<?xml version="1.0" encoding="UTF-8"?><!-- generator="wordpress/2.1.1" -->
<rss version="2.0" 
	xmlns:content="http://purl.org/rss/1.0/modules/content/">
<channel>
	<title>Comments on: Dynamically defined Smoke Test Suites for Java</title>
	<link>http://grahamis.com/blog/2007/09/11/dynamically-defined-smoke-test-suites-for-java/</link>
	<description>It's too real to be true</description>
	<pubDate>Fri, 21 Nov 2008 19:52:09 +0000</pubDate>
	<generator>http://wordpress.org/?v=2.1.1</generator>

	<item>
		<title>By: Pete Moore</title>
		<link>http://grahamis.com/blog/2007/09/11/dynamically-defined-smoke-test-suites-for-java/#comment-401</link>
		<author>Pete Moore</author>
		<pubDate>Wed, 12 Sep 2007 07:11:07 +0000</pubDate>
		<guid>http://grahamis.com/blog/2007/09/11/dynamically-defined-smoke-test-suites-for-java/#comment-401</guid>
					<description>Just want to say that the "tops new version" of Clover is out (creatively called Clover 2) and it does generate per test coverage info. e.g. http://www.cenqua.com/clover/20/samples/index_test_summary.html :)

We definitely plan to do some kind of test optimization in the next release or two. In fact the Bamboo (Atlassian's CI offering) guys are pretty keen on making test optimization an out of the box option if you also have Clover installed which could be really nice.</description>
		<content:encoded><![CDATA[<p>Just want to say that the &#8220;tops new version&#8221; of Clover is out (creatively called Clover 2) and it does generate per test coverage info. e.g. <a href="http://www.cenqua.com/clover/20/samples/index_test_summary.html" rel="nofollow">http://www.cenqua.com/clover/20/samples/index_test_summary.html</a> <img src='http://grahamis.com/blog/wp-includes/images/smilies/icon_smile.gif' alt=':)' class='wp-smiley' /> </p>
<p>We definitely plan to do some kind of test optimization in the next release or two. In fact the Bamboo (Atlassian&#8217;s CI offering) guys are pretty keen on making test optimization an out of the box option if you also have Clover installed which could be really nice.</p>
]]></content:encoded>
				</item>
	<item>
		<title>By: Brendan Humphreys</title>
		<link>http://grahamis.com/blog/2007/09/11/dynamically-defined-smoke-test-suites-for-java/#comment-402</link>
		<author>Brendan Humphreys</author>
		<pubDate>Wed, 12 Sep 2007 08:30:09 +0000</pubDate>
		<guid>http://grahamis.com/blog/2007/09/11/dynamically-defined-smoke-test-suites-for-java/#comment-402</guid>
					<description>Hi Josh,

Good to meet you back at CITCON Sydney. Clover 2 will include similar functionality - the ability to determine a subset of tests to run based on what source files have changed. 

JTestMe certainly was an elegant use of Aspects - one that could potentially replace the obligatory "logging" example that Aspectinado's like to proffer ;-) We don't use Aspects simply because we don't need to - our instrumentation already collects the necessary data to determine which tests hit which statements. 

Of course, as we discussed at the time, the assumption being made here - tests that execute over changed code are the only ones that will test that code - doesn't always hold. Obviously running all the tests regularly is still prudent :-)

Clover 2 will also provide something similar to ProTest - allowing prioritising based on changed code, coverage and speed, so that tests that provide the most coverage the quickest are run first, the idea here being that your test suite should fail fast.

Cheers,
-Brendan</description>
		<content:encoded><![CDATA[<p>Hi Josh,</p>
<p>Good to meet you back at CITCON Sydney. Clover 2 will include similar functionality - the ability to determine a subset of tests to run based on what source files have changed. </p>
<p>JTestMe certainly was an elegant use of Aspects - one that could potentially replace the obligatory &#8220;logging&#8221; example that Aspectinado&#8217;s like to proffer <img src='http://grahamis.com/blog/wp-includes/images/smilies/icon_wink.gif' alt=';-)' class='wp-smiley' /> We don&#8217;t use Aspects simply because we don&#8217;t need to - our instrumentation already collects the necessary data to determine which tests hit which statements. </p>
<p>Of course, as we discussed at the time, the assumption being made here - tests that execute over changed code are the only ones that will test that code - doesn&#8217;t always hold. Obviously running all the tests regularly is still prudent <img src='http://grahamis.com/blog/wp-includes/images/smilies/icon_smile.gif' alt=':-)' class='wp-smiley' /> </p>
<p>Clover 2 will also provide something similar to ProTest - allowing prioritising based on changed code, coverage and speed, so that tests that provide the most coverage the quickest are run first, the idea here being that your test suite should fail fast.</p>
<p>Cheers,<br />
-Brendan</p>
]]></content:encoded>
				</item>
</channel>
</rss>
