Thanks for the writeup, these kinds of explanations are extremely helpful. On Feb 6, 2011, at 10:46 PM, Kalyan Akella wrote:
> A few days back, I was working with the Liquibase Maven > Plugin<http://www.liquibase.org/manual/maven>for doing DB upgrades in > my project. This plugin has 2 specific goals, among > others, which I was using, > update<http://www.liquibase.org/manual/maven_update>& > rollback <http://www.liquibase.org/manual/maven_rollback>. 'update' goal > executes all the Liquibase change-sets & 'rollback' undoes the applied > change-sets. > > So here's my scenario. I need update & rollback to happen on my local test > DB during the integration-test phase. Moreover, I want them to execute only > when I introduce a Liquibase changeset. The idea was to verify whether my > change-set is a valid one by doing an upgrade->rollback->upgrade cycle on > the test DB. Because, there were a large number of updates, I only wanted > this to happen if I introduce a change-set in the first place. > > Initially, I wanted to control this behavior at the plugin level (in the > POM) where I choose to execute the Liquibase plugins or not based on whether > I pass in a system property (ala, -DskipTests) or through some other > mechanism. Unfortunately, both the Liquibase plugins do not support the skip > flag (as others do, like the surefire plugin). I had to later achieve this > result by moving these two plugins inside a Maven profile and skip the > profile altogether using the activation tag. > > So, here's my point. Whatever reasoning was behind the introduction of > 'skip' flag functionality in some plugins like surefire, is the same very > reason we need this support to be built across all plugins in a generic way. > We can achieve this at the Maven level either through the CLI or through the > POM as my implementation proves. It is of course, up to the community to > decide when this should be rolled out as a feature. So you implemented this by skipping whole phases, and this is one approach. Just thinking off the top of my but what if you could define a set of plugin executions you declared to skip? Or a combination of both where you can skip whole phases if you like, but would it be useful to skip a selected set of executions. Then we could support things like just skipping Surefire, like we do now, except that each plugin author wouldn't have to implement this logic. I agree with you that providing some plumbing for this in the core would make more sense. My concern still remains the portability of builds. Thanks again for the explanation. > > Sincere Regards, > Kalyan C. Akella > > On Sun, Feb 6, 2011 at 8:32 PM, Jason van Zyl <ja...@sonatype.com> wrote: > >> Out of curiosity what is the first thing you used this feature for? >> >> On Feb 6, 2011, at 8:06 AM, Kalyan Akella wrote: >> >>> Hi, >>> >>> Thank you for the comments. >>> >>> First of all, these changes DO NOT affect the existing integration test >>> suite. I verified this by running the ENTIRE suite with & without my >>> patches. The *results are the same*. >>> >>> Secondly, I am in no urgent need to get these changes to Maven 3.0.x. Do >> it >>> as you see fit. My request for applying the patches was only from a pure >>> code review perspective. >>> >>> Lastly, I shall update the JIRA ticket with the specific use cases where >>> this feature helps. If my time permits, I'll update the usage & >>> documentation. >>> >>> Sincere Regards, >>> Kalyan C. Akella >>> >>> On Sun, Feb 6, 2011 at 3:28 AM, Jason van Zyl <ja...@sonatype.com> >> wrote: >>> >>>> >>>> On Feb 5, 2011, at 7:43 AM, Kalyan Akella wrote: >>>> >>>>> Hi, >>>>> >>>>> I attached the revised patches to the issues that implement the >>>> following: >>>>> >>>>> 1. ability to skip multiple plugins & their executions at CLI, >>>>> 2. ability to skip plugins & their executions in POM using the <skip> >>>>> element >>>>> >>>>> I'm still working on the documentation. Please apply the patches and >> let >>>> me >>>>> know your comments. >>>>> >>>> >>>> These patches are not going to be applied to trunk right away. >>>> >>>> I would suggest using the Github Maven tree, make a branch