On Oct 6, 2006, at 7:47 AM, Ken Tam wrote:
I was blocked this morning by an autowiring error running the Spring
integration BootstrapTestCase -- it kept complaining that it couldn't
resolve an autowire to BuilderRegistry. After rolling back pretty
much all the local changes that I thought might be even vaguely
relevant, the problem remained, so I pulled down a pristine tree to
see if it manifested there. It didn't, but doing that build seemed to
have some effect (perhaps through my local maven repo) since then the
formerly "dirty" tree also started to work.
Thought I would mention it in case we're accumulating evidence for
some race condition or weird corner case in the build/bootstrap
sequence.
I had similar weird build issues last night making the changes to
SCATestCase to look for the application SCDL using a specific
classloader. A build would work, but make a small change and things
would break downstream.
I put it down to a lack of certainty about which SCDL was being found
on the classpath - I don't think Maven guarantees the order in which
dependencies get added which means there's no guarantee which
resource will be returned by getResource() if two resources exist
with the same path.
The changes I made should resolve that as they now look for SCDL in
the same location as a specific class rather than anywhere on the
classpath.
--
Jeremy
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
