[Bug 12740] New: Wine won't start (regression against 0.9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

Summary: Wine won't start (regression against 0.9.59)
           Product: Wine
           Version: 0.9.60
          Platform: PC
        OS/Version: Solaris
            Status: UNCONFIRMED
          Severity: blocker
          Priority: P1
         Component: loader
        AssignedTo: [e-mail]
        ReportedBy: [e-mail]

Created an attachment (id=12395)
 --> (http://bugs.winehq.org/attachment.cgi?id=12395)
winecfg error output

With version 0.9.60, Wine will immediately crash. See attached output from
winecfg.

Previous version 0.9.59 works just fine.

[Bug 12740] Wine won't start (regression against 0.9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

Petr Sumbera  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |regression

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

Petr Sumbera  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Wine won't start (regression|Wine tries to execute non-
                   |against 0.9.59)             |executable data (regression
                   |                            |against 0.9.59)

--- Comment #1 from Petr Sumbera   2008-04-22 16:43:50 ---
Wine tries to execute non-executable data. As can be seen also from system log:

Apr 22 22:53:23 ferrari genunix: [ID 269049 kern.notice] NOTICE: wine[5044]
attempt to execute non-executable data at 0x0 by uid xxxx
Apr 22 22:53:23 ferrari last message repeated 1 time
Apr 22 22:53:23 ferrari genunix: [ID 269049 kern.notice] NOTICE: wine[5033]
attempt to execute non-executable data at 0x0 by uid xxxx

This is related to libwine as Wine will run when libwine.so.1.0 from version
0.9.59 is preloaded.

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

--- Comment #2 from Austin English   2008-04-22 23:49:37 ---
Please run a regression test:
http://wiki.winehq.org/RegressionTesting

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

Petr Sumbera  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |[e-mail]
          Component|loader                      |winex11.drv

--- Comment #3 from Petr Sumbera   2008-04-23 14:54:52 ---
I ran regression tests and the winner is...

--

185157cb6eb25bc5e75e9a260b593e98153d2255 is first bad commit
commit 185157cb6eb25bc5e75e9a260b593e98153d2255
Author: Kusanagi Kouichi slash at ma.neweb.ne.jp
Date:   Fri Apr 4 20:44:36 2008 +0900

winex11: Connect to XIM server dynamically.

:040000 040000 4df81736d311abdd4b603e64987fb90e4e5c9a71
7999fe03f030755e4e731116a1701b0779e07ab7 M      dlls

--

Together with following proposed fix:

http://www.winehq.org/pipermail/wine-devel/2008-April/064839.html

--

...basically it's partly bug against my bad testing (I just verified it
builds).

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

Marcus Meissner  changed:

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

--- Comment #4 from Marcus Meissner   2008-04-25 09:11:55 ---
looks like a NULL ptr jump.

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

--- Comment #5 from Petr Sumbera   2008-04-25 12:11:04 ---
It's somehow related to using XRegisterIMInstantiateCallback(). Disabling it on
following place will make wine (winecfg) start:

--- dlls/winex11.drv/xim.c.orig Fri Apr 25 18:51:50 2008
+++ dlls/winex11.drv/xim.c      Fri Apr 25 18:52:21 2008
@@ -484,7 +484,7 @@
 void X11DRV_SetupXIM(void)
 {
     wine_tsx11_lock();
-    XRegisterIMInstantiateCallback(thread_display(), NULL, NULL, NULL,
X11DRV_OpenIM, NULL);
+    //XRegisterIMInstantiateCallback(thread_display(), NULL, NULL, NULL,
X11DRV_OpenIM, NULL);
     wine_tsx11_unlock();
 }

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

Petr Sumbera  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|blocker                     |critical

--- Comment #6 from Petr Sumbera   2008-04-28 10:43:40 ---
With recent sync to Solaris (build 85) Xlib i18n routines with the community
this problem doesn't appear again (Wine is running and XICCallback is defined).

