Thanks, Kathey, I re-read your proposal. I guess in principle I think this is a good thing, but it seems like an awful amount of work. But I am open to holding this as a vision and supporting it; backward compatibility is a very important feature, I think we all agree.

I think ConnectionEnv is a better name than Behavior(iour/ioure)Checker

David

Kathey Marsden wrote:
Daniel John Debrunner wrote:


David W. Van Couvering wrote:




I'll check this diff, but one quick comment: a lot of these diffs look
like formatting diffs.  Shouldn't we be using the 10.2 canons even if we
are using applications built against 10.1?

No. Think of the 10.1 tests as an existing customer application. Any
changes seen by the tests may also be seen by real customer
applications. I'm suprised there were that many failures, though I
haven't seen the breakdown. I could see the addition of SQLStates for
client messages would make a number of tests fail.





I posted a proposal earlier on this thread  regarding how to possiblly
automate this testing for java based tests and would appreciate your
input.  This  proposal offers no solution for .sql tests, which present
a large number of diffs, but my thought is that moving forward we will
be moving mostly to java assertion based testing.
http://www.mail-archive.com/[email protected]/msg12668.html

In the context of the JUnit discussion, I was thinking that instead of BehaviourChecker the class be called ConnectionEnv so it can
potentially be expanded to include JVM based logic as well.  So your
test would have somthing like:

import org.apache.derby.functionTests.util.ConnectionEnv;

ConnectionEnv connenv = new ConnectionEnv(conn);
if (connenv.supportsXXX())
    ......

in ConnectionEnv,  supportsXXX()  will make the decision based on
driver, database version, driver version, jvm etc.


Kathey


begin:vcard
fn:David W Van Couvering
n:Van Couvering;David W
org:Sun Microsystems, Inc.;Database Technology Group
email;internet:[EMAIL PROTECTED]
title:Senior Staff Software Engineer
tel;work:510-550-6819
tel;cell:510-684-7281
x-mozilla-html:TRUE
version:2.1
end:vcard

Reply via email to