Skip to content

cran-comments.md

Latest commit

 

History

History
52 lines (45 loc) · 2.61 KB

cran-comments.md

File metadata and controls

52 lines (45 loc) · 2.61 KB
Raw

Test environments

  • local macOS (Darwin 24.6.0), R 4.5.x
  • GitHub Actions: ubuntu-latest (devel, release, oldrel-1), windows-latest (release), macOS-latest (release)
  • ggplot2-devel job (deprecation-as-error)

R CMD check results

There were no ERRORs or WARNINGs.

There was 1 NOTE:

  • "checking for future file timestamps ... unable to verify current time" — this is an environment issue (the check host could not reach the remote time server) and is unrelated to the package.

Downstream dependencies

survminer has 72 reverse dependencies on CRAN. This release is additive: new functionality is provided through new functions and opt-in arguments, and the default output of existing functions is unchanged.

I ran a scoped reverse-dependency check: I downloaded all 72 source packages and searched them for the small set of functions whose behaviour changed this cycle (surv_pvalue() gained two output columns; ggcompetingrisks() was modernized with its default output unchanged; ggcoxfunctional() no longer warns on the formula form). Four reverse dependencies call one of these -- clustcurv, Coxmos and ggquickeda use surv_pvalue(), and simstudy uses ggcompetingrisks(). I checked each: all read surv_pvalue() results by column name (the new columns are additive), and simstudy's ggcompetingrisks() call reproduces unchanged. No problems are introduced by this update.

Update

This is a major update (0.5.2 -> 1.0.0; see NEWS.md), consolidating a large body of new functionality while keeping existing defaults backward compatible. Key additions:

  • One-call publication presets: ggsurvplot(preset = "publication" | ...) plus the theme_surv_classic/minimal/bold() companions.
  • Modern effect measures: restricted mean survival time (ggrmst(), ggrmst_difference()), milestone survival (ggmilestone()), and landmark analysis (gglandmark()), all computed internally (no new run-time dependency).
  • Cox-model tools: subgroup and multi-model forest plots (ggforest_subgroup(), ggforest_models()), a non-proportional-hazards diagnostic panel (ggcoxnph()), and ggadjustedcurves() gains a hazard-ratio annotation and a number-at-risk table.
  • A composable single-ggplot return (ggsurvplot(output = "ggplot")), parametric-over-KM overlays (ggsurvparametric()), a CDISC ADTTE helper (surv_adtte()), a batch weighted-log-rank wrapper (weighted_logrank()), reverse-KM median follow-up (surv_median_followup()), and plotmath support in legend.labs.
  • Modernized competing-risks visualization (ggcompetingrisks()).
  • New teaching vignettes and various bug fixes and deprecation cleanups.