If nothing else comes out of this discussion thread, we will at least hear build-process opinions from above-average developers. Thank you in advance to all that reply.
In response to Filip : On 10/17/07, Filip Hanik - Dev Lists <[EMAIL PROTECTED]> wrote: > lemme give you my feedback and some history I sincerely appreciate feedback and history. Additionally, thank you for caring. :) > we just refactored everything from being "component/module" based into a > single source tree. > Everyone at the time agreed that it would make life easier, for me > personally, it was a huge improvement. I was unaware of the prior unification process, and I thought I noticed a sense of frustration :) Perhaps modularizing is not one of the features that we need then. Seeing as I am still new here, I am unaware of any current component delegation (i.e. John Smith works on catalina-ha only, Jane Doe works on jasper only). I have had to assume that all active developers just simply work on every single component in the project. In other projects that I have worked on, a clear separation of responsibility is vital to success and a more prosperous growth. > > - Model-Based builds - Automatic packaging for the individual modules. > not sure what this is, even though we have a single source tree, we do > generate a list of jars. This would not have to be an all-or-nothing separation. But, if a component is just an API, the packaging model is a JAR; the current webapp models and future webapp modules could even package as WARs. If everyone did in fact agree to modularize the separate components as mentioned above, each component would be responsible for its own packaging process instead of managing all packaging and distribution within the single build/dist files. In many cases I have found this to be a great feature. Of course, in others, it can be a nightmare. If Tomcat was modularized and failed as such, I will respect the single source tree. > > - Dependency Management - Whether it is Apache or another third-party, > > dependencies can all easily be plugged in. > > > we do that today, crude but working, ANT just adopted Ivy, a dependency > manager for ANT. I was unaware of the ASF adoption of Ivy 2.0. If it is, as you say, 'crude', why was it adopted? I thought Ivy was just drowning in the wake of Maven 2, but it seems to have more features than when I first evaluated it. Thank you for the heads up; I will be keeping an eye on it. > > - Distribution Management - Packaging and Deployment - Although Tomcat > > has a structured distribution model with Ant, Maven could make this > > easier with its assembly plugin. > > > We currently have a "distribute to Maven repo" in place. > The most current version is in the sandbox, that would allow us to > publish to the central ASF repo with signed JAR's. > This allows(will allow) other projects that do use Maven, to integrate > tomcat into their system. > You can glance over it here > http://svn.apache.org/viewvc/tomcat/sandbox/gdev6x/res/maven/ While reviewing the project structure I did in fact notice that you added the poms there. I assumed that this was just an orphaned process because the 6.0.14 files did not publish to the main repository at : http://repo1.maven.org/maven2/org/apache/tomcat/. I am clearly unaware of the distribution schedule, and I apologize for my confusion. > So, speaking for myself, I have yet not seen a benefit of Maven over our > current ANT build. And I wouldn't be up for eating dog food. > water and cracker, although simple, have sustained us very long. I am typically a big supporter of the KISS Principal (http://en.wikipedia.org/wiki/KISS_principle), too. Of course, I also believe that this principle, when applied to IT, often marries itself to a resistance-to-change. I agree that Maven does not in fact suit every build, and, if done wrong, it can be detrimental to the health of a programmer. If 20 active developers know and prefer Ant, and 1 knows and prefers Maven....well, I am going to promote Ant all the way. The purpose here was solely to generate some feedback. Both Ant and Maven will surely be around for awhile, but it is only right to give each a fair chance to excel at whatever each is good at. > I'd vote against the proposal, maybe cause I'm just getting to old to > spend hours with Maven, but you should collect feedback from the others > as well, and maybe there is a majority one way or the other. You are only old if you admit to yourself that you are old. Again, whether the votes come in as a +1 or -1, I am honored to have an opportunity to work with you. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]