[Biojava-dev] Build system changes

[Thomas Down]

Once upon a time, Matthew Pocock wrote:
> OK - if this is a real problem, could we have an
> AbstractBioJavaTest class that nullifies all fields in
> tearDown using introspection? Not sure how the java
> securtiy access will play with this, but anything to
> reduce coding...

I think that might be the most practical solution right now.

One worry, though, is that if you watch memory usage during
an 'ant runtests', it creeps up pretty rapidly even when
running tests which only use a few small objects.  I'm wondering
how much of the problem is actually data, and how much is due
to keeping lots of classes in memory.  BioJava is a pretty big
library, and once you count the classes themselves, plus lots of
tables to make the interpreter run fast, plus at least one
(and sometimes multiple) JITted machine code versions, it all
adds up.  If all the test classes are hanging around indefintiely,
and referncing lots of classes in BioJava, we might well have
a bit source of leaking memory right there :-(.

> Are the JUnit guys aware of this? Is it a bug,
> feature, or like the human chin - just a side-effect
> of the rest of the system?

I was looking at the source for the ant junit task last
night trying to work this one out, but it just made my head
hurt.  I'll try again at some point.

     Thmas.