Nice, looking forward to the ever improving SHACL features !
As for 4.0, it would be a good occasion to look into GH2392 aka the ES/Solr/Lucene upgrades since those may break some stuff,
so I’ll try if I can come up with some tests.
Another thingie might be an overhaul of the console (GH1394): PicoCLI + JLine seems like a very nice combo,
and would make it easier to add some features (i18n, ANSI color) or to remove some boilerplate code…
Best regards,
Bart
From: rdf4j-dev-bounces@xxxxxxxxxxx <rdf4j-dev-bounces@xxxxxxxxxxx>
On Behalf Of Jeen Broekstra
Sent: donderdag 3 december 2020 0:02
To: rdf4j developer discussions <rdf4j-dev@xxxxxxxxxxx>
Subject: Re: [rdf4j-dev] timeline for 3.5.0 release
In the week of December 7, I'll do another 3.5 milestone build. This will not yet include the new SHACL AST. I may in parallel release a 3.4.5 patch release, btw (haven't decided yet).
Final release of 3.5 is planned for December 23. Havard is doing his utmost to get the AST in shape (hah!) to make that release deadline, so the plan is to include that as part of 3.5.
If for some unforeseen reason the SHACL AST rewrite does not make the release deadline, we will go ahead with a 3.5 release without it, and schedule a 3.6 release, possibly quite shortly after.
I also want us to start thinking about a 4.0 major release, ideally somewhere in Q1 of 2021. Most of the issues currently planned for it (see https://github.com/eclipse/rdf4j/milestone/30)
are about code cleanup and removal of legacy stuff. Now is the time to start proposing new features or improvements that require a major release though, and see if we have the resources to start working on them.
One major open question is also still where we stand on migrating to Java 11 (see https://github.com/eclipse/rdf4j/issues/2046). This is currently not planned for 4.0, but
there may be significant performance benefits that we should consider. I'd invite further disussion on that issue, and your preferences for a timeline on this.
On Fri, Nov 27, 2020, at 21:18, Jeen Broekstra wrote:
I'm aiming for a final 3.5 release in the week of December 7.
There's a couple of outstanding things, if we can get them stable and merged before that, great, if not, no worries, we just reschedule them.
On Thu, Nov 12, 2020, at 19:54, Jeen Broekstra wrote:
Didn't get around to the milestone build today unfortunately, I will take care of it tomorrow.
Havard, I had one or two minor comments on your PR for the 3.4.4 release notes, but apart from that it looks good to from my POV.
On Sat, Nov 7, 2020, at 18:45, Håvard Ottestad wrote:
Unless there are any objections, I'll do a 3.5.0 Milestone build this week (Thursday at the latest). There's enough merged and/or as-good-as merged stuff to warrant a milestone, and I'd like to able to test a few things in our (metaphacts)
platform without having to rely on snapshots.
Havard, I think you were also considering doing a patch release this weekend. I wasn't planning on doing a 3.4.4 patch release myself but if you want pick this up, that'd be great. If you do, can you make sure to include the outstanding
PR for GH-2604 (the RDF* transaction bug)?
It's currently waiting for final approval, if the build succeeds I'm happy for it to be merged (I'll merge it myself tomorrow). But I would appreciate it being included in a 3.4.4 patch release if we do one.
On Fri, Oct 23, 2020, at 14:56, Jeen Broekstra wrote:
No worries Havard, as said there's no pressure on this from my end, in your own time.
Also, fwiw: there's always the option of doing a 3.5 release without the AST changes, and then just reschedule that for a 3.6 release in another few weeks. So either way we're not blocked here.
On Fri, Oct 23, 2020, at 14:11, Håvard Ottestad wrote:
In light of my previous email, off mailing-list, I’m not sure when I’ll manage to finish the new AST. That being said there is really only one thing missing and that is the support for advanced targeting with target shapes.
Could you guys give me a couple of weeks to see how things shape up? I have every intention to finish the new AST sooner rather than later!
I thought it might be good to start thinking about a timeline for a 3.5.0 release, and what should go into that release.
If we look at
the project board we can see that there are quite a few issues done, but some really big features planned for 3.5.0 (including the SHACL AST rewrite, the Elasticsearch upgrade, and the Rio character encoding fix) are not quite finished.
The Elasticsearch upgrade (#2392) is currently marked as blocked. I believe though that all CQs have been approved so the only thing stopping us from merging is, I think, some documentation on how to migrate an existing index. Bart, is
that your understanding as well? I know you are currently focused on other responsibilities, do you think you will have some availability later on to help carry this over the line?
Regarding the SHACL AST rewrite, there is a massive "work in progress" pull request up. Havard, what is currently your personal timeine for getting this in a release-ready state? Can you give a ballpark estimate (no pressure, it's just
so we all know what to expect)?
The Rio character encoding issue is something that has been lingering in "blocked" for a while now, I will try and put some effort into gettting that over the line this weekend. I also want to spend a little time on the recent contributions
from Allesandro (and some of the followup issues flowing from that), so we have all that ready to go in a next release.
_______________________________________________
_______________________________________________
_______________________________________________
_______________________________________________
_______________________________________________
_______________________________________________
_______________________________________________
|
