Raymond, Sorry I didn't realize that Felix gets added into the SCA manfest.jar file, causing all samples to load Felix even when OSGi contributions are not used.
The change you made to ClassReferenceModelResolver tries to load the class using the classloader first. For OSGi contributions, it should be the other way round - OSGi bundle resolver should be called first, otherwise potentially the wrong classloader could be used to load the classes. I have added a patch for an alternative fix under TUSCANY-1757. It doesn't need to be fixed for 1.0, but I think it should be fixed in the main branch. Could you take a look at it? The change is in OSGiBundleModelResolver, which doesn't start the runtime, but looks for the bundle if the runtime already been started. Thank you... Regards, Rajini On 9/19/07, kelvin goodson <[EMAIL PROTECTED]> wrote: > > RC3 is much faster now, it now runs in 2 seconds --- > > C:\Release\SCA\1.0\RC3\bin\tuscany-sca-1.0-incubating\samples\calculator>ant > run > Buildfile: build.xml > > run: > [java] 3 + 2=5.0 > [java] 3 - 2=1.0 > [java] 3 * 2=6.0 > [java] 3 / 2=1.5 > > BUILD SUCCESSFUL > Total time: 2 seconds > C:\Release\SCA\1.0\RC3\bin\tuscany-sca-1.0-incubating\samples\calculator> > > > for me, thanks. > > On 18/09/2007, Raymond Feng <[EMAIL PROTECTED]> wrote: > > Hi, > > > > I guess I make all of you anxious to know what's wrong :-). > > > > Here is the problem, the OSGi contribution modules starts the OSGi > runtime > > in its contribution listener/class reference resolver. But there is no > code > > to shutdown the OSGi runtime. As a result, there are a few felix threads > > when the runtime is stopped and it hangs. > > > > I'm fixing the problem as follows: > > 1) Make sure the OSGi runtime won't be inited until it's required > > 2) Add a module activator for osgi-runtime to shutdown the OSGi runtime > upon > > the stopping of the Tuscany runtime. > > > > I'm running a build now and will commit the fix soon (if the maven build > > doesn't overheat my laptop this time :-(, It took me almost one hour to > > recover from last crash). > > > > Thanks, > > Raymond > > > > ----- Original Message ----- > > From: "Simon Laws" <[EMAIL PROTECTED]> > > To: <tuscany-dev@ws.apache.org>; <[EMAIL PROTECTED]> > > Sent: Monday, September 17, 2007 3:01 PM > > Subject: Re: sample start up time? > > > > > > > On 9/17/07, ant elder <[EMAIL PROTECTED]> wrote: > > >> > > >> Wonderful. > > >> > > >> ...ant > > >> > > >> On 9/17/07, Raymond Feng <[EMAIL PROTECTED]> wrote: > > >> > > > >> > Yes, that's what I meant. > > >> > > > >> > Thanks, > > >> > Raymond > > >> > > > >> > ----- Original Message ----- > > >> > From: "ant elder" <[EMAIL PROTECTED]> > > >> > To: <tuscany-dev@ws.apache.org> > > >> > Sent: Monday, September 17, 2007 1:54 PM > > >> > Subject: Re: sample start up time? > > >> > > > >> > > > >> > > Is that also going to help the shutdown hang problem? > > >> > > > > >> > > ...ant > > >> > > > > >> > > On 9/17/07, Raymond Feng <[EMAIL PROTECTED]> wrote: > > >> > >> > > >> > >> Hi, > > >> > >> > > >> > >> I found the issue and will check in a fix soon. > > >> > >> > > >> > >> Thanks, > > >> > >> Raymond > > >> > >> > > >> > >> ----- Original Message ----- > > >> > >> From: "kelvin goodson" <[EMAIL PROTECTED]> > > >> > >> To: <tuscany-dev@ws.apache.org> > > >> > >> Sent: Monday, September 17, 2007 8:01 AM > > >> > >> Subject: Re: sample start up time? > > >> > >> > > >> > >> > > >> > >> > Hmm, I just re-ran this against RC1b and the calculator > sample > > >> hung > > >> > >> > after producing its results. All the output appeared pretty > > >> quickly, > > >> > >> > but then it failed to stop. > > >> > >> > > > >> > >> > C:\Release\RC1b\tuscany-sca-1.0-incubating\samples\calculator > >ant > > >> ru > > >> > >> > Buildfile: build.xml > > >> > >> > > > >> > >> > run: > > >> > >> > [java] 3 + 2=5.0 > > >> > >> > [java] 3 - 2=1.0 > > >> > >> > [java] 3 * 2=6.0 > > >> > >> > [java] 3 / 2=1.5 > > >> > >> > [java] Java Result: 1 > > >> > >> > > > >> > >> > BUILD SUCCESSFUL > > >> > >> > Total time: 8 minutes 4 seconds > > >> > >> > Terminate batch job (Y/N)? y > > >> > >> > > > >> > >> > On 17/09/2007, kelvin goodson <[EMAIL PROTECTED]> wrote: > > >> > >> >> I can run the calculator in 4 seconds, -- time until I see > the > > >> "Run" > > >> > >> >> word on the screen approx 2 secs and then a further 2 before > I > > >> > >> >> see > > >> > the > > >> > >> >> result output. > > >> > >> >> > > >> > >> >> Kelvin. > > >> > >> >> > > >> > >> >> On 17/09/2007, Simon Laws <[EMAIL PROTECTED]> wrote: > > >> > >> >> > I made a local distro from the branch and the samples seem > to > > >> take > > >> > >> ages > > >> > >> >> > to > > >> > >> >> > get going. A really comical amount of time! It takes 25 > seconds > > >> > >> before > > >> > >> >> > I see > > >> > >> >> > any output after running the calculator sample. Anyone else > > >> seeing > > >> > >> >> > that? > > >> > >> >> > > > >> > >> >> > Simon > > >> > >> >> > > > >> > >> >> > > >> > >> > > > >> > >> > > > >> --------------------------------------------------------------------- > > >> > >> > To unsubscribe, e-mail: [EMAIL PROTECTED] > > >> > >> > For additional commands, e-mail: > [EMAIL PROTECTED] > > >> > >> > > > >> > >> > > >> > >> > > >> > >> > --------------------------------------------------------------------- > > >> > >> To unsubscribe, e-mail: [EMAIL PROTECTED] > > >> > >> For additional commands, e-mail: [EMAIL PROTECTED] > > >> > >> > > >> > >> > > >> > > > > >> > > > >> > > > >> > > > That's made a big (+ve) difference. What was the problem Raymond? > > > > > > Simon > > > > > > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [EMAIL PROTECTED] > > For additional commands, e-mail: [EMAIL PROTECTED] > > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [EMAIL PROTECTED] > For additional commands, e-mail: [EMAIL PROTECTED] > >
