Planning the 2.34 release

Developers
1

Since 2.33 was released on September 5th, we’re coming up on the four month cycle for a next release. To avoid the upcoming winter holidays, it seems prudent to delay the release by a week or so.

@rok_cesnovar and I are proposing the following dates:

  • Feature freeze on on January 8th, 2024
  • Release on January 16th, 2024 (the 15th is a holiday in the US)

Please use this thread if you have any thoughts on those dates, or if you need a pull request reviewed or think an issue is a high priority for this release.

3 Likes
2

Some issues that seem like they’d be nice to knock off:

3

Hey @WardBrian,

now all tests for the 7-parameter diffusion model run successfully. This is PR #2822 and issue #2682

Could we include this PR in the next release?

2 Likes
4

If the Math PR gets approved before the 6th or so, I think we can get a PR into the compiler to expose it before the freeze, and documentation can be written during the freeze

1 Like
5

I would very much like to see this go in

  • Add resampling options for multi-Pathfinder · Issue #3242 · stan-dev/stan · GitHub
    Now if the posterior is high-dimensional and far from normal, it is likely that multi-Pathfinder returns only copies of one draw2, which makes further diagnostics and proper multi-chain-MCMC initialization impossible. Thus, having an option to turn off the resampling would be very useful to increase the usefulness of Pathfinder. This doesn’t affect the algorithm otherwise, the option would just control whether the last step resampling is done.

Related but less critical issue is

6

Some PRs related to issues brought up in this thread:

7

The release candidates are out!

If nothing comes up in the next week, the release will occur on Tuesday. Further discussions can be had on GitHub

1 Like