On 5/29/07, Stephane Bailliez <[EMAIL PROTECTED]> wrote:
Gilles Scokart wrote: > The mail of Jean-Baptiste raises a question: What is our published API? > > For the moment we only considered the backward compatibility of our ant > interface. But that should maybe change. > > Shouldn't we start to define what is our published API, what are the > internal classes subject to changes, and what are the 'stable' interface > that can be reused? > > I guess that's an important information for an open source project. But it > is also something difficult and it might quickly be heavy. > IMHO the code is not ready to effectively have a published API yet
Agreed and
if you do that this will affect further your ability to evolve it without breaking everything. (and it's hard enough from a behavior point of view) You'll have to be pretty strict on what is published and I'm not sure it can be done right now. This innocent change of setting 2 public getters to get the results post execution is actually interesting in itself.- - it does not need to be public but more protected. - I find it a bit unusual to have getters for post execution results in Ant tasks, I'm very uncomfortable with this generally
How would you make those results available? With a reference in the Project? Or maybe we should recommend to use the base unstable API instead of trying to use the ant task for that? Xavier - you may have to live with it for quite some time
-- stephane
-- Xavier Hanin - Independent Java Consultant Manage your dependencies with Ivy! http://incubator.apache.org/ivy/
