[wtp-dev] Status of our builds as we approach RC1

[David M Williams <david_williams@xxxxxxxxxx>]

Yes, that's right, July 13th is coming
right up. Remember we are looking to *produce* the high quality RC1 on
Wednesday.

First, I'd like to thank Naci for improving
our build mechanics this weekend. Having tagged builds and a tagged "builder"
project is much more professional. 

Second, please be aware that we've to
setup builds to be "continuous" (meaning, immediate if modifications
are released). This means if you release something, it will trigger an
I-build. 
I'm still trying to understands CruiseControl's
concept of interacting time intervals, but the basic idea is it checks
at least once per hour and then waits a few minutes for no
changes, then does the build, if one
is not already in progress, and since a build-in-progress might take up
to two hours to finish  ... 

... this means, basically, once you
release, your changes should be available in an I-build in from 2 to 4
hours later. There's still plenty of room for improvement, name changing,
etc., so don't hesitate to contribute. 

Third, I've updated to Xerces 2.7.0
.. speaking of which, one of the JUnit failures over the weekend (with
XSD Validation tests) was entirely my fault, see https://bugs.eclipse.org/bugs/show_bug.cgi?id=103279
tor details and my penance. 

As luck would have that pesky erratic
org.eclipse.jst.ws.tests_
 failed in the last build of
the weekend. I know Bug 103167  has been opened for this, and
undoubtedly needs some fixes and investigations there, but I did the obvious
thing, and tried to run J2EEUtilsTests
all by itself and it fails ALL THE TIME, when ran all by itself. It fails
for different reasons than what's in the build logs ... but, it just makes
me wonder ... shouldn't all JUnit tests be "stand alone"? And
if one test depends on another test, or test suite, I imagine that would
make that test more fragile? Am I missing something? Or can this test be
improved?

Thanks all.