Tags give the ability to mark specific points in history as being important
-
v0.1.2
protectede1ed138b · ·raddr 0.1.2 First published release. 0.1.0 and 0.1.1 were both tagged during development, on 2026-07-31 and 2026-08-02, and neither was submitted to CRAN or published anywhere else. Both tags are retained as honest records of what was checked on those days rather than moved, and this tag supersedes both. Note that v0.1.1's own tag message opens "First published release". That was the intent at the time and it did not come true: the submission did not happen, and the tree that tag points at cannot be the one submitted. Tag messages are published records and are not rewritten, so the correction is recorded here rather than by editing v0.1.1. WHY A NEW VERSION RATHER THAN A MOVED TAG 1cf7fb8 -- the commit repointing URL: and BugReports: off the suspended GitHub account and onto GitLab -- is not an ancestor of eb5d5c4, the v0.1.1 tag. A tarball built from v0.1.1 therefore still declares: URL: https://github.com/bart-turczynski/raddr -> 404 BugReports: https://github.com/bart-turczynski/raddr/issues -> 404 Submitting that tag would have re-shipped the exact defect RADD-yrppvxdi spent three weeks discharging. Moving the tag was the alternative and is rejected for the reason eb5d5c4 gave when it rejected moving v0.1.0: the tag exists on origin and on the backup mirror, so moving it is a published-history rewrite, and v* is a protected ref. WHAT DIFFERS FROM v0.1.1 DESCRIPTION the version, and the two URL fields repointed to GitLab man/raddr-package.Rd regenerated for those URLs NEWS.md a 0.1.2 section; the "first published" claim moves here R/codes.R `since` defaults to 0.1.2, and comments only besides inst/WORDLIST GitLab tests/ new assertions from 7015880 No executable code changed. Stripping comment and blank lines from `git diff v0.1.1..dev -- R/` leaves an empty diff, so behavior at this tag is identical to behavior at v0.1.1. CHECK STATUS AT THIS TAG R CMD check --as-cran on macOS 26.4.1 / R 4.6.0: 0 errors, 0 warnings, 1 NOTE. GitLab CI, native x86_64 Linux, R release and R-devel: both success (pipeline 2822889532 on merge request !33, at this commit). The NOTE is New submission plus one invalid URL -- https://gitlab.com/bart-turczynski/raddr/-/issues, status 404. GitLab serves 404 for the /-/issues path of any project to a signed-out client; re-measured 2026-09-05 against gitlab.com/gitlab-org/gitlab, which answers the same way, while this project's own root returns 200. The address is correct and is kept. CRAN.R-project.org/package=raddr is deliberately absent from URL:. It 404s until acceptance, re-measured 2026-09-05. It goes in afterwards. NOT YET SUBMITTED. Windows via win-builder has not been re-run on this tarball; the prior runs in docs/win-builder.md were against raddr_0.1.1.tar.gz. cran-comments.md lists what is still owed (RADD-yrppvxdi).
-
v0.1.1
protectedeb5d5c43 · ·raddr 0.1.1 First published release. 0.1.0 was tagged during development on 2026-07-31 and never submitted anywhere; v0.1.0 is retained as an honest record of what was checked that day rather than moved, and this tag supersedes it. The difference from v0.1.0 is two dependency floors, one of which is a correctness fix: Imports: vctrs (>= 0.7.0) -- vctrs < 0.7.0 modifies vctrs_rcrd types in place in vec_assign(), and raddr_address is a vctrs_rcrd, so reading one dialect off a parse corrupted every later read of the same vector (RADD-vppmbsia). Pinned by a non-mutation test, not by the floor alone. Imports: rlang (>= 1.1.7) -- real but previously inherited from vctrs rather than declared. Both floors are measured rather than declared: data-raw/check-dep-floor.sh runs R CMD check --as-cran against vctrs at exactly 0.7.0 on R 4.0.0, the declared R floor. Status OK, transcript in docs/dep-floor-check.md. Not yet on CRAN. DESCRIPTION's URL and BugReports still 404 while the GitHub account is suspended, which is the remaining submission blocker (RADD-yrppvxdi). -
v0.1.0
protected7fc7b8b2 · ·raddr 0.1.0 First release. Retagged from 4239c75: that commit was tagged 0.1.0 but never submitted or pushed anywhere, and 25 commits of Epic N and later work landed after it. Since nothing was ever published under the old tag, 0.1.0 is this.