Ahh... that makes sense. And in fact, it sounds like why Github skeletons are preferable to most Maven archetypes.
On Sun, May 31, 2015 at 1:44 PM, Matt Burgess <[email protected]> wrote: > Sorry I worded that wrong, I meant that archetypes and templates often > generate empty skeletons, where your sample project would have your > existing working code. > > Sent from my iPhone > > > On May 31, 2015, at 1:43 AM, Ted Dunning <[email protected]> wrote: > > > > Empty skeletons? > > > > This project has working code. > > > > Besides, in my limited survey of development methods, git clone dominates > > over maven archetypes among developers I know. > > > > > >> On Sun, May 31, 2015 at 1:39 AM, Matt Burgess <[email protected]> > wrote: > >> > >> How about Maven archetype(s) and/or Lazybones templates, with working > code > >> examples instead of empty skeletons? Your examples below would make a > >> perfect project template, it works out of the box yet it is easily > >> refactored with an IDE for someone else's project. > >> > >> Sent from my iPhone > >> > >>> On May 30, 2015, at 6:36 PM, Ted Dunning <[email protected]> > wrote: > >>> > >>> Hey, > >>> > >>> I just built the beginnings of a project to contain sample UDF > >>> implementations to minimize the startup time of people wanting to build > >>> their own. > >>> > >>> See https://github.com/mapr-demos/simple-drill-functions > >>> > >>> Ping me if there are improvements to be had. > >>> > >>> Also, I am curious how we should make this available. I would like for > >> it > >>> be outside of the normal drill distribution so that people can see > what a > >>> stand-alone project should look like. > >>> > >>> This is related to the documentation which I found to be correct, but > >> less > >>> specific than it should be at the end (or beginning) where I think it > >>> should say: clone this repo, run these two commands and see something > >> work > >>> in drill. > >>> > >>> Any thoughts here? > >> >
