[wtp-dev] Time to step up our versioning compliance checks?

[David M Williams <david_williams@xxxxxxxxxx>]

In sorting through the versioning anomalies
as we approached M1, it's made me realize that, as a project, perhaps we
should increase our vigilance on making sure things are versioned correctly
even for every build. 

The issue can be illustrated by the
points below. 

Notice that even in our final M1 candidate,
there is an error produced in the qualifier suffix for 
org.eclipse.wst.ws_tests.feature 
2.0.0.v200706041905-6-CN9oA55P5O987GB_ (current) 
2.0.0.v200706041905-6-CN9oA55P5O987G_L (reference)

What ever the reason, we can live with
it for our M1, no need to respin!

And I'm sure at some level this is a
limitation (or bug) with the qualifier suffix computation algorithm 

But, on the other hand, it may also
be due to the fact that we allow lots of bundle "qualifier-only"
errors to exist.

There are 3 in 2.0.1 recent build, 
http://download.eclipse.org/webtools/committers/drops/R2.0/M-2.0.1-200708230357/versioningReport_2.0.1.html#bundlequalifier

And a few dozen in our M1 candidate
http://download.eclipse.org/webtools/committers/drops/R3.0/S-3.0M1-200708231155/versioningReport_3.0M1.html#bundlequalifier


This may not be too much of a practical
problem right now, but may be eventually. 

My main concern is that even though
WTP 3.0 M1 is correct for our relationships in 2.0.1 build right now, 
that next week, as we do more 2.0.1
work, it may quickly become incorrect. 

Hence, I recommend we start counting
"qualifier only" warnings as "errors" for our JUnit
tests work discipline ourselves 
to increase service versions more appropriately
... .namely, when the service first goes into a stream? (Or, in some cases,
these 
"qualifier only" increases
might actually deserve a minor version increase!?). 

Let me know if any suggestions for other
options, otherwise I'll change our tests next week to be stricter. 
And, as always, let me know if questions,
especially if/when errors start to pop-up next week. 

Thanks,