Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Consider deprecating support for R versions older than 4 #13

Open
jdhoffa opened this issue Jan 23, 2024 · 8 comments
Open

Consider deprecating support for R versions older than 4 #13

jdhoffa opened this issue Jan 23, 2024 · 8 comments

Comments

@jdhoffa
Copy link
Member

jdhoffa commented Jan 23, 2024

Version 4 has been available for around 4 years now. I'm not sure how long we want to offer support.
Probably dangerous, and user-facing but we should at least discuss it.

We already skip a lot of tests on anything older than 4, which may give us fall security into thinking that things "work"

@cjyetman
Copy link
Member

relevant read https://blog.r-hub.io/2022/09/12/r-dependency/#:~:text=this%20is%20the%20strategy%20used,up%20to%205%20years%20old.

@AlexAxthelm
Copy link
Contributor

@jdhoffa Are we skipping the tests because of the code used in testing, or because of the code used in the app?

@jdhoffa
Copy link
Member Author

jdhoffa commented Jan 23, 2024

Not sure in every case to be honest. In the most recent case (that I added), it was because of code used in testing.

@jdhoffa
Copy link
Member Author

jdhoffa commented Jan 24, 2024

Transferring this issue to https://github.com/RMI-PACTA/demo_actions/issues

@jdhoffa jdhoffa transferred this issue from RMI-PACTA/r2dii.plot Jan 24, 2024
@AlexAxthelm AlexAxthelm transferred this issue from another repository Jan 24, 2024
@AlexAxthelm
Copy link
Contributor

One more thought here: When I was working in government, it was not uncommon for some of the software I used to be very outdated. If the P4B code supports the R 3.x releases, maybe we should keep it that way. 3.0.3 came out ~10 years ago (which is coincidentally how old the version of SQL Server I was using was before it got end-of-life'd on security updates, and we were forced to update).

Point being: in a security controlled environment, new features aren't compelling reasons to upgrade (I would not be surprised if some of those environments are still running R 3.0.0).

@cjyetman
Copy link
Member

I feel like R v3.5 (April 2018) was a really significant release with the introduction of the ALTREP framework and the introduction of v3 serialization for native R objects (e.g. <3.5 cannot read the default versions of RDS/RDA files created by 3.5+)... so that feels like a good benchmark minimum.

@jdhoffa
Copy link
Member Author

jdhoffa commented Jan 26, 2024

All valid points. And Banks are likely in a similar setup to governments with restricted setups that may only support older versions, so perhaps 3.5 is a good baseline.

@AlexAxthelm
Copy link
Contributor

3.5.0 makes sense in that case. In any case, I think this relates to #33, where a good check here is to try running things on 3.5 and see how many things don't work.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

3 participants