Reverse depends check results in preparation of 1.0.14

[eddelbuettel]

January is around the corner, and a new release date for Rcpp is coming. So I ran another set of reverse depends (summary available as usual). (And I am currently running a follow-up run partly in response to a new r-devel issue). This finished run had a couple of i) missing dependencies for newer packages, ii) known unrelated issues (here: BH and RcppEnsmallen upgrades leaving some packages in need of love), iii) 'random' (?) network issues, iv) packages also failing at CRAN we can ignore and v) residuals to lok at.

Here we see two residuals:

• recurse (version 1.4.0, maintained by @cbracis, emailed 2024-12-21)
  (IMHO a false positive, sent PR on 2025-01-11 to address)
• symengine (version 0.2.6, maintained by @Marlin-Na, emailed 2024-12-21)
  (not an issue at CRAN on 2025-01-12 in their reverse-depends checks but consistently for me to maybe just "fragile")

I will attempt to also email the maintainers with a copy of the error I saw and will ask for help but as a first step I also include the errors below under 'details'.

Release candidates of Rcpp can as always be installed via

install.packages('Rcpp', repos = c('https://rcppcore.r-universe.dev', 'https://cloud.r-project.org'))

as detailed in the R-universe page for Rcpp.

recurse

One test fails on Debian with g++-14 in

* checking tests ... [15s/15s] ERROR
  Running ‘testthat.R’ [14s/15s]
Running the tests in ‘tests/testthat.R’ failed.
Last 13 lines of output:
  [ FAIL 1 | WARN 0 | SKIP 2 | PASS 147 ]
  
  ══ Skipped tests (2) ═══════════════════════════════════════════════════════════
  • empty test (2): 'test_recurse.R:116:1', 'test_recurse.R:181:1'
  
  ══ Failed tests ════════════════════════════════════════════════════════════════
  ── Error ('test_recurse.R:231:7'): polygon ─────────────────────────────────────
  Error in `UseMethod("getRecursionsInPolygon", trajectory)`: no applicable method for 'getRecursionsInPolygon' applied to an object of class "NULL"
  Backtrace:
      ▆
   1. └─recurse::getRecursionsInPolygon(createMove2Obj(track), polyc) at test_recurse.R:231:33
  
  [ FAIL 1 | WARN 0 | SKIP 2 | PASS 147 ]
  Error: Test failures
  Execution halted
* checking for unstated dependencies in vignettes ... OK

symengine

One test fails on Debian with g++-14 in

* checking tests ... [21s/22s] ERROR
  Running ‘testthat.R’ [21s/22s]
Running the tests in ‘tests/testthat.R’ failed.
Last 13 lines of output:
    'test-vecbasic_s4bindings.R:9:1'
  • symengine_have_component("llvm") is not TRUE (2):
    'test-double_visitor.R:23:5', 'test-double_visitor.R:34:5'
  
  ══ Failed tests ════════════════════════════════════════════════════════════════
  ── Error ('test-ODESystem.R:20:5'): ODESystem and predict works ────────────────
  Error in `Rcpp::sourceCpp(code = rcpp_code, env = compile_sys_env, verbose = FALSE)`: Error 1 occurred building shared library.
  Backtrace:
      ▆
   1. └─symengine::ODESystem(eq1, eq2) at test-ODESystem.R:20:5
   2.   └─Rcpp::sourceCpp(code = rcpp_code, env = compile_sys_env, verbose = FALSE)
  
  [ FAIL 1 | WARN 0 | SKIP 11 | PASS 644 ]
  Error: Test failures
  Execution halted
* checking for unstated dependencies in vignettes ... OK

This could of course also be another required package being out of sync.

[eddelbuettel]

Rcpp 1.0.14 is now getting onto CRAN without any rev.dep issues ¹ so closing this. @Marlin-Na what I reported happens on the machine I tested on but apparently not on the CRAN machine that matters :) If you have some time maybe we can try to sort out why it fails for me? It seems Boost related...

Footnotes

1. There was one other false positive / repeat occurence that also happens with 1.0.13 and is hence ignored. ↩