You're right...I forgot the shutdown hook....Here 'tis with the shutdown
hook:
import org.ops4j.pax.runner.platform.internal.Pipe;
public class BrowserStarter {
static Pipe err;
static Pipe out;
static Pipe in;
public static final void main(String[] args) throws Exception {
try {
Process p = Runtime.getRuntime().exec(args[0]);
p.waitFor();
err = new Pipe( p.getErrorStream(), System.err ).start( "error" );
out = new Pipe( p.getInputStream(), System.out ).start( "out" );
in = new Pipe( p.getOutputStream(), System.in ).start( "in" );
destroyFrameworkOnExit(p, new Pipe[] {err,out,in});
} finally {
if(err != null) err.stop();
if(out != null) out.stop();
if(in != null) in.stop();
}
}
private static void destroyFrameworkOnExit( final Process process,
final Pipe[] pipes )
{
Runtime.getRuntime().addShutdownHook( new Thread( new Runnable()
{
public void run()
{
try
{
for ( Pipe pipe : pipes )
{
pipe.stop();
}
}
finally
{
process.destroy();
}
}
}
)
);
}
}
I had high hopes that this would be it. But when I started the service,
it spiked (a little higher this time at 63%), then back down to 0%.
So...I suppose that I should try profiling it. Great suggestion. I'll
give it a shot and report back what I find.
Alin Dreghiciu wrote:
> Looks like the only thing is not there now is the shutdown hook. which
> is a simple thread that stops the pipes and destroys the started
> process.
> But, can't we figure out what is the tread that is consuming the CPU
> using a java profiler? We have an open source license for JProfiler.
>
> On Thu, Jul 31, 2008 at 3:27 PM, Craig Walls <[EMAIL PROTECTED]> wrote:
>
>> So far, no good. Here's what BrowserStarter.java looks like now:
>>
>> import org.ops4j.pax.runner.platform.internal.Pipe;
>>
>> public class BrowserStarter {
>> static Pipe err;
>> static Pipe out;
>> static Pipe in;
>>
>>
>> public static final void main(String[] args) throws Exception {
>> try {
>> Process p = Runtime.getRuntime().exec(args[0]);
>> p.waitFor();
>>
>> err = new Pipe( p.getErrorStream(), System.err ).start( "error" );
>> out = new Pipe( p.getInputStream(), System.out ).start( "out" );
>> in = new Pipe( p.getOutputStream(), System.in ).start( "in" );
>> } finally {
>> if(err != null) err.stop();
>> if(out != null) out.stop();
>> if(in != null) in.stop();
>> }
>> }
>> }
>>
>> And my jsl.ini's cmdline looks like this:
>>
>> cmdline = -cp c:/semantra/bin;c:/semantra/bin/pax/pax-runner-0.12.0.jar
>> BrowserStarter "java -jar
>> c:/semantra/bin/org.eclipse.osgi_3.4.0.v20080605-1900.jar -console 8888"
>>
>> As before, when I start the service, there's a small spike (about 40% or
>> so), then it settles back down to 0%. Of course, that's a good thing,
>> except that it does nothing to explain why Pax Runner starts up as a
>> service consuming 100% of the CPU.
>>
>> I'm still tinkering and looking around at Pax Runner source on my end to
>> see if I can figure it out. But I wanted to report the status of this
>> experiment to see if you guys have any ideas.
>>
>>
>>
>> Niclas Hedhman wrote:
>>
>>> On Thursday 31 July 2008 05:44, Craig Walls wrote:
>>>
>>>
>>>> public class BrowserStarter {
>>>> public static final void main(String[] args) throws Exception {
>>>> Runtime.getRuntime().exec(args[0]);
>>>> }
>>>> }
>>>>
>>>>
>>> I am also not on Windows (who uses that anyway? ;-) )
>>>
>>> Perhaps change to;
>>>
>>> public class BrowserStarter {
>>> public static final void main(String[] args) throws Exception {
>>> Process p = Runtime.getRuntime().exec(args[0]);
>>> p.waitFor();
>>> }
>>> }
>>>
>>> And if that doesn't do it, look up the Pipe class in Pax Runner and add that
>>> by;
>>>
>>> static Pipe err;
>>> static Pipe out;
>>> static Pipe in;
>>>
>>> err = new Pipe( process.getErrorStream(), System.err ).start( "error" );
>>> out = new Pipe( process.getInputStream(), System.out ).start( "out" );
>>> in = new Pipe( process.getOutputStream(), System.in ).start( "in" );
>>>
>>> and the Shutdown hook to cleanup properly.
>>>
>>>
>>> Cheers
>>>
>>>
>> _______________________________________________
>> general mailing list
>> general@lists.ops4j.org
>> http://lists.ops4j.org/mailman/listinfo/general
>>
>>
>
>
>
>
_______________________________________________
general mailing list
general@lists.ops4j.org
http://lists.ops4j.org/mailman/listinfo/general
