Little worried about the ListView/TreeView/TableView classes
jonathan.giles at oracle.com
Tue Jul 10 13:08:21 PDT 2012
Of course bug fixing is very high priority for us. We try our very best
to ship releases with the most critical bugs fixed, and as few of the
other level of bug as is possible. The flow of incoming bugs never
stops, and sometimes we just have to draw an imaginary line in the sand.
It really has no reflection on you or the quality of your bug report -
it's just that if we didn't do this, we'd never make a release.
What isn't entirely obvious to the rest of the community is the speed at
which our releases start to ramp down, entering what we call high
resistance mode. In this stage of the release, it becomes more and more
difficult to make fixes as the risk / reward ratio has to be conveyed to
a team of people who form what we call 'bug court'. The ratio has to
better as the weeks wind on.
Many of the issues that have been filed against me recently would be
very risky in my opinion, and whilst the reward is also high, I could
not comfortably defend making a change at this stage of the release
without potentially causing follow-on issues that would be worse and /
or be considered showstoppers - bugs that stop a release from happening
or from other teams (such as the Scene Builder team) from being able to
build with the latest promoted code. This is why I am cautious with
these bugs. Trust me, I wish I could fix them all right now. I really do.
We are already in high resistance mode for JavaFX 2.2. You've probably
noticed the number of emails coming from our pushes to the repo has been
diminishing in the last few weeks.
Of course, the flip side is that you can be far less cautious at the
start of a release, as it is a long way off from the next release (no
matter how short the cycles are). I hope to find time to really get
through my bug backlog - and I hear you loud and clear what issues are
causing you pain. I just hope I don't need to jump straight into feature
work - trust me, I've been asking for more focus on bug fixes for a long
JavaFX releases are synced with Java update releases, which I believe
are every three months.
Personally I would love to do a minor 2.x release prior to the next
major release so that we can iterate on bug count and a few minor
features. It probably won't surprise you to say I've already mentioned
this numerous times to the people that make these kind of decisions but
I'm not aware of the current plans - I'm just a lowly engineer here,
trying to give you guys the best darn API (with as few bugs as possible) :-)
I hope that clears up some things.
On 11/07/2012 6:57 a.m., Christian Schudt wrote:
> John, I share your worries. During the last two months I reported 13 bugs. 5 of these were targeted for Lombard, 6 are still unscheduled and only 2 were scheduled for 2.2.
> Generally speaking I got the impression, that there are more important issues, as can be also seen here in the last comment:
> http://javafx-jira.kenai.com/browse/RT-20616 (which is already Priority Major)
> I don't know what these priorities are, but in my opinion fixing bugs and stabilizing existing code, should be high priority for a new release, especially if they are a regression from a prior version, like this one:
> http://javafx-jira.kenai.com/browse/RT-21936 (worked in 2.1 beta and is now targeted for Lombard, too)
> I would appreciate, if the JavaFX Team let the community know, what these priorities for 2.2 are and when it is supposed to be released.
> Have you (Oracle?) considered to release further minor 2.x releases until JavaFX 3.0? I worry, that 2.2 won't be stable enough to work with until Q3 2013.
> Best regards,
More information about the openjfx-dev