Beer definitely owed - on 2 counts. 1stly, I flagged this null set up as a thing to do before I checked in, and then didn't do it. 2ndly, I still haven't got the test harness to work on my development machine (how bad is that).
Thanks for finding this. Isn't it great to have other people find/fix bugs rather than having to do it all yourself. ;-) -M Keith James wrote: >>>>>>"Matthew" == Matthew Pocock <[EMAIL PROTECTED]> writes: >>>>> > > [...] > > Matthew> I am going to check the isUnchangeable code-base in this > Matthew> evening. Many many files are touched, so I would be very > Matthew> supprised if all old applications ran without a glitch > Matthew> against the new code-base without modification. Please > Matthew> send the list bug reports if you experience problems. > > I guess you checked in without running 'ant runtests' as there are > roughly 140 (yes, really) tests failing. > > Most of the problem seems to be the ChangeSupport constructors > chaining through with a hard-coded null value for a Set which gets > passed to the HashSet constructor. I've replaced this with > Collections.EMPTY_SET for now and there are 7 tests still failing, > mostly in feature projection. > > Beer owed! ;) > > These are cool improvements, by the way. > _______________________________________________ Biojava-l mailing list - [EMAIL PROTECTED] http://biojava.org/mailman/listinfo/biojava-l
