Hi neko list, hi Nicolas,

I come back to this question because I really need to implement a thread dump mechanism for a nekoVM applications.

The concerned application runs two ThreadRemotingServer objects at the same time and I need to know what is happening in all threads at a t instant.

No issue to get the statcktraces of all the threads of the both ThreadRemotingServer objects. However, I want to dump the threads when these remoting servers seem to hang so I cannot send the thread dump instructions through haxe remoting. A solution could be to trigger the thread dumps when receiving a system event on the nekoVM.

My question is thus : how can I catch system events in my haxe/neko code to trigger a function of my application ?

(By system events, I mean what we send with the "kill" shell instruction on Linux, which is not only used to kill a process, depending on which signal we send... Such a mecanism would be similar to what most JVMs propose to take thread dumps)

Thanks in advance !
Thomas Fétiveau Développeur Indépendant Mobile : +33 623 110 922 Tél : +33 297 667 674 web : http://www.tofee.fr/ skype : thomas.fetiveau.tech
Le 04/01/2012 18:24, Nicolas Cannasse a écrit :

Le 04/01/2012 17:13, Thomas Fétiveau a écrit :

Hi,

I would like to know how to take thread dumps on the nekoVM (I'm running
it on Linux).

Is there any web page summing up all diagnosis and troubleshooting
tools/procedures for nekoVm applications ?

There are not many tools for debugging in Neko so far.

If you have a neko.vm.Thread running, you can get it's current callstack by doing the following :

untyped haxe.Stack.makeStack(neko.Lib.load("std", "thread_stack", 1)(t.handle));

Best,
Nicolas

--

-- 
Neko : One VM to run them all
(http://nekovm.org)