harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Geir Magnusson Jr." <g...@pobox.com>
Subject Re: [drlvm][testing] Regression Testing Framework announcement
Date Wed, 27 Dec 2006 10:47:52 GMT
Thanks for the announcement, but I'm a little confused - we've had  
this in DRLVM for a little while now?  Or is it something new?

On Dec 27, 2006, at 4:49 AM, Alexander Kleymenov wrote:

> Hello,
> With this letter I'd like to announce another step made to robust and
> trustworthy DRLVM – its regression testing infrastructure. Recently it
> was committed into DRLVM workspace and now it is ready to be used. To
> get started with framework there is a README file placed under
> src/test/regression directory. Along with the README file you can see
> examples of the regression tests.
> Here I am listing the main features of the framework:
>    - The regression test format is JUnit. (Recommended, but not  
> limited by).
>    - Support for custom test launchers allowing to configure the way
> to execute the test (JVM arguments specifying, environment variables
> presetting, some preconditions checking and so on)
>    - Easy new test integration. Just write it, and place it in to the
> right place (actually corresponding to JIRA report describing the
> problem)
>    - Support for test consisting of Java/C/C++/Java Assembler  
> source code.
> Now, to successfully proceed with DRLVM regression testing:
> - From formal point of view: we should take an agreement for each
> closing JIRA to provide (whenever it's possible) regression test
> complying with format of the framework. For more information on "Good
> Issue Resolution Guideline" see [1].
> - From technical point of view: we should include automated
> regression testing into CC based code integrity testing.
> I'd like to thank and encourage the people who already use the
> framework. I see 2 new regression tests implemented: for HARMONY-2259
> [2] (by George Timoshenko) and for HARMONY-2261 [3] (by Egor Pasko).
> Great stuff! Let's go in such a way! Let's provide regression tests
> without describing of launch preconditions (as for example in
> HARMONY-2872 [4]) but saying for themselves! And, I believe, soon
> DRLVM will become even more reliable and stable!
> Thank you,
> Alexander
> [1] http://harmony.apache.org/issue_resolution_guideline.html
> [2] http://issues.apache.org/jira/browse/HARMONY-2259
> [3] http://issues.apache.org/jira/browse/HARMONY-2261
> [4] http://issues.apache.org/jira/browse/HARMONY-2872

View raw message