How hard would it be to abstract the filesystem access completely and
go directly against in-memory implementations, rather than "getting
sloppy" with disk i/o?
6 seconds per test is still pretty expensive time wise.
-Brian
On Feb 14, 2005, at 1:44 PM, Sunitha Kambhampati wrote:
Daniel John Debrunner wrote:
Dain Sundstrom wrote:
That looks like it would do it. Did you time the code? I'm curious
how much faster is database creation (create db, tablses, insert
data) and how much faster is say some select and update queries.
I ran a quick check with a single Derby test suite and overall it was
around twice as fast, database was created and test cases run and
results checked in under six seconds per test.
So I made the necessary changes per Dan's email; built insane jars
and then ran the entire derbyall test suite.. With these options set
, derbyall test suite ran around twice as fast on my laptop (t40;ide
disks).
Sunitha.