and apply >> your >>>> changes there if you would like people to try them. This is something >> that >>>> would greatly change the behavior of Maven so it's not going into 3.0.x. >> It >>>> may go into Maven 3.1 but that's a discussion. Changes like this aren't >>>> going in without a discussion. I also asked that explanations be >> provided >>>> about why this patch is useful and the use cases surrounding its >>>> functionality. >>>> >>>> If you haven't already, we can show you how the integration test suite >>>> works, as any changes you make must not affect the operation of the >> existing >>>> integration tests. >>>> >>>>> Sincere Regards, >>>>> Kalyan C. Akella >>>>> >>>>> On Wed, Feb 2, 2011 at 9:27 AM, Kalyan Akella <kalyan.ake...@gmail.com >>>>> wrote: >>>>> >>>>>> Hi, >>>>>> >>>>>> Thank you for the patch review comments. I hope it is integrated to >> the >>>>>> trunk. >>>>>> >>>>>> My next steps: >>>>>> >>>>>> 1. Implement support for skipping multiple plugins/executions at the >>>> CLI. >>>>>> 2. Implement ability to skip plugin execution through the POM as >>>> required >>>>>> by http://jira.codehaus.org/browse/MNG-3102 >>>>>> 3. Add documentation (with the proper reasoning) for both these >>>> approaches. >>>>>> >>>>>> >>>>>> Sincere Regards, >>>>>> Kalyan C. Akella >>>>>> >>>>>> On Wed, Feb 2, 2011 at 3:17 AM, Kristian Rosenvold < >>>>>> kristian.rosenv...@gmail.com> wrote: >>>>>> >>>>>>> Our main topic of discussion on irc was initially if this *should* be >> a >>>>>>> 3.0.X or 3.1 issue due to being somewhat of a borderline between >>>>>>> improvement/new feature. It seems to me like the way this discussion >> is >>>>>>> heading, it should probably be tagged 3.1. >>>>>>> >>>>>>> I agree very much about getting some well defined use cases and >>>>>>> John's comments on the issue are also very good. >>>>>>> >>>>>>> I will leave the issue at "3.1" for now, and request well defined >>>>>>> use-cases on the issue. >>>>>>> >>>>>>> Kristian (I have all the skip options I need already) >>>>>>> >>>>>>> >>>>>>> ti., 01.02.2011 kl. 15.44 -0500, skrev Jason van Zyl: >>>>>>>> So the technical work was done implementing this but no where in the >>>>>>> issue does it describe why this feature is useful. >>>>>>>> >>>>>>>> This fundamentally changes the behaviour of the lifecycle and the >>>>>>> example listed in the issue is skipping tests. Some explanation of >> why >>>> this >>>>>>> is good would be nice, there's nothing about the reasoning of the >>>> feature in >>>>>>> the issue. Possibly something like: >>>>>>>> >>>>>>>> - As a user I would like to skip the test phase locally, and only >> from >>>>>>> the command line so it's easier while I experiment putting new test >>>> features >>>>>>> in place. But I would like to guarantee that these can only be done >>>> from the >>>>>>> command line and never in the presence of the release plugin so that >>>>>>> skipping phases is not done inadvertently during a release. >>>>>>>> >>>>>>>> The second case in a issue doesn't strike me as useful without an >>>>>>> explanation, to run to the install phase but not package the JAR? >>>>>>>> >>>>>>>> I think a fundamental alteration of behaviour requires an >> explanation >>>> of >>>>>>> reasoning. I believe that is reasonable. It initially strikes me as a >>>>>>> potentially bad idea. Maybe you talked about the reasoning in IRC, >>>> which I >>>>>>> think is perfectly fine, just toss the discussion or salient points >> of >>>> the >>>>>>> reasoning into the issue. >>>>>>>> >>>>>>>> On Feb 1, 2011, at 3:00 PM, Kristian Rosenvold wrote: >>>>>>>> >>>>>>>>> I have looked at the patch diffs and it looks really great, one of >>>> the >>>>>>>>> better patches I've seen in the issue trackers ! >>>>>>>>> >>>>>>>>> We had a quick chat about this patch on irc and we figured we could >>>>>>> take >>>>>>>>> this patch for 3.0.x. >>>>>>>>> >>>>>>>>> One thing we would like though, is to be able to specify multiple >>>>>>>>> exclusions, not just one plugin/execution. Any chance you could fix >>>>>>>>> that ? >>>>>>>>> >>>>>>>>> As for the documentation, that would be something >>>>>>>>> like https://svn.apache.org/repos/asf/maven/site/trunk >>>>>>>>> >>>>>>>>> and the document >>>>>>>>> >>>>>>>>> src/site/apt/guides/mini/guide-configuring-maven.apt >>>>>>>>> >>>>>>>>> Sorry for not replying immediately; been busy preparing >>>>>>>>> my trip to the alps tomorrow ;) >>>>>>>>> >>>>>>>>> Kristian >>>>>>>>> >>>>>>>>> >>>>>>>>> ti., 01.02.2011 kl. 09.16 +0530, skrev Kalyan Akella: >>>>>>>>>> Hi, >>>>>>>>>> >>>>>>>>>> Any updates on the patch. Did it apply cleanly ? Any comments, >>>>>>> changes ?? >>>>>>>>>> >>>>>>>>>> Sincere Regards, >>>>>>>>>> Kalyan C. Akella >>>>>>>>>> >>>>>>>>>> On Sat, Jan 29, 2011 at 8:01 AM, Kalyan Akella < >>>>>>> kalyan.ake...@gmail.com>wrote: >>>>>>>>>> >>>>>>>>>>> Hi, >>>>>>>>>>> >>>>>>>>>>> Thank you for the comments. I updated the issue, >>>>>>>>>>> http://jira.codehaus.org/browse/MNG-3321 with the modified >>>> patches. >>>>>>>>>>> Regarding your point about documentation, will you please point >> me >>>>>>> to the >>>>>>>>>>> exact location where I should include the usage instructions for >>>>>>> this >>>>>>>>>>> feature and other info (perhaps a wiki ?). >>>>>>>>>>> >>>>>>>>>>> >>>>>>>>>>> Sincere Regards, >>>>>>>>>>> Kalyan C. Akella >>>>>>>>>>> >>>>>>>>>>> On Fri, Jan 28, 2011 at 12:55 PM, Kristian Rosenvold < >>>>>>>>>>> kristian.rosenv...@gmail.com> wrote: >>>>>>>>>>> >>>>>>>>>>>> Thanks for the patch, which looks good. I added some comments to >>>>>>> the >>>>>>>>>>>> issue, http://jira.codehaus.org/browse/MNG-3321 >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> Kristian >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>> --------------------------------------------------------------------- >>>>>>>>>>>> To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org >>>>>>>>>>>> For additional commands, e-mail: dev-h...@maven.apache.org >>>>>>>>>>>> >>>>>>>>>>>> >>>>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> >> --------------------------------------------------------------------- >>>>>>>>> To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org >>>>>>>>> For additional commands, e-mail: dev-h...@maven.apache.org >>>>>>>>> >>>>>>>> >>>>>>>> Thanks, >>>>>>>> >>>>>>>> Jason >>>>>>>> >>>>>>>> ---------------------------------------------------------- >>>>>>>> Jason van Zyl >>>>>>>> Founder, Apache Maven >>>>>>>> http://twitter.com/jvanzyl >>>>>>>> --------------------------------------------------------- >>>>>>>> >>>>>>>> First, the taking in of scattered particulars under one Idea, >>>>>>>> so that everyone understands what is being talked about ... Second, >>>>>>>> the separation of the Idea into parts, by dividing it at the joints, >>>>>>>> as nature directs, not breaking any limb in half as a bad carver >>>> might. >>>>>>>> >>>>>>>> -- Plato, Phaedrus (Notes on the Synthesis of Form by C. Alexander) >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>> >>>>>>> >>>>>>> >>>>>>> --------------------------------------------------------------------- >>>>>>> To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org >>>>>>> For additional commands, e-mail: dev-h...@maven.apache.org >>>>>>> >>>>>>> >>>>>> >>>> >>>> Thanks, >>>> >>>> Jason >>>> >>>> ---------------------------------------------------------- >>>> Jason van Zyl >>>> Founder, Apache Maven >>>> http://twitter.com/jvanzyl >>>> --------------------------------------------------------- >>>> >>>> We all have problems. How we deal with them is a measure of our worth. >>>> >>>> -- Unknown >>>> >>>> >>>> >>>> >> >> Thanks, >> >> Jason >> >> ---------------------------------------------------------- >> Jason van Zyl >> Founder, Apache Maven >> http://twitter.com/jvanzyl >> --------------------------------------------------------- >> >> We all have problems. How we deal with them is a measure of our worth. >> >> -- Unknown >> >> >> >> Thanks, Jason ---------------------------------------------------------- Jason van Zyl Founder, Apache Maven http://twitter.com/jvanzyl --------------------------------------------------------- In short, man creates for himself a new religion of a rational and technical order to justify his work and to be justified in it. -- Jacques Ellul, The Technological Society
