RE: [nebula-dev] Nebula Build

["Peter Centgraf" <peter@xxxxxxxxxxxx>]

> I don't have a strong opinion here but I don't know that using 0 as the 
> major ver is a good idea.  When developing a set of inter-related 
> plugins you want to be able to specify plugin dependencies as you 
> develop.  That is, I want to say plugin A requires plugin B 1.0.  If 
> plugin b is 0.7 then that won't work. 
> 
> Also, is this an Eclipse convention?  I haven't seen other plugins start 
> at 0.x when developing.

There doesn't appear to be a way to express "no guarantees" using the
Eclipse versioning conventions, so using zero would be a Nebula quirk.
Frankly, I don't think any other projects are this lax about API
compatibility.  I thought this would be an easy hack to bypass the version
checking during build.  It's also reasonably clear that no formal release
has been made yet.  I hope that our widgets will all move to a 1.0 release
soon, but I don't think anyone is ready to commit to that yet.

I don't have a strong objection to starting with 1.0, as long as the minor
version gets incremented when an API change is made.

> I created the feature when I started the build work but in the end I did 
> not use it.  It could also be deleted.  I don't think we ever expect all 
> the widgets to be updated as a group.  I don't even think most people 
> will want to use the update manager to update individual widgets.  Most 
> end-users probably don't want to think about downloading individual 
> widgets.  I'm not sure if the ROI is there for any work on Nebula 
> features.

I would have found this convenient a few months ago.  Back then, I didn't
have access to external CVS servers through my company's firewall, so I had
to constantly re-download and extract the jars for each widget.  An update
site would make this a lot easier.  The main reason to do it would be to
allow other Eclipse projects to consume Nebula widgets, though.  It's also a
requirement for the coordinated releases. 

> I would definitely appreciate work on these things.  Nebula suffers a bit
> because none of the committers have much time.  None of us work on Nebula
> full-time.  If you make these changes, submit patches then we have no 
> excuse :)

I'm not full-time on this, either, but I'm trying to make the most of my
newbie enthusiasm + anal-retentive personality.  :-)

> One last thought on the moon-of-Jupiter release, its of course necessary 
> that each plugin actually prepare a release for next summer.  Individual 
> plugins may decide to release earlier or later.  So I'm not sure how 
> easy it would be for us to jump on those releases.  I don't think 
> they'll know how to handle us.

I actually think it will be fairly easy for us to participate, since our
projects tend to be very small and have light dependencies.  If we follow
the conventions, we can ride the coattails of the more complex projects and
use their build tools.  Of course, we'd have to get the widgets up to some
standard of "release quality", but we have a year to do that.  If any of the
widgets are still not ready for prime time by then, we can exclude those
specific plugins/features from the release.

--
Peter