[vbox-dev] python: launchVMProcess(session, 'sdl', None) segfaults

[Nikolay Igotti]

    Hi Francisco,


 Reproduced and fixed issue, fix will appear in next maintenance 
release. But as you're Gentoo guy :), try this patch:


 Index: src/VBox/Frontends/VBoxSDL/VBoxSDL.cpp
===================================================================
--- src/VBox/Frontends/VBoxSDL/VBoxSDL.cpp      (revision 69582)
+++ src/VBox/Frontends/VBoxSDL/VBoxSDL.cpp      (working copy)
@@ -1332,10 +1334,22 @@
     }
 
     /*
-     * Do we have a name but no UUID?
+     * Do we have a UUID?
      */
-    if (vmName && uuidVM.isEmpty())
+    if (!uuidVM.isEmpty())
     {
+        rc = pVirtualBox->FindMachine(uuidVM.toUtf16().raw(), 
pMachine.asOutParam());
+        if (FAILED(rc) || !pMachine)
+        {
+            RTPrintf("Error: machine with the given ID not found!\n");
+            goto leave;
+        }
+    }
+    else if (vmName)
+    {
+        /*
+         * Do we have a name but no UUID?
+         */
         rc = pVirtualBox->FindMachine(Bstr(vmName).raw(), 
pMachine.asOutParam());
         if ((rc == S_OK) && pMachine)
         {


  Thanks for report,
      Nikolay


Nikolay Igotti wrote:
> 20.01.2011 13:09, Francisco J. Vázquez пишет:
>>> Any chance to inspect core generated by
>>> VBoxSDL (as it was killed with SIGSEGV).
>>>
>> I would if I knew how :). Running gdb python and pasting the previous
>> commands doesn't work. gdb doesn't "see" the segfault:
> Nope, that wouldn't help, VM process is launched with fork/exec or 
> spawn. Usually, if core dump generation is enabled,
> see http://www.virtualbox.org/wiki/Core_dump, you'll have core file 
> generated for crashed processes.
> Open it with gdb -c .  Hope that'll help.
>
>   Nikolay
>
>
>
> _______________________________________________
> vbox-dev mailing list
> vbox-dev at virtualbox.org
> http://vbox.innotek.de/mailman/listinfo/vbox-dev