Skip to main content

Please review bugfixes for ISSUES #30 and #31

No replies

Hi Allen, team,

Could you please review a couple of (related) bugfixes for the following issues:

Issue #30: REGRESSION: backward compatibility with MIDP 1.0 RI broken

REGRESSION: SATSA 1.0b TCK doesn't run tests in MIDP mode with ME Framework 1.2

In both cases, the root of the problem is the relationships between
J2meBaseTestSuite and J2meTestBundlingService. More specifically, the
test suite delegates getAgentClass() and getAgentJar() to the service,
which is wrong. These methods are being invoked by the interview code,
and at the moment when the interview needs this info, the service
might not be even created! This leads to NPEs in the Interview, or,
worse, to outdated values in the exported test environment.

Basically, lifecycles of Interview/TestSuite
and the Service are different, and the Interview should not depend on
the Services.

Also, there is no point to define default MIDP agent class names and
jars in the Bundling Service, they should be defined where all other
class and jar names defined, in the Interview.

So, the getAgentClass and getAgentJar methods in the TestSuite are
preserved, but they do not delegate to the Bundling, they just provide
extensions points for those test suites that would like to quickly
adjust midp agent class and jar. The default values for names and jars
have been moved to the Interview.

Also, Harness.getEnv() is used instead of
Harness.getParameters.getEnv(), since we need some values exported in

Attached is the webrev with bugfixes.

I'd appreciate your quick response, so that we could start regression
testing of this code as soon as possible.

To unsubscribe, e-mail:
For additional commands, e-mail: