I sent a query to derby-user about a user complaining about startup time
of Derby. I thought I would add to this list that I think we're
experiencing the same problem in our testing, except that our way of
fixing it is not to try and fix the startup time of Derby but to change
the way we test so that we don't start Derby multiple times.
I think this is one way to work around the problem, but I also think
it's worth understanding the problem. I thought I'd send an email about
this first to get everyone's take on this, but will likely log a JIRA
for this at some point, even though I don't have a reproduction nor do
we have a specific way to measure it. Perhaps the right way to do this
is to add a regression test that test startup time and that way we can
start measuring it and shoot for reducing it or at least not making it
worse across releases.
David
