Release ggplot2 3.1.1

[thomasp85]

Prepare for release:

• devtools::check()
• devtools::check_win_devel()
• rhub::check_for_cran()
• revdepcheck::revdep_check(num_workers = 4)
• Polish NEWS
• Polish pkgdown reference index
• Draft blog post

Submit to CRAN:

• usethis::use_version()
• Update cran-comments.md
• devtools::submit_cran()
• Approve email

Wait for CRAN...

• Accepted 🎉
• usethis::use_github_release()
• usethis::use_dev_version()
• Finish blog post
• Tweet
• Add link to blog post in pkgdown news menu

[thomasp85]

@hadley @karawoo @clauswilke @yutannihilation we will make a minor release of the current master shortish, in order to make the integration of the new unit system in grid as smooth as possible so it can be part of the R 3.6 release... There are no breaks in reverse dependencies, so it should be smooth sailing

Please hold of on merging anything into master until the release process has ended

[clauswilke]

I think we should address #3158 before a 3.2.0 release, so we don't release a new API we immediately invalidate with the next release.

I would delete these five lines and not make any other changes:

ggplot2/R/layer-sf.R

Lines 3 to 7 in 033fb52

	#' @export
	#' @include layer.r
	#' @rdname ggsf
	#' @usage NULL
	#' @format NULL

@yutannihilation suggested to remove the layer_class argument from layer(), but my preference is to leave it in. Even if it isn't useful outside of ggplot2 at this point, it will be useful for internal ggplot2 code and substantially reduce code duplication. We could add "for ggplot2 internal use only" to the documentation to make this clear.

[thomasp85]

I'll make sure to remove those lines

[clauswilke]

@thomasp85 Thanks! I'm traveling through the end of the week and won't have much time to look into anything before the weekend.

[thomasp85]

No problem... there is not much to do as there is zero revdep problems :-)

[dpseidel]

Great to see another release is coming together! Perhaps it's too late now with the already revdeps done, but is there any interest in merging #3040 to fix the sec.axis issues before release? It's been ready for review since February but we never got it merged.

[thomasp85]

I’m afraid we’ll have to postpone it. The thing has gotten a little stressed due to syncing with grid and the next r release and we can’t really take bigger code changes in now. We’ll have it in the next release🙂

[yutannihilation]

@thomasp85 Do you know why ggplot2 stays in the CRAN pending queue? Are there anything we can do if there's some issue?

ftp://cran.r-project.org/incoming/pending/

[thomasp85]

My guess is that they are busy with the 3.6 release process

[yutannihilation]

Aw..., OK, let's just wait.

[thomasp85]

Ok, quick update on this. The first release (current master) had a surprising number of breaking reverse dependencies so we decided to make a hotfix 3.1.1 release instead, only including the changes necessary to play nice with the updated grid. It is now on CRAN 🎉

This gives us some more time to make a proper 3.2.0 release and I’ll create a milestone soon where we can begin to attach issues/PRs we want in. I still think we should go for a real update soonish 1-1.5 months, but the 3.1.1 release has removed some timing pressure from the process

[clauswilke]

Thanks for the update. I think a proper release cycle with a set cutoff date for new features, then a few weeks later a cutoff date for bugfixes, then release is the way to go.

[yutannihilation]

the 3.1.1 release has removed some timing pressure from the process

Cool. Thanks for handling the process!

[lionel-]

the 3.1.1 release has removed some timing pressure from the process

Just FYI, on R 3.6, CRAN ggplot2 produces messages at load-time because it overwrites quosure methods in rlang: r-lib/rlang#669

[lock]

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/