Hans Dockter schrieb:
Adam and I were talking yesterday about how to improve the skipping. We have discussed the idea that it would be nice to skip complete subsections of the DAG. One idea was to somehow find implicit rules on which to decide what makes up such a subsection.

For example:

clean<-resources<-compile<-testResources<-test

install dependsOn compile, test

In such a situation we could figure out, that testResources is only needed by the test task, and in the case the test task is skipped we could skip testResources as well.

so that is what you can do already...

But to just define 'install dependsOn test', with the knowledge that compile is also executed when test is executed, would define any subsections to skip. But in this case I see dependsOn compile, test and dependsOn test as equivalent. I think this should not lead to different behavior.

if you divide the tasks in those that are helpers and those that are intended to be called, then you could say, that for example test and compile are callable tasks, while testResource is not. If you now have 'install dependsOn test' and test is skipped, then following the DAG, we can skip any not callable task.so the next task we actually execute would be compile and all it depends on. Of course the question if that can be realized easily very much depends on how the DAG is build, if it is build and how a task is executed

An alternative would be to introduce the concept of task groups. Each task can belong to 0..n task groups and we allow to skip specific tasks or task groups.

not sure, but that sounds a bit complicated to me

bye blackdrag

--
Jochen "blackdrag" Theodorou
The Groovy Project Tech Lead (http://groovy.codehaus.org)
http://blackdragsview.blogspot.com/


---------------------------------------------------------------------
To unsubscribe from this list, please visit:

   http://xircles.codehaus.org/manage_email


Reply via email to