Testing the Fuseki fixes has been done - Laura Morales has used the development build to check out the fixes and confirmed they are fixed.

The Jenkins job, other than the Shiro issue, is building cleanly. There are no changes around OS issues except TDB2 fixes (and TDB2 is "experimental").

More testing is always good but it takes time. The minimum is some testing and 3 +1 votes on process and legal; after that more is better and I'd say the criteria is "is it better than 3.5.0?", not some notion of "perfect".

If we have this week before a release can start, some final things for 3.6.0: (this is 3.6.0 so a few actual changes can happen, not a 3.5.1)

1/ The jena-text documentation improvements
2/ Downgrade shiro to 1.2.6
3/ riot: status code on warnings (#315)
4/ Ideally, dataset assembler (#314) [might be too tight for time].

Anything else?

Rob - I can merge #315 and we can sort out the implementation stuff later.

    Andy

On 25/11/17 23:45, ajs6f wrote:
Ditto, except for me it's the 8th.

ajs6f

On Nov 25, 2017, at 6:12 PM, Bruno P. Kinoshita 
<brunodepau...@yahoo.com.br.INVALID> wrote:

I can run the build and verify signatures any day in the next weeks. Just not 
much time to properly test Fuseki and review changes until after Dec 3rd.
CheersBruno

      From: Andy Seaborne <a...@apache.org>
To: "dev@jena.apache.org" <dev@jena.apache.org>
Sent: Sunday, 26 November 2017 12:02 PM
Subject: Jena 3.6.0?

The bug in Fuseki that causes UI uploads to fail, and some other UI
issues, is a bit annoying.

Is there the energy and time to vote on a 3.6.0 release if I build one?
Please respond if you'll be able to vote in the next few weeks.

If there is - from our experience last time, we can test the latest
development builds now, before a formal VOTE which will shorten the time
in case there is any problems to address.

     Andy

The build is complaining about a Shiro issue - this is harmless and a
problem somewhere in the Fuseki tests. Some state is getting initialized
twice.  It does not happen when Fuseki is run nor does it cause any
tests to fail.  It happens because of the 1.2.4->1.4.0 Shiro upgrade ;
it comes in at 1.2.6 -> 1.3.0. Solution: ship with 1.2.6

"""
[...] IniRealm  WARN  Users or Roles are already populated.  Configured
Ini instance will be ignored.
"""

     Andy




Reply via email to