[DISCUSS] 3.5.0 Release Plan

[Stephen Mallette <sp...@gmail.com>]

3.5.0 has been in the works for a really long time. Well over a year. It
hasn't really grown into everything I thought it would given the attention
paid to 3.3.x and then 3.4.x releases at the early part of the year. That
said, I think it has many important dependency/runtime upgrades that I
believe are important. Some examples include: Java 11, dropping Python 2,
JS Node 10, .NET Standard 2.0 and many other sorts of upgrades, like Spark
3.0.

We also have a much better testing framework around GLVs to focus on with
Translators driving that functionality. We also got rid of the GLV
templating system which should make it easier for developers to contribute
to them.

There are a few new features as well and I think that if we put some energy
into it over the course of our typical release cycle we could get a few
more neat things in there. Ultimately, though, I think those
dependency/runtime upgrades are what really drive us to a release here...if
we can make it a bit more than that, I think that's great but let's think
of 3.5.0 as a "heavy maintenance release" with the idea that heavier
breaking changes we had originally planned head to 3.6.0.

As of right now, just in studying the Upgrade Documentation, I don't see a
lot of risk for upgrading for either providers or users. There are some
obvious areas where compile issues will ensue but we don't seem to have
gross incompatibility. I think Gremlin Server/Driver protocol remained
unchanged between 3.4.x and 3.5.0 (though we may have gotten rid of some
deprecated settings), but with some testing we can confirm what's happening
there for sure.

I think we should look to release 3.5.0 in the next release cycle with
3.4.10. Let's avoid any breaking changes, especially in serialization and
server protocols (apart from perhaps deprecation removal). Finally, let's
try to drop in a few new features to make 3.5.0 good base of work 2021
while opening up for breaking changes again in 3.6.0 as soon as we release.
I would think our next release would be in the March/April timeframe as we
typically drop once every three or four months.

Happy to hear any thoughts on the matter...discuss away!

Re: [DISCUSS] 3.5.0 Release Plan

[Stephen Mallette <sp...@gmail.com>]

As there haven't been any objections or additional comments here, I think
we can assume consensus and look forward to seeing 3.5.0 in the next few
months. Nice to see this moving forward in this fashion. Thanks!

On Fri, Dec 18, 2020 at 3:19 PM Stephen Mallette <sp...@gmail.com>
wrote:

> 3.5.0 has been in the works for a really long time. Well over a year. It
> hasn't really grown into everything I thought it would given the attention
> paid to 3.3.x and then 3.4.x releases at the early part of the year. That
> said, I think it has many important dependency/runtime upgrades that I
> believe are important. Some examples include: Java 11, dropping Python 2,
> JS Node 10, .NET Standard 2.0 and many other sorts of upgrades, like Spark
> 3.0.
>
> We also have a much better testing framework around GLVs to focus on with
> Translators driving that functionality. We also got rid of the GLV
> templating system which should make it easier for developers to contribute
> to them.
>
> There are a few new features as well and I think that if we put some
> energy into it over the course of our typical release cycle we could get a
> few more neat things in there. Ultimately, though, I think those
> dependency/runtime upgrades are what really drive us to a release here...if
> we can make it a bit more than that, I think that's great but let's think
> of 3.5.0 as a "heavy maintenance release" with the idea that heavier
> breaking changes we had originally planned head to 3.6.0.
>
> As of right now, just in studying the Upgrade Documentation, I don't see a
> lot of risk for upgrading for either providers or users. There are some
> obvious areas where compile issues will ensue but we don't seem to have
> gross incompatibility. I think Gremlin Server/Driver protocol remained
> unchanged between 3.4.x and 3.5.0 (though we may have gotten rid of some
> deprecated settings), but with some testing we can confirm what's happening
> there for sure.
>
> I think we should look to release 3.5.0 in the next release cycle with
> 3.4.10. Let's avoid any breaking changes, especially in serialization and
> server protocols (apart from perhaps deprecation removal). Finally, let's
> try to drop in a few new features to make 3.5.0 good base of work 2021
> while opening up for breaking changes again in 3.6.0 as soon as we release.
> I would think our next release would be in the March/April timeframe as we
> typically drop once every three or four months.
>
> Happy to hear any thoughts on the matter...discuss away!
>