Anyone still sitting on the fence at this point is probably going to sit there through the final release, or would poke around for weeks before looking at it.

Personally, I say we fix the 8 issues, release B4, and if nothing critical comes up in a week or ten days, go to Struts 1.1 final. (Hey, how about that for a Valentine's Day present? =:0)

I do believe that we have a high-quality release that fulfills the Apache standards. But, realistically, the codebase is closer to a Struts 2.0 release than Struts 1.1. Things are likely to come up after the final release, regardless of how many or few release candidates we publish.

This codebase has been in a beta for going on a year, and five different books have been published about it. I should hope it's ready for prime time by this point =:0)

So, I'd say lets cut to the chase. Do B4, and if it's good, let's just go with it. If the fence sitters come up with anything once final ships, we go with an early Struts 1.1.1. Let's get the momentum up, and trust ourselves to do the right thing when the time comes.

-Ted.


James Turner wrote:
Ted said (I just love that aliteration...):

My suggestion would be to schedule a Beta 4 against the nightly build, and then to not hesitate releasing B4 as Struts 1.1. final if it flies. The idea being we suspect that B4 is a "defacto" release candidate, and may go from B4 to Release, if appropriate.

The only different is that if we call it an RC, people who have been
sitting on the fence waiting to adopt might actually try 1.1 and we can
flush out any hidden bugs that are only uncovered by strenuous
applications testing.

One thing I would discourage would be cutting a Release Candidate if we "strongly suspect" that we will need another RC.

I strongly suspect only because I trust in comrade Murphy, and that
there's Always One More Bug.


One definition of Release Candidate is

o "Release candidate" means we think this build is it, and will only change high priority bugs -- and API changes are totally verbotten.

I think that's where we are, aren't we?  I wouldn't be shocked if an RC
based on a currently nightly build couldn't become the final.  Declaring
it a release candidate is more a signal than anything, it says: "The
didling is done, we intend to release the final product very shortly,
this might be it, start using it so you'll be ready."

James



--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>



--
Ted Husted,
Struts in Action <http://husted.com/struts/book.html>


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to