Setting up Aggregate 1.0 GAE in Eclipse

The time has come for me to port my work in building a custom External
Connection from Aggregate 0.9.3 to 1.0. To date, I'd been developing in
Eclipse with a Google App Engine local dev server. I got my dev server going
with the instructions available here:
http://code.google.com/p/opendatakit/wiki/AggregateCodeCheckout

I tried the same instructions to set up 1.0, but without success. Questions:

  • Are the instructions in the link above still valid for setting up a
    Aggregate 1.0 GAE local instance?
  • There are three branch options for the Mercurial checkout:
    uiexperiment, v1_0, and default. Which do I need?
··· -- Kevin Harvey Web Developer

The code checkout instructions are modified for 1.0. We haven't updated
that page since we haven't officially released Aggregate 1.0.

··· ------------------------------------------------------------------------ w.r.t. setting up your Aggregate 1.0 build environment:

See CONFIGURE.TXT:
http://code.google.com/p/opendatakit/source/browse/CONFIGURE.txt?repo=aggregate&name=v1_0

Eclipse development requires GWT 2.3 and GAE 1.5.1 or 1.5.2 (depending upon
the code branch and environment).

Maven 3 is used for command line builds and running unit and integration
tests (a small set right now). If anyone has cycles to complete the set-up
for emma code coverage tooling and reports across all the maven
sub-projects, that would be absolutely wonderful to have. We have a Jenkins
(nae Hudson) build server and it would be great to have html build reports
with code coverage stats.

Why is the configuration so complicated?

Limitations in the GWT plug-in and where it requires certain resources to
be, and its lack of support for facets, limits the Eclipse development
environment to a non-Maven-managed non-facet-based project
(eclipse-aggregate-gae). Because of this, we have only set up a GAE-based
development environment in Eclipse. Setting up a MySQL-based development
environment should be doable by replicating the directory structures of the
eclipse-aggregate-gae project (e.g., making duplicates of all the
war/WEB-INF/lib jars).


w.r.t. branches:

  • default -- Aggregate 0.9.x tip
  • v1_0 -- Aggregate 1.0 release tree
  • uiExperiment -- Aggregate 1.0 development tree -- this can be unstable
    with sparsely tested new features that may be undergoing a high rate of
    change.

For integrating independent changes, you should work against the v1_0 tip;
we are trying to keep that as stable as possible.
If you want to see what is coming down the pike, you can then try out the
uiExperiment tip.

Note that the database schema has changed between v1_0 and uiExperiment.
You will need to use different appspot instances if you try to build and
deploy both, and whenever 1.0 comes out, it will require a clean datastore,
as it will have the uiExperiment schema change.

Mitch

On Thu, Aug 11, 2011 at 10:18 AM, Kevin Harvey kcharvey@gmail.com wrote:

The time has come for me to port my work in building a custom External
Connection from Aggregate 0.9.3 to 1.0. To date, I'd been developing in
Eclipse with a Google App Engine local dev server. I got my dev server going
with the instructions available here:
http://code.google.com/p/opendatakit/wiki/AggregateCodeCheckout

I tried the same instructions to set up 1.0, but without success.
Questions:

  • Are the instructions in the link above still valid for setting up a
    Aggregate 1.0 GAE local instance?
  • There are three branch options for the Mercurial checkout:
    uiexperiment, v1_0, and default. Which do I need?

--
Kevin Harvey
Web Developer

--
Mitch Sundt
Software Engineer

University of Washington
mitchellsundt@gmail.com