I committed your latest patch to the trunk. It seemed like a number
of issues were starting to back up behind this one, so felt it
reasonable to get this change into the trunk where a number of people
could experiment with it, and verify if it fixes similar issues of if
the other ones are different. It also means we get a number of
nightly tests run in different environments sooner than later.
I also think it is reasonable to hold up on updating the branch until
you have done the work you plan.
Bryan Pendleton wrote:
Mike Matrigali wrote:
Bryan do you think it is a problem checking in your patch for 1583
prior to doing the 1724 research? Are you planning to commit, or are
you looking for someone else to commit the patch?
Hi Mike, thanks for the feedback.
I don't think it would be a problem to commit the patch now. I think the
patch is straightforward and it certainly prevents the NPE, and it has some
regression tests to help us keep an eye on this code in the future.
derbyall
runs are completely clean.
The questions that Mamta and Yip have raised are good ones, and they
indicate
that I (we) don't really understand the circumstances that cause some
ColumnDescriptor instances to have valid TableDescriptor values, while
other
ColumnDescriptor instances don't.
If you would like to commit the patch now, to resolve the urgent problem
and get us moving forward, that would be great.
I am still intending to study the ColumnDescriptor handling more, and will
report on what I learned at that time.
If the patch hasn't been committed by the end of the weekend, and if my
study goes well, I was intending to commit it then.
But committing it now is fine with me!
thanks,
bryan