See:

http://www.opensolaris.org/jive/message.jspa?messageID=228337#228337

But this still should be fixed as Wine should run also on older Solaris
versions (Solaris 10, ..).

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

--- Comment #7 from Kusanagi Kouichi   2008-04-29 02:53:57 ---
Created an attachment (id=12564)
 --> (http://bugs.winehq.org/attachment.cgi?id=12564)
patch

Does this solve the problem?
I wish there were a better solution.

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

--- Comment #8 from Petr Sumbera   2008-04-29 14:21:47 ---
Hi Kusanagi,

I can run Wine with your patch just fine. Not sure what is exact limitation but
I think it's the only way how to solve it.

Please submit this patch. Thanks!

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

Albert Lee  changed:

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

--- Comment #9 from Albert Lee   2008-04-29 17:47:32 ---
Please see http://www.opensolaris.org/ive/thread.jspa?messageID=229930 for
info. The crash is caused by XRegisterIMInstantiateCallback being totally
broken. It's fixed in snv_85 because of the new l18n code. A runtime check to
revert to immediate XIM should be possible.

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

--- Comment #10 from Kusanagi Kouichi   2008-04-30 08:20:21 ---
Created an attachment (id=12592)
 --> (http://bugs.winehq.org/attachment.cgi?id=12592)
runtime check patch

This might be wrong or imcomplete.

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

Austin English  changed:

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

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

--- Comment #11 from Albert Lee   2008-05-01 03:44:50 ---
Created an attachment (id=12612)
 --> (http://bugs.winehq.org/attachment.cgi?id=12612)
Second try at runtime check

The runtime check should not assume you run Nevada... I've modified it to use
dlsym to check for a symbol I assume might be removed in the new library.

I've also added a suggestion for the missing struct definitions (instead of
using #define).

Unfortunately I don't have an snv_85+ system handy to check on.
Someone with snv_85+ or Indiana should make sure nm -g /usr/lib/libX11.so |
grep _XInitGenericIM doesn't return anything.

http://www.opensolaris.org/jive/thread.jspa?threadID=58305&tstart=0 (I typoed
the previous link) is the thread on xwin-discuss for this problem. I have
attached  a testcase for the crash there.

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

Albert Lee  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #12612|0                           |1
        is obsolete|                            |

--- Comment #12 from Albert Lee   2008-05-13 20:31:40 ---
Created an attachment (id=13040)
 --> (http://bugs.winehq.org/attachment.cgi?id=13040)
Updated fallback-to-XIM runtime check.

Based on the information provided by    
xsumbe00 on http://www.opensolaris.org/jive/thread.jspa?messageID=231794 
that patch was the "correct" way to avoid the problem.

I've updated it for 1.0-rc1. The same patch is in SFEwine.

[Bug 12740] Wine tries to execute non-executable data (regression against 0 .9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

--- Comment #13 from Petr Sumbera   2008-05-14 03:22:02 ---
I have filed bug against implementation of XRegisterIMInstantiateCallback in
Solaris 10 and 9:

http://bugs.opensolaris.org/view_bug.do?bug_id=6698109

If it turns out that it can be fixed in Solaris 10 and 9 I would prefer not to
have special patch for this in Wine.

[Bug 12740] Wine tries to execute non-executable data (regression against 0.9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

Alexandre Julliard  changed:

What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |1.0.1

--- Comment #14 from Alexandre Julliard   2008-06-21 05:47:13 ---
Should be mostly fixed by 5b7bec34c1b550e26744fe632a55e196a751eab4. It will
still crash when switching XIMs dynamically but that's probably good enough for
now.

[Bug 12740] Wine tries to execute non-executable data (regression against 0.9.59)

http://bugs.winehq.org/show_bug.cgi?id=12740

Alexandre Julliard  changed:

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

--- Comment #15 from Alexandre Julliard   2008-06-27 10:15:28 ---
Closing bugs fixed in 1.1.0.