\rundll32.exe": Module not found
wine: wineprefixcreate failed while creating '/home/testuser/.wine'.

The directory /home/testuser/.wine is never created and when I create it
manually I get an error

Warning: the specified Windows directory L"c:\windows" is not accessible.
Warning: the specified System directory L"c:\windows\system32" is not
accessible.
Warning: could not find DOS drive for current working directory
'/home/testuser/Download', starting in the Windows directory.
wine: could not load L"c:\windows\system32\setup_v_2_9.exe": Module not
found

It seems like it is a problem of that specific version. Is there a trick to
avoid that? Elsewhere I'll restart regression with a different bad release
number.

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #11 from Austin English   2008-08-18 10:30:12 ---
Disable tests with the patch on the regression testing page.

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #12 from Hadron   2008-08-19 00:54:57 ---
Seems like that patch doesn't works:

patch -p1 < test.patch.diff && CC="ccache gcc" ./configure && make clean &&
make depend && make && patch -p1 -R < disable_test.diff
patching file dlls/Makefile.in
patch: **** malformed patch at line 12: diff --git a/programs/Makefile.in
b/programs/Makefile.in

I'm sorry but what you expect here from somebody who is not a
full-time-programmer is really too much. I'm stumbling from one non-working
manual to the next and don't get closer to what I really want: a Wine version
that works as good as the old 0.9.36!

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #13 from Hadron   2008-08-19 07:48:22 ---
So what else can I do here?

- I know version 0.9.36 is working, version 0.9.50 and later are not
- regression testing fails because after some time I create a version that
doesn't works with a strange error where nobody can help me (.wine directory
error, see above)
- applying a special regression testing patch fails because that patch doesn't
works (malformed patch, please see above)
- reverse regression testing is running but for the moment I'm searching a
minor compile problem in order to find a patch where I would guess that doesn't
works on the older releases of the software - means I'm far away from the real
regression test that will show me the patch that damaged Wine
> From reading that I think it would be much more easier for a developer to check

the bug directly within the sources instead for me as non-developer to waste
hours and hours and days after days with manuals and testing methods  that do
not work for me.

I'm really willing to help here but I afraid I have reached the limit of my
personal knowledge and capabilities...

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #14 from Louis Lenders   2008-08-19 08:34:58 ---
Does the app have a free download? Then we could test it ourselves

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #15 from Austin English   2008-08-19 09:53:08 ---
(In reply to comment #13)
> So what else can I do here?
> - regression testing fails because after some time I create a version that
>
> doesn't works with a strange error where nobody can help me (.wine directory
> error, see above)

$ rm -rf ~/.wine
> - applying a special regression testing patch fails because that patch doesn't
>
> works (malformed patch, please see above)

Try downloading the patch from the bottom link. Cutting and pasting it gets
malformed for some reason...or manually make the changes (there are only a
couple lines to change...

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #16 from Hadron   2008-08-20 00:17:39 ---
The application is from http://www.scaps.com/downloads/index.php , the direct
download link is
http://download.scaps.net/downloads/v_2_9_5/?c=1&f=sc_sam_setup_v_2_9_5_...

Btw: I removed .wine for every test - that did not help really...

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                URL|                            |http://download.scaps.net/do
                   |                            |wnloads/v_2_9_5/?c=1&f=sc_sa
                   |                            |m_setup_v_2_9_5_20080819.exe
           Keywords|                            |download

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #17 from Louis Lenders   2008-08-20 04:53:53 ---
Created an attachment (id=15499)
 --> (http://bugs.winehq.org/attachment.cgi?id=15499)
reverted patch

Here you go:
54f2f60b5489c31188eae8d25871ed19ffe96b31 is first bad commit
commit 54f2f60b5489c31188eae8d25871ed19ffe96b31
Author: Huw Davies 
Date:   Wed Aug 8 12:36:37 2007 +0100

ole32: The default handler needs to have its own implementation of
IPersistStorage.

For now we forward the methods to the data cache as before.

:040000 040000 17e1c586ceeaa927bf7312d8897eb6bbc448117d
80531c5e09b4dad1e60736b06532234f6dbe0282 M      dlls

The attached patch is basically the revert of that patch against current git,
and makes the app start fine for me.

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

Louis Lenders  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]
             Status|UNCONFIRMED                 |NEW
          Component|-unknown                    |ole32
     Ever Confirmed|0                           |1

--- Comment #18 from Louis Lenders   2008-08-20 04:54:25 ---
added author patch

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #19 from Damjan Jovanovic   2008-08-21 08:49:13 ---
Created an attachment (id=15521)
 --> (http://bugs.winehq.org/attachment.cgi?id=15521)
do not release the DataCache

It's some kind of reference leak, hacking it to never release the DataCache
fixes the bug.

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

Damjan Jovanovic  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

[Bug 14863] Crash in application / works with version 0.9.36

http://bugs.winehq.org/show_bug.cgi?id=14863

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

Lei Zhang  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Crash in application / works|Crash in SCAPS SAM
                   |with version 0.9.36         |

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #20 from Hadron   2008-08-26 00:44:52 ---
Created an attachment (id=15649)
 --> (http://bugs.winehq.org/attachment.cgi?id=15649)
Crash Backtrace

I tried it with the latest development version wine-1.1.3-g9ce31cf but the
problem still seems to be there? Please refer to the attached crashlog, for me
it seems like it is the same problem.

Or was'nt the bugfix included yet? :-)

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #21 from Hadron   2008-10-14 03:33:15 ---
Could somebody with developer privilieges please put the provided patch to the
main development branch? The latest versions do not contain it so that this bug
is still there...

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #22 from Dmitry Timoshkov   2008-10-14 03:40:09 ---
(In reply to comment #21)
> Could somebody with developer privilieges please put the provided patch
> to the main development branch? The latest versions do not contain it so
> that this bug is still there...

First of all there was no any feedback on the possible workaround, does it
really help? Secondly, that's just a hack, no a real solution.

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #23 from Damjan Jovanovic   2008-10-14 03:44:00 ---
In case I wasn't clear, my hack to never release the DataCache does work.

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #24 from Hadron   2008-10-17 06:14:59 ---
So Damjans patch could be added!?

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #25 from Austin English   2008-10-17 11:29:13 ---
(In reply to comment #24)
> So Damjans patch could be added!?

It's a hack, not a real solution. If someone wants to take it and flesh it out
better, then yes.

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #26 from Hadron   2008-10-20 00:20:33 ---
Hm, so it is better to have non-working code there that causes a crash instead
of a (not so perfect) solution that lets Wine work?

Btw: did the original author of the buggy code ever gave a statement?

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

Hadron  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
         Difficulty|---                         |Minutes

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #27 from Huw Davies   2008-11-07 03:32:44 ---
(In reply to comment #26)
> Hm, so it is better to have non-working code there that causes a crash instead
> of a (not so perfect) solution that lets Wine work?
>
> Btw: did the original author of the buggy code ever gave a statement?

I've made quite a few changes to the default handler recently, have you tried
the current git?

And please don't set the difficulty level of the bug if you don't understand
it.  It's clearly not fixable in minutes.

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #28 from Hadron   2008-11-07 05:43:52 ---
I did not notice any changes in GIT regarding that problem, I'll test them as
soon as possible.

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

Dmitry Timoshkov  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
         Difficulty|Minutes                     |---

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #29 from Hadron   2008-11-10 02:50:44 ---
No, still the same problem, application crashes during startup.

You can check it easily on your own system: install the application from
http://download.scaps.net/downloads/v_2_9_5/?c=1&f=sc_sam_setup_v_2_9_5_...
(it's free) and then execute the installed application SAMLight to see the
problem.

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #30 from Hadron   2008-12-04 03:51:33 ---
> And please don't set the difficulty level of the bug if you don't
> understand it. It's clearly not fixable in minutes.

There is a patch available - so why can't it be used within minutes? Or how
about switching back to the previous state where the software without that
modification worked?

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #31 from Austin English   2008-12-04 12:40:44 ---
(In reply to comment #30)
> > And please don't set the difficulty level of the bug if you don't
> > understand it. It's clearly not fixable in minutes.
>
> There is a patch available - so why can't it be used within minutes? Or how
> about switching back to the previous state where the software without that
> modification worked?

The patch is a hack.

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #32 from Hadron   2008-12-08 00:39:39 ---
> The patch is a hack.

I understand that. But the current implementation is not working. So is a crash
better than a hack that would work?

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #33 from Jeff Zaroyko   2008-12-08 00:50:35 ---
(In reply to comment #32)
> > The patch is a hack.
>
> I understand that. But the current implementation is not working. So is a crash
> better than a hack that would work?

You can apply it to your version of Wine, the patch that broke your application
may have fixed or vastly improved others.  It's called progress.  If you can't
provide a real solution then you will just have to wait for someone else to
look at the issue.

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

Austin English  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]

--- Comment #34 from Austin English   2008-12-08 15:10:17 ---
*** Bug 15870 has been marked as a duplicate of this bug. ***

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

--- Comment #35 from Rob Shearman   2009-11-16 17:19:20 ---
Bug 12419 may be related to this bug.

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

Rob Shearman  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED

--- Comment #36 from Rob Shearman   2009-11-24 15:27:51 ---
Fixed by this commit:
http://source.winehq.org/git/wine.git/?a=commit;h=6592c25bc749373cad0e7584ef...

[Bug 14863] Crash in SCAPS SAM

http://bugs.winehq.org/show_bug.cgi?id=14863

Alexandre Julliard  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |CLOSED

--- Comment #37 from Alexandre Julliard   2009-12-04 12:15:39 ---
Closing bugs fixed in 1.1.34.