On Tue, 3 Apr 2018 17:33:09 +0200 Michael Vorburger <[email protected]> wrote: > On Tue, Apr 3, 2018 at 5:18 PM, Stephen Kitt <[email protected]> wrote: > > On Sun, 25 Mar 2018 11:33:27 -0700 > > Casey Cain <[email protected]> wrote: > > > At today's DDF Michael and Stephen had a very good discussion > > > regarding tools for evolving ODL. One of these potential changes > > > is migration to Gradel / Bazel. This of course is not a simple > > > change. Not only would it significantly affect ODL but other LFN > > > projects as well. As such, I'm suggesting that ODL appoint a > > > representative to work the LFN community as a whole to > > > investigate the pros, cons, work effort and desire to migrate. > > > > On the topic of Gradle (v. Maven), I came across the following this > > weekend: > > https://blog.philipphauer.de/moving-back-from-gradle-to-maven/ > > > > It makes for interesting reading, even though it *is* just one > > anecdote and not generalisable. > > There will always be pro and con blog posts for any new tool and > technology... ;-) > > Steep learning curve probably with anything new; we're just more used > to Maven's quirks? :)
I don’t think that’s the main point of the article above re. the steep learning curve. As I understand it, the main issue that the author encountered is that *everyone* involved in the project had to learn Gradle and learn quite a lot of it, because unlike Maven it can’t be handled automagically by the IDE. If you look at ODL currently, most developers don’t know much about Maven (if anything at all), and they don’t need to. (That doesn’t mean it’s not useful to know Maven, but because IDEs know how to massage the POM files, it’s not necessary. IDEs also know how to manage some Gradle stuff, but because it’s a full-blown programming language, they can’t handle everything and the scope of what they *do* handle is narrower than in the Maven case.) > As to the why at all and is it worth it, for me the main driver > really is performance because of properly working incremental builds, > even on Jenkins for Gerrit. > > Re. Groovy instead XML, their idea seems to be Kotlin (as blog > mentions); BUT my PoV is that probably either is wrong - I expect in > a big project like ours one would have custom tasks and plugins > (which we can write in Java...) and should aim to have minimal build > scripts which are much more POM like declarative than imperative > after all, just for depenencies and the like. > > Ultimtely we'll have to figure out for ourselves if and when this > makes sense for us, if ever. Yup, we will indeed. I wasn’t trying to say the article above means we should drop it; I just found it interesting to add to the debate because it’s the first meaningful article I’ve come across which presents a project returning to Maven with some good arguments. Regards, Stephen
pgpDD2n7dGH9Y.pgp
Description: OpenPGP digital signature
_______________________________________________ Discuss mailing list [email protected] https://lists.opendaylight.org/mailman/listinfo/discuss
