Re: Why would native code in a java process fail - but only if started by wotaskd?
Re: Why would native code in a java process fail - but only if started by wotaskd?
- Subject: Re: Why would native code in a java process fail - but only if started by wotaskd?
- From: Hugi Thordarson <email@hidden>
- Date: Mon, 24 Aug 2009 18:23:47 +0000
Hi Aaron,
Yet another very good suggestion - but all my apps run with
java.awt.headless=true, so unfortunately that's not the culprit :-/.
The real headscratcher is that everything works fine the first time
wotaskd starts the applications after a clean reboot. I'm not seeing
what happens with a manual stop/start in JavaMonitor that causes the
app to fail. However, I hope I can tell you soon :).
Cheers,
- hugi
// Hugi Thordarson
// http://hugi.karlmenn.is/
On 24.8.2009, at 14:36, email@hidden wrote:
Hi All,
Here's a thought. When you start up from Terminal do you see
anything new bouncing in the dock? If so, that means that it
requires access to window services to launch.
Launching via JavaMonitor as "root" often means you do not have
access to window services. That could cause an immediate crash when
it tries to obtain it. Java AWT calls often require this, your
native code might too. Usually there are ways to get around this but
its situation is different. If it is a carbon app you are connecting
to there are a few options in the info.plist of the carbon app
bundle you need to flip to make it run "headless".
-- Aaron _______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Webobjects-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden