Re: [picocontainer-dev] patches for *full* binding-annotation capability

--- Paul Hammant <[email protected]> wrote:


> > - IMO, breaking unit tests should not be
> acceptable under refactor  
> > (except for very specific and special cases, but
> not wholesale  
> > breaking).  If tests don't pass - refactor should
> not be committed.

Well, there are different kinds of test. Tests for
some functionality / class which is gone and replaced
by something else  going to be broken anyway - as they
become obsolete.   Test assuring overall behaviour
have to stay and be satisfied


> One outcome is that Konstantin has learned what the
> desired final  
> state is, and then performs it again as a series of
> refactorings on  
> trunk with no tests breaking at any point.  He
> already knows the risk  
> that he may get himself in a knot on the other
> branch that he can't  
> get out of :-)

Problem is, that parameter/adapter stuff is pretty
central to everything - so refactoring is bound to 
touch anything. 

> Konstantin - in principal - how unhappy would you be
> with reapplying  
> the changes @ apache based on the wisdom you gleaned
> from  
> experimentation ?

Maybe apache project would be a good starting point
for it - as we also have to cope with demands of
plexus. 

for now, I'd fix the branch.

Anyway, this refactoring is not useless - it already
unearthed at least one serious bug not seen previously
;) 

regards,

----[ Konstantin Pribluda http://www.pribluda.de ]----------------
JTec quality components: http://www.pribluda.de/projects/


      ____________________________________________________________________________________
Be a better pen pal. 
Text or chat with friends inside Yahoo! Mail. See how.  http://overview.mail.yahoo.com/

---------------------------------------------------------------------
To unsubscribe from this list please visit:

    http://xircles.codehaus.org/manage_email