I have started to change the SCA code.

Initially, the objects instantiated dynamically were not being instantiated,
because info about which class to instantiate is extract from some text
files contained in SCA modules' jars. Unfortunately, when Android converter
converts the SCA modules' jars to .dex files, it ignores the non-class
files, and they are not included. It happens because Android has a very
specific way to store the app resources, which is completely different from
Java.

Though, as Luciano suggested, I started to modify these dynamic
instantiation for static. It's working for now : ). All these modifications
were made only on host-embedded module code, which I have copied into [1]
and renamed to host-android.

Also, one of those static instantiated classes was trying to access the
javax.naming.InitialContext in its constructor, but it's not supported by
Android and I commented the code. As this class belonged to SCA core module,
I also copied this module into the mobile-android sandbox and renamed the
module to core-android.

So, all the modifications made so far are in the [2].

Now, as I was expecting, I got some ClassNotFoundExceptions on SCA code when
it tries to use the Java XML API. It happens because Android implements only
a small part of this API. So, next step is try to adapt the SCA XML access
into Android way to access XML. Does anybody know some other XML API that is
javax.xml independent I could use instead of Android one?

Suggestions? : )

Adriano Crestani

[1]
https://svn.apache.org/repos/asf/incubator/tuscany/sandbox/mobile-android/host-android/
[2]
https://svn.apache.org/repos/asf/incubator/tuscany/sandbox/mobile-android/<https://svn.apache.org/repos/asf/incubator/tuscany/sandbox/mobile-android/host-android/>



On Mon, Mar 17, 2008 at 1:41 PM, Luciano Resende <[EMAIL PROTECTED]>
wrote:

> I'd probably categorize the experiment we are doing now, as a learning
> experience to familiarize with the Mobile environment available in
> Google Android. For this, we have choosen to try running one of the
> simplest samples we had (calculator) in the Android JVM
> environment.... In the long run, the idea is to have a more
> light-weight "sca mobile core/runtime" available, similar to what we
> did for Web 2.0 (implementation-widget) and described by Mike below.
>
> But it's indeed good to discuss these items with the community, would
> be also good to have the community contributing requirements/scenarios
> for a Tuscany mobile story. And we are all open and welcoming any
> help.
>
>
> On Mon, Mar 17, 2008 at 8:22 AM, Mike Edwards
> <[EMAIL PROTECTED]> wrote:
> > Brandon Werner wrote:
> >  > Can you explain the benefit of moving SCA/SDO over to the Android
> platform?
> >  > I can understand wanting to have the ability to consume simple XML
> services
> >  > and perhaps consume disconnected datagraphs, but I don't understand
> why the
> >  > world wouldn't want a more light-weight way of doing that than
> ripping out
> >  > something that was meant for the JEE / JSE.
> >  >
> >  > <snip>
> >  Brandon,
> >
> >  So you think that a device using the Andriod platform is likely to want
> >  to use services and a service-oriented approach to building
> >  applications.  But you think that SCA is too "heavyweight"?  Is that
> >  simply a question of the size of the code involved (I note that the
> core
> >  of Tuscany is quite small) or is the SCA approach too "heavyweight" in
> >  some other sense?
> >
> >  I wonder if you've taken a look a the implementation.widget exemplified
> >  in the Tutorial demo code?  This applies SCA principles to AJAX widgets
> >  running in the browser - simplifying the connectivity a great deal (in
> >  my opinion) and yet adding little overhead.
> >
> >  In the longer run, we may want to look to do something similar for the
> >  Android platforms, perhaps with Java rather than the more limited
> >  JavaScript components.  Starting with the current SCA package seems
> >  reasonable to me - once we have experience with how that works, we will
> >  have a better idea what needs to be changed and tailored for the
> Android
> >  platform.
> >
> >
> >  Yours,  Mike.
> >
> >
> >
> >  ---------------------------------------------------------------------
> >  To unsubscribe, e-mail: [EMAIL PROTECTED]
> >  For additional commands, e-mail: [EMAIL PROTECTED]
> >
> >
>
>
>
> --
> Luciano Resende
> Apache Tuscany Committer
> http://people.apache.org/~lresende <http://people.apache.org/%7Elresende>
> http://lresende.blogspot.com/
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>

Reply via email to