Release network 1.17

[krivit]

Checklist before submitting to CRAN

Note that the final submission must be done by @CarterButts.

Directions for use

1. Create an appropriately named issue (e.g., "Release package version version.").
2. Copy-paste this file into the issue text to provide the checklist functionality.
3. After a commit directly pursuant to a checkbox item, uncheck all checkboxes below it.
4. After any other commit to master branch, uncheck all boxes from the "master branch" section down.
5. After any other commit to *release* branch, uncheck boxes from the "Release branch" section down.
6. Close the issue after all boxes are checked.

master branch

• DESCRIPTION: Package's CRAN dependencies require a recent CRAN version.
• DESCRIPTION: Package's dependencies that are a part of R core (base, datasets, grDevices, methods, splines, stats4, tools, utils, compiler, graphics, grid, parallel, stats, tcltk, translations) do not specify a version.
• DESCRIPTION: Package's dependencies that are usually shipped with R (KernSmooth, Matrix, MASS, class, nnet, spatial, boot, cluster, codetools, foreign, lattice, mgcv, nlme, rpart, survival) do not specify a version unless necessary.
• All tickets associated with the release's milestone have been resolved.
• The stubs file in inst/include/ is up to date if needed.
• Roxygen has been run on the package since last commit to code, documentation, or DESCRIPTION.
• Relevant tutorials run satisfactorily.
• inst/NEWS.Rd: file is up to date: none of the commits since the last commit to explicitly mention updating the NEWS file are newsworthy.
• LICENSE: license refers to the correct package.
• LICENSE: affiliations are up to date.
• LICENSE: copyright year range is correct.
• All files that should have them have copyright headers. Generally, it never hurts to rerun addcopyheaders to ensure this.
• R CMD check passes with ENABLE_statnet_TESTS=yes environment variable and --run-donttest.
• rchk static code analysis passes.
• Reverse dependency checks pass unless breakage known.

Release branch (v1.17)

• Release branch has had the latest master merged into it. (git checkout *release*; git merge master)
• DESCRIPTION: There are no Remotes: directives (unless releasing simultaneously).
• R-CMD-check.yaml: rcmdcheck() is called with error_on = "warning".
• DESCRIPTION: Release branch's version number does not have commit number attached to it.
• DESCRIPTION: Date is today.
• CITATION: update_CITATION has been run after installing all the mentioned packages from the release branch.
• All files in the release branch have been committed and pushed to GitHub.
• *pkg*.tar.gz: the candidate submission file is from the latest version of the release branch.
• *pkg*.tar.gz: R CMD check passes with ENABLE_statnet_TESTS=yes environment variable and --run-donttest.
• *pkg*.tar.gz: R CMD check passes with --as-cran option without WARNINGs.
• *pkg*.tar.gz: R CMD check passes Win-Builder (https://win-builder.r-project.org).
• *pkg*.tar.gz: R CMD check passes on the previous stable release (unless there is a good reason to require the latest R).

After release

• Release branch corresponding to the package on CRAN has been tagged with the exact CRAN version and pushed to GitHub.

[krivit]

I have created a release branch v1.17, in which I have eliminated dependence on statnet.common 4.5: it now works with the CRAN version. I had informed the reverse-dependency developers of the changes and the breakages on 20 May.

As far as I know, there is nothing blocking the release of network now. Given all the other packages in the pipeline, the sooner we release, the better.

Some of the items on this checklist are not applicable, and others I can't complete because network uses "nonstandard" LICENSE file and file headers, I am not very familiar with its C API or tutorials, and I don't know what should be done with #59.

@CarterButts, @mbojan, @chad-klumb , can I ask you to please it the rest of the way?

[CarterButts]

v1.17 is now passing checks on Rhub and all three versions of winbuilder, plus my at-home checks. So I think it's good to go. Unless I hear immediate objections, I will begin the transfer. Thanks @krivit and @martinamorris for your help w/pre-checks!

[CarterButts]

Package has been pushed. Will update and close when/if it passes muster on their end...

[CarterButts]

Argh. JSS has moved from http to https, so it failed because the URLs weren't of the correct type. And, of course, because it is CRAN, no other testing utilities but their own internal one flagged this as an issue. After all these years, this remains a constant: they always apply super secret squirrel tests on packages that are never covered in their other testing utilities (including WinBuilder and friends). Why do they make such work for themselves? Like wavefunction collapse, it is a deep and impenetrable mystery.

It does appear, in any event, that one's external measurement of a package status and that of CRAN do not diagonalize in the same basis.

Fixed the issues and am trying again.....

[CarterButts]

OK, they may have taken it. Not yet clear. Will check tomorrow, and update things accordingly.

[martinamorris]

lol, super secret squirrels. fingers crossed.

[CarterButts]

The latest hurdle is the revdep issue. Have explained it. Hopefully, they will weigh us lightly on the scales of Ma'at.

[drh20drh20]

Scales of Ma'at... Now there's a reference I had to look up.

Indeed, may release 1.17 be judged lighter than the proverbial feather by the priests of CRAN!

[CarterButts]

The package has passed the final acceptance phase. The package will propagate to CRAN in a few hours. You may start your landing.