james-server-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Bagnara <apa...@bago.org>
Subject 2.3.0a3 release plans
Date Sat, 13 May 2006 11:35:33 GMT
Hi all,

As you probably noticed a2 release proposal "failed" early.

I first tagged "r400018" as a2 candidate, then we found a Javamail 1.4 
regression, fixed it and tagged r400140 for release.

Starting new refactoring work I found out a new bug in the 
MimeMessageWrapper/CopyOnWriteProxy things.

A little history of that bug is here:

The commits involved in the fix are:
r404802 - Few more tests (work in progress) for NPE and IOExceptions in 
message cloning/sharing code (JAMES-474)
r404787 - Added one level of indirection (field=>getter) over the 
wrapped message. Now the wrapped message is kept inside the 
referencetracker (more clean, easier to read)
r404783 - Previous commit was not a solution, trying a new one after a 
new failing test has been added (JAMES-474)
r404782 - One more test for the MMCoW object (still investigating on 
r404780 - Possible fix for an NPE whose test has been just added to 
MimeMessageCopyOnWriteProxy (maybe related to JAMES-474)
r404779 - Further changes to MimeMessageCopyOnWriteTest to better check 
the expected behaviour and to avoid leaving garbage in the temp folder.
r404773 - Refactored the MimeMessageCopyOnWriteProxyTest and added a 
proof of a bug (maybe related to JAMES-474)

How do you prefer to proceed?

a. Should we branch from r400140, merge the above changes and try an 
"a3" release?

 From there the only known showstopper is JAMES-490:
I already solved this for a SocketFactory I use in a custom 
remotedelivery implementation and it seems to work, so I'll commit a fix 

b. Should we otherwise try to release as alpha3 the current trunk as 
soon as JAMES-490 will be fixed? This time I would identify the "stable" 
revision, run a build of that revision and publish it on my apache home 
for people to test, if no showstoppers arise I'll tag it as alpha3 and 
ask Noel to prepare the release and start the vote.

Imho it does not make much sense to "branch" for alpha releases so I 
would go for the latter solution, but I understand that you may have 
different opinions.


To unsubscribe, e-mail: server-dev-unsubscribe@james.apache.org
For additional commands, e-mail: server-dev-help@james.apache.org

View raw message