You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tuscany.apache.org by Luciano Resende <lu...@gmail.com> on 2010/08/20 09:05:39 UTC
Binding Comet location
Should we move the binding-comet-* from [1] to [2], this makes it
easier to others to load and give feedback to the new code. Or even
better, are we ready to move it to trunk ?
[1] https://svn.apache.org/repos/asf/tuscany/sca-java-2.x/contrib/modules/
[2] https://svn.apache.org/repos/asf/tuscany/sca-java-2.x/trunk/contrib/modules/
--
Luciano Resende
http://people.apache.org/~lresende
http://twitter.com/lresende1975
http://lresende.blogspot.com/
Re: Binding Comet location
Posted by Florian MOGA <mo...@gmail.com>.
Hi Luciano,
I'm afraid it's been a little misunderstanding as I totally agree with
moving the code to trunk/contrib and I was just detailing why I did the copy
of poms from trunk to contrib. It's just that the same build problem was
affecting the other modules and samples in contrib/ so that solved them too.
I've committed new versions of them (contrib/pom.xml and
contrib/modules/pom.xml) which has an empty <modules> section so that they
won't report they can't find the modules. Let me know if you agree with the
present solution or feel free to let me know if you have other alternatives
for it. In the future, is it wanted to move all the work from contrib to
trunk/contrib?
Meanwhile, I moved the comet binding to trunk/contrib. Could you give me
some more details on the stability and readiness of a new extension. I'm not
sure what are the next steps I need to do in order to get the binding ready
for the release.
Thanks,
Florian
On Fri, Aug 27, 2010 at 8:17 PM, Luciano Resende <lu...@gmail.com>wrote:
> On Fri, Aug 27, 2010 at 4:59 AM, Florian MOGA <mo...@gmail.com> wrote:
>
> > Regarding the binding, I've just fixed the last issues and details so if
> > that's ok with everybody I can proceed moving it to trunk these days.
> Will
> > it enter the 2.0 release or are there any further steps needed?
> > Florian
>
> I guess my initial suggestion was much simpler, just move from
> contrib/modules [1] to trunk/contrib/modules [2], to simplify how
> others would checkout and load the extension to their IDE. As for
> including in the release, I guess you would be the best to access the
> stability and readiness of the new extension and if all tests are
> passing, etc
>
>
> [1] https://svn.apache.org/repos/asf/tuscany/sca-java-2.x/contrib/modules/
> [2]
> https://svn.apache.org/repos/asf/tuscany/sca-java-2.x/trunk/contrib/modules/
>
>
> --
> Luciano Resende
> http://people.apache.org/~lresende
> http://twitter.com/lresende1975
> http://lresende.blogspot.com/
>
Re: Binding Comet location
Posted by Luciano Resende <lu...@gmail.com>.
On Fri, Aug 27, 2010 at 4:59 AM, Florian MOGA <mo...@gmail.com> wrote:
> Regarding the binding, I've just fixed the last issues and details so if
> that's ok with everybody I can proceed moving it to trunk these days. Will
> it enter the 2.0 release or are there any further steps needed?
> Florian
I guess my initial suggestion was much simpler, just move from
contrib/modules [1] to trunk/contrib/modules [2], to simplify how
others would checkout and load the extension to their IDE. As for
including in the release, I guess you would be the best to access the
stability and readiness of the new extension and if all tests are
passing, etc
[1] https://svn.apache.org/repos/asf/tuscany/sca-java-2.x/contrib/modules/
[2] https://svn.apache.org/repos/asf/tuscany/sca-java-2.x/trunk/contrib/modules/
--
Luciano Resende
http://people.apache.org/~lresende
http://twitter.com/lresende1975
http://lresende.blogspot.com/
Re: Binding Comet location
Posted by Florian MOGA <mo...@gmail.com>.
Hi Luciano,
There was no pom in the contrib/ folder as well (I've committed it at that
time) so the only way to have it working without copying poms was either
setting the parent to ../../trunk/modules/pom.xml or not using a parent
which wouldn't bring some tuscany dependencies.
Before making the changes, I've checked the other modules found in
contrib/modules and found out that they also have the parent set to
../pom.xml so this workaround fixes the build problem on the other modules
too. I agree that duplicating poms from trunk is probably not the nicest
solution but otherwise we'll need to modify all the other poms in each
module.
Changing the poms from contrib/ and contrib/modules not to define any child
modules like in trunk seems like a clean solution for now. Any other ideas?
It is possible to somehow import the poms from trunk so that when global
changes are made, they won't have to be made in both trunk and contrib?
Regarding the binding, I've just fixed the last issues and details so if
that's ok with everybody I can proceed moving it to trunk these days. Will
it enter the 2.0 release or are there any further steps needed?
Florian
Re: Binding Comet location
Posted by Luciano Resende <lu...@gmail.com>.
On Tue, Aug 24, 2010 at 5:22 PM, Florian MOGA <mo...@gmail.com> wrote:
> Hi,
> Sorry for the late post, been out of town the last few days. Today I've
> committed some changes to the binding and the sample consisting in
> formatting the code with the Tuscany Eclipse formatter, using the Tuscany
> code template and doing a little bit of refactoring where needed. This being
> done, we're good to go from my side to get the modules (binding-comet and
> binding-comet-runtime) and the sample (comet-webapp) into trunk. If somebody
> is willing to check them out and make some code review, that would be even
> better.
> A few observations regarding the module would be the following:
>
> The unit test from binding-comet-runtime consists in starting a node which
> defines a contribution with a comet service. This will fail if the machine
> on which this is executed has already taken the 8080 port. Is this feasible?
> I don't know how the Hudson environment will behave in such a situation. If
> problems occur we could simply rename CometTestCase to CometTest and leave
> the unit test to be run manually when needed.
We usually setup the unit tests to run on port 8085, which will then
run with no problem on the hudson environment.
> I've also commited a copy of trunk/pom.xml and trunk/modules/pom.xml in the
> contrib and contrib/modules folders as a temporary solution as modules in
> the contrib folder define a parent pom with relative path "../pom.xml" which
> will lead to missing parent pom errors when no prior Tuscany trunk module
> builds are performed. This affected pretty much all the modules in contrib
> from what I've seen. Performed a build in a clean environment and didn't get
> any parent pom errors anymore.
I'm not sure I understood what's the issue you were facing here, if
the problem is with relative path, you could easily just fix the
relative path from ../pom.xml to maybe ../../pom.xml ?
> It is possible that some changes will be needed in pom.xml when moving to
> trunk but I'm not that sure. Most probably, the only updates that will need
> to be done are located in parent poms.
>
If you are having issues moving the project, please let me know and I
can give it a try.
> Feel free to give feedback regarding this new binding.
I have plans to take a deeper look at it, but i'm not finding much
free time on my nightly shift :)
--
Luciano Resende
http://people.apache.org/~lresende
http://twitter.com/lresende1975
http://lresende.blogspot.com/
Re: Binding Comet location
Posted by Florian MOGA <mo...@gmail.com>.
Hi,
Sorry for the late post, been out of town the last few days. Today I've
committed some changes to the binding and the sample consisting in
formatting the code with the Tuscany Eclipse formatter, using the Tuscany
code template and doing a little bit of refactoring where needed. This being
done, we're good to go from my side to get the modules (binding-comet and
binding-comet-runtime) and the sample (comet-webapp) into trunk. If somebody
is willing to check them out and make some code review, that would be even
better.
A few observations regarding the module would be the following:
- The unit test from binding-comet-runtime consists in starting a node
which defines a contribution with a comet service. This will fail if the
machine on which this is executed has already taken the 8080 port. Is
this feasible? I don't know how the Hudson environment will behave in such a
situation. If problems occur we could simply rename CometTestCase to
CometTest and leave the unit test to be run manually when needed.
- I've also commited a copy of trunk/pom.xml and trunk/modules/pom.xml in
the contrib and contrib/modules folders as a temporary solution as modules
in the contrib folder define a parent pom with relative path "../pom.xml"
which will lead to missing parent pom errors when no prior Tuscany trunk
module builds are performed. This affected pretty much all the modules in
contrib from what I've seen. Performed a build in a clean environment and
didn't get any parent pom errors anymore.
- It is possible that some changes will be needed in pom.xml when moving
to trunk but I'm not that sure. Most probably, the only updates that will
need to be done are located in parent poms.
Feel free to give feedback regarding this new binding.
Thanks,
Florian
On Fri, Aug 20, 2010 at 10:05 AM, Luciano Resende <lu...@gmail.com>wrote:
> Should we move the binding-comet-* from [1] to [2], this makes it
> easier to others to load and give feedback to the new code. Or even
> better, are we ready to move it to trunk ?
>
> [1] https://svn.apache.org/repos/asf/tuscany/sca-java-2.x/contrib/modules/
> [2]
> https://svn.apache.org/repos/asf/tuscany/sca-java-2.x/trunk/contrib/modules/
>
> --
> Luciano Resende
> http://people.apache.org/~lresende
> http://twitter.com/lresende1975
> http://lresende.blogspot.com/
>