Maybe I need to rephrase this a bit. Your line of reasoning is absolutely
correct from a process point of view (like a CI build). But rigorously
applying this principle across the board denies developers the comfort of
selectively running subsets of tests. Developer productivity should count
for something as well.


Kristof Vanbecelaere wrote:
> 
> I disagree. Have you ever written a selenium test? This is trial and
> error. I have not touched any "real" code, only test code. So I know my
> unit tests succeed. All I want to do is run the integration-test phase
> without unit tests.
> 
> 
> Stephen Connolly-2 wrote:
>> 
>> 2008/11/27 Kristof Vanbecelaere <[EMAIL PROTECTED]>
>> 
>> Actually, that is the aim.
>> 
>> You run all the unit tests to make sure that the code is good enough to
>> try
>> and run the integration tests.
>> 
>> If your unit tests fail, your code is broken and you know it, so fix your
>> code.
>> 
>> If your unit tests pass, now lets see if it integrates correctly, hence
>> run
>> the integration tests.
>> 
>> If the integration tests pass, we can publish the project (i.e. install
>> or
>> deploy to maven repo)
>> 
>> This is what the lifecycle is all about... a well defined sequence of
>> phases, all the previous phases must complete successfully before the
>> next
>> phase starts.
>> 
>> 
> 
> 

-- 
View this message in context: 
http://www.nabble.com/maven-surefire-plugin-configuration-for-unit-integration-tests-tp20724622p20730746.html
Sent from the Maven - Users mailing list archive at Nabble.com.


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to