[scilab-Users] Scilab 5.2.2 fails to start on Mac OS 10.6.3

[Jeff DuMonthier]

Unfortunately the issue of being unable to start scilab on some Mac's, including my iMac is still present, even in a beta's of 5.3.

I ran 5.3 beta 3 in debug mode and after it stopped responding did a backtrace in gdb.  This was the output:

. done
Reading symbols for shared libraries .... done
JavaVM: requested Java version (1.5) not available. Using Java at "/System/Library/Frameworks/JavaVM.framework/Versions/1.6.0/Home" instead.
Reading symbols for shared libraries . done
Reading symbols for shared libraries . done
Reading symbols for shared libraries .. done
Reading symbols for shared libraries . done
Reading symbols for shared libraries . done
Reading symbols for shared libraries . done
^C
Program received signal SIGINT, Interrupt.
0x00007fff80a0f2fa in mach_msg_trap ()
(gdb) bt
#0  0x00007fff80a0f2fa in mach_msg_trap ()
#1  0x00007fff80a0f96d in mach_msg ()
#2  0x00007fff8312c3c2 in __CFRunLoopRun ()
#3  0x00007fff8312b84f in CFRunLoopRunSpecific ()
#4  0x00007fff80f8491a in RunCurrentEventLoopInMode ()
#5  0x00007fff80f8471f in ReceiveNextEventCommon ()
#6  0x00007fff80f845d8 in BlockUntilNextEventMatchingListInMode ()
#7  0x00007fff8166729e in _DPSNextEvent ()
#8  0x00007fff81666bed in -[NSApplication nextEventMatchingMask:untilDate:inMode:dequeue:] ()
#9  0x00007fff8162c8d3 in -[NSApplication run] ()
#10 0x0000000132edc4b0 in +[AWTStarter startAWT:] ()
#11 0x0000000132ea02c9 in -[CPerformer perform] ()
#12 0x00007fff8719c63b in __NSThreadPerformPerform ()
#13 0x00007fff8312de91 in __CFRunLoopDoSources0 ()
#14 0x00007fff8312c089 in __CFRunLoopRun ()
#15 0x00007fff8312b84f in CFRunLoopRunSpecific ()
#16 0x00007fff8312b5d6 in CFRunLoopRun ()
#17 0x0000000100015fdb in initMacOSXEnv ()
#18 0x0000000100003bbd in ?? ()
#19 0x00000001000039a4 in ?? ()


I did try reinstalling Java 1.5 from a previous OS update using Pacifist and it just crashed in a different way.  I think that Java 1.5 may not be entirely compatible with OS 10.6, which is probably why it was removed.

If anyone has suggestions on how/where to set breakpoints and what else to look for, I'd be glad to try it.  


On May 10, 2010, at 3:46 PM, Sylvestre Ledru wrote:

> Le lundi 10 mai 2010 à 12:49 -0400, Jeff DuMonthier a écrit :
>> I was running Scilab 5.2.1 successfully, but after installing 5.2.2 and a few other things, Scilab will no longer start up.  I've tried going back to 5.2.1 and have the same problem, so the issue is probably with something else I installed, not a change from 5.2.1 to 5.2.2.  The most likely culprits are an Apple system update combined with an extension, or Python, which I installed to get something else to work.  I'm not sure if Scilab uses Python or something that gets stepped on by the installation, but it is one of the things that changed.
>> 
>> When I attempt to start Scilab, the icon in the dock changes to a generic Java icon and it basically hangs until I do a force quit.  These are the messages I see in the console:
>> 
>> 5/10/10 12:27:58 PM	[0x0-0x72072].org.scilab.modules.jvm.Scilab[606]	JavaVM: requested Java version (1.5) not available. Using Java at "/System/Library/Frameworks/JavaVM.framework/Versions/1.6.0/Home" instead.
>> 5/10/10 12:28:24 PM	com.apple.spindump[632]	ConnectPort: can't find mach server port name = jackdmp_entry.0_default err = unknown error code
>> 5/10/10 12:28:24 PM	com.apple.spindump[632]	Cannot connect to server Mach port
>> 5/10/10 12:28:24 PM	com.apple.spindump[632]	jack server is not running or cannot be started
>> 5/10/10 12:28:25 PM	com.apple.launchd.peruser.502[153]	([0x0-0x72072].org.scilab.modules.jvm.Scilab[606]) Exited with exit code: 143
>> 
>> 
>> The first message appears immediately.  The rest appear when I force quit Scilab.  I also have Scilab installed and working on another Mac running the same system version.  The message about the Java version also shows up in the console on that machine, but since it works there I don't think that's the issue.  The other messages seem to be referring to JackOSX, which is a configurable audio router.  I don't know why Scilab would be trying to start or connect to that.  It's not set up as the default audio device and Scilab was working with the same version of that installed before.  I can only guess that an Apple OS update or the Python install have created an extension conflict with an extension that wasn't a problem before.  The machine where Scilab is working has the same system version but I did not install Python or JackOSX.  Does anyone have any suggestions as to what I could try to debug this further?  I'd rather not have to start ripping out extensions or cleaning out the Python install, which there doesn't seem to be an easy way to do.
> 
> We don't have this issue on our testing Mac OS X machines and I confirm
> that we are not doing anything with the sound card on the start up of
> the software.
> However, we noticed some random behaviors under Mac. Please let me know
> if you find a workaround on this issue.
> 
> Sylvestre
> 
> -- 
> -------------------------
> Sylvestre Ledru
> -------------------------
> The Scilab Consortium
> Digiteo
> Domaine de Voluceau
> Rocquencourt - B.P. 105
> 78153 Le Chesnay Cedex
> France
>