As far as I know, there was no "vote" to block this fix, but there WAS a
vote on what would be in the release, and this was either approved by
omission or simply slipped though the cracks.
On that ground, I do disagree with Serge on one issue: "I think holding
people accountable for knowing 100% of what needs to be done in an onsite
project is unrealistic at best." I think that we all ought to become more
familar with the James area of bugzilla. Is it possible, someone please
tell me, to have an report generated and e-mailed reflecting open bugs on,
say, a weekly (serious bugs) or monthly (other requests) basis? Sort of a
tickler to remind us of what is outstanding, and annoy us into fixing them?
I think that IF we can fix this in an isolated fashion, then we might
consider fixing it for the same reason that we fixed the Oracle bug. This
appears to impact a large number of users, is a show stopper (AIUI) if you
reboot the server, and is a Top-5 FAQ by my informal guestimation. Yes, I'd
love to change the default config to use HypersonicSQL, but that is an even
bigger issue than fixing the bug.
The only decent way I can think of to fix this problem is as follows:
1. Take the code from Avalon for the version of the broken classes that
we are using.
2. Move those into a James package: org.apache.james.<wherever>
3. Change the imports, and use our classes.
4. Test.
5. Fix those classes. We have a proposed change ALREADY SUBMITTED.
6. Re-test.
This is the only way I can think of to make this a "simple bug" because I
would NOT want to fix those classes in the Avalon code, and have to rebuild
their jars. That would be far more risky, I think.
IF we follow the above steps, what exposure do we have with this fix?
Please let me know what issues I will need to test for, so that I can
accommodate them.
I further want to know if EVERYONE agrees that IF we can fix this cleanly,
that we have a Release. I can appreciate that people are frustrated. We
ALL want v2.1 out the door, but it has not been delayed by creeping
bug-fix-itis. The primary delay has been that the docs needed work, and no
one seems to have had the time to work on them other than Peter (and, to a
minor extent, me). From the little feedback of any kind that we've had to
them, the new docs are considered a vast improvement, and my hat is off to
Peter for all of his hard work on them.
In the future, I hope that we can keep the docs and code changes more in
synch. I'd also like to see us do minor releases with more frequency,
similar to tomcat, rather than go the better part of a year between major
releases. But that is another discussion.
EVERYONE (Peter included) agrees that this bug is "very bad." I am off-line
for the most part for another few days, but if no one else has the time to
attempt this bugfix, I will personally do it next week. I think that we are
ready for the Release, we have a good product, we do want to be able to
SUPPORT it, and based upon that last issue, I am willing to take the time to
fix this rather than immediately put out a 2.1.1 (think of that marketing
message -- an immediate patch for a long-standing bug of this nature), but I
don't want this to inspire any further delay. I, as does EVERYONE, want to
ship this excellent product.
And, again, for future releases I would like to discuss that weekly/monthly
status e-mail so that we are ALL more aware of defects.
--- Noel
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
