RDFLib 6.2.0 release planning

[aucampia]

Starting this discussion so we can stipulate what should be in 6.2.0 - please add any things you would like to see before the release as comments and then we can try make issues for it and try and implement it all.

CC: @gjhiggins @nicholascar @white-gecko @ashleysommer

[aucampia]

I think we should aim for a release more or less on 2022-06-01 - I would like earlier but I think that is most realistic. Ideally we should tag all issues we think we can fix within this time-frame.

[ghost]

I think we should aim for a release more or less on 2022-06-01 - I would like earlier but I think that is most realistic. Ideally we should tag all issues we think we can fix within this time-frame.

How about 2022-06-04? It would coincide with the 20th. anniversary of the initial release of RDFLib 0.9 after being split from redfootlib.

Also, because of the changes in API resulting from the "identifier-as-context-and-remove-ConjunctiveGraph" work, perhaps we should consider a 7.0 release?

[aucampia]

The date I had in mind was more considering just fixing a couple of bugs, and maybe getting #1686 in, and maybe some typing, basic quality of life stuff which I think is having a detrimental effect on our users. I think if we want to include #1814 this will not be realistic, though I may be wrong. I also think that it may be good to at least have everything we plan to deprecate marked as deprecated in at least one release before removing them, but I can also live without this.

[ghost]

No worries, just an off-the-wall suggestion.

[aucampia]

I'm up for it though, depending on how others feel, though I think there are many things we want to get into 7.0 - the other approach is to basically just kind of ramp up the interface breaking instead of waiting, though we may end up releasing a new major version every quarter if we do this.

[nicholascar]

I think we had previously sort of decided that changes to the graph identifiers would result in a 7.0.0 release, not a 6.x.

Other things to include (and I think these will easily be included, perhaps before 7.0.0 in a 6.1.x or a 6.x release due to current PRs) are:

• completion of prefix binding issues
• completion of test reorg & modernisation
• add more in here @nicholascar

We also have work on RDF* going on now (rdflib-rdfstar) and something may be ready for 7.0.0 release...

[ghost]

I think we had previously sort of decided that changes to the graph identifiers would result in a 7.0.0 release, not a 6.x.

That was my understanding, if it makes the cut.

[aucampia]

I think we had previously sort of decided that changes to the graph identifiers would result in a 7.0.0 release, not a 6.x.

I agree on this point, was thinking before this, I would like to get into a monthly, or at least quarterly release cadence if possible. We can maybe hold off on releases somewhat but we should also work on reducing the effort to make a release.

[nicholascar]

I definitely support a quarterly release cadence! I think monthly's a bit much but certainly having more frequent releases will reduce the effort needed for each release. The effort has reduced a lot with 5.0.0 > 6.0.0 > 6.0.1 > 6.1.0 > 6.1.1 effort, but we still do have a lot of stuff in master not yet in a release.

[nicholascar]

@gjhiggins can you please email me on [email protected] so I can send you emails?

[ghost]

I'm [email protected], your email addy is whitelisted.