On Tue, Feb 3, 2009 at 11:50 AM, Simon Laws <[email protected]>wrote:
> > > On Tue, Feb 3, 2009 at 11:37 AM, ant elder <[email protected]> wrote: > >> >> >> On Tue, Feb 3, 2009 at 7:31 AM, Luciano Resende <[email protected]>wrote: >> >>> We have made good progress, and we should be really near getting a >>> good Java SCA 2.0-M1 RC ready for voting. >>> >>> In the mean time, I have created some stable distributions [1] so >>> people can start looking and making sure we are all comfortable with >>> the direction we are taking with the distributions, it's structure, >>> etc. >>> >>> I'll start reviewing the open JIRAS to see what is a must for this >>> release, and start moving things to the 2.0 queue. >>> >>> [1] >>> http://people.apache.org/~lresende/tuscany/20-M1-Stable/<http://people.apache.org/%7Elresende/tuscany/20-M1-Stable/> >>> >>> -- >>> Luciano Resende >>> Apache Tuscany, Apache PhotArk >>> http://people.apache.org/~lresende<http://people.apache.org/%7Elresende> >>> http://lresende.blogspot.com/ >>> >> >> >> One thing i noticed in the src distro is that lots of things seem to be >> included three times in the archive, i think we've had that in the past >> caused by something like overlapping includes in the src assembly xml. I've >> tried building the binary all distro today and it goes into a loop saying: >> >> at >> org.codehaus.plexus.archiver.AbstractArchiver$1.hasNext(AbstractArchiver.java:406) >> >> and then just ends with no message. >> >> Looking at the 2.0-M1 JIRAs there are two I think we should try really >> hard to fix for M1 - TUSCANY-2800 and TUSCANY-2805. >> >> ...ant >> >> >> >> >> > I'm up for a bit of sample build fixing as, with the latest build changes, > the binding-ws-calculator ant build has stopped working. Not only should we > fix the sample builds but instigate a module (or other device) to test the > samples from the distribution structure in as close a way as we can get to > how the user will run them. > > My initial thought, a week back, was to include testing in each sample pom > that relies on the "dir" version of the distro that we create in the main > maven build. The sample pom can then go run the ant build and run the > sample. I like this approach as it's sample specific. Someone (Ant?) said > that they wanted this to be in a separate test which I'm not opposed too if > we think sample testing makes the sample poms too complicated. > > I'm going to make a new module (itest-samples?) to try separating this > testing from the samples themselves.Basically going through the loop. > > for each sample > run ant compile > run the sample > > Simon > That itest sounds good to me. Binidng-ws-calculator is working for me just now. I've also been seeing intermittent fails in calculator-equinox, has anyone else see that? I've also just raised TUSCANY-2806 about the order things are getting built in which i think is one of the reasons we're seeing so many build issues. ...ant
